Browse the corpus

not available in India but can be designed and fabricated by the stroke survivors themselves and their caregivers, for example, the universal cuff that can assist the stroke survivor to use their affected hand for feeding, brushing, writing and grooming. The key topics covered under this section is provided in table 5. The primary objective of this section is to inform the stroke survivors of the importance of assistive devices that can be used to perform everyday activities independently and safely. Assistive devices can boost the confidence of the stroke survivor to engage in their everyday tasks. It also reduces the assistance and support provided by the caregivers, thereby reducing the physical strain in providing care and support for the stroke survivor in their daily living tasks. Description of the intervention Naming the application This application was intended to educate the stroke survivors and their caregivers to manage their physical disabilities following stroke. Therefore, the web-based application was named ‘Care for Stroke’ to emphasise the importance of enhancing the life of individuals experiencing stroke and continuum of care that is essential for a stroke survivor.

o educate the stroke survivors and their caregivers to manage their physical disabilities following stroke. Therefore, the web-based application was named ‘Care for Stroke’ to emphasise the importance of enhancing the life of individuals experiencing stroke and continuum of care that is essential for a stroke survivor. Logo and tagline of the application The logo of the application was created by the principal investigator himself under the supervision of experts from the field of disability, rehabilitation and design (figure 1). The logo depicts a stroke survivor accepting support from another person in a home environment and trying to mobilise himself/herself. The design of the logo stresses the importance of the stroke survivor accepting support from another person and actively engaging in functional activities while staying at home. Figure 1 Logo of the application. The tagline of the application is ‘Think Smart—Take Control’. This tagline emphasises the importance of proactive, innovative and smart planning for therapy and rehabilitation services that the stroke survivor and their caregivers should execute, outside the hospital environment. It also encourages the stroke survivors to take control of their problems following stroke and work towards an independent life after a stroke.

Introduction Obstetricians have encouraged early, or even immediate, cord clamping since the 1960s as part of the active management of the third stage of labour to prevent maternal haemorrhage.1 For term births, timing of cord clamping does not have a clear effect on postpartum blood loss. However, immediate clamping reduces neonatal haemoglobin and neonatal jaundice.2 Although these effects are generally transient and well tolerated, reduced iron levels are still seen at age 3–6 months, an effect associated with developmental delay.3 4 The WHO,5 the International Federation of Gynaecology and Obstetrics,6 and the National Institute for Health and Care Excellence (NICE) in the UK7 now recommend deferring cord clamping. In the UK, about a third of newborn babies are attended at birth by neonatal resuscitation staff. For most, all that happens is an assessment, stimulation, thermal care and simple airway management. Around 15% of babies receive active resuscitation at birth, such as mask ventilation, intubation, cardiac massage or drug administration.

a third of newborn babies are attended at birth by neonatal resuscitation staff. For most, all that happens is an assessment, stimulation, thermal care and simple airway management. Around 15% of babies receive active resuscitation at birth, such as mask ventilation, intubation, cardiac massage or drug administration. When a baby requires assessment or stabilisation and support at birth, standard practice heretofore is to clamp and cut the cord immediately and then take the baby to a resuscitation platform, usually at the side of the room or in another room. This means that the period of transitional circulation is foreshortened. In addition, parents are often not able to see or touch their baby at birth.8 Evidence from other areas of adult and child resuscitation has shown that family presence is preferred by relatives and by staff9–13 and is now standard in these settings. This issue has not been explored for care after birth, and to date newborn resuscitation has always been away from the woman and her partner.

irth.8 Evidence from other areas of adult and child resuscitation has shown that family presence is preferred by relatives and by staff9–13 and is now standard in these settings. This issue has not been explored for care after birth, and to date newborn resuscitation has always been away from the woman and her partner. The need for immediate neonatal care and support increases with increasing prematurity. For preterm infants, deferred cord clamping is associated with decreased transfusion for anaemia, decreased low blood pressure requiring inotropic support and less low-grade intraventricular haemorrhage compared with immediate clamping.14 There is an increase in jaundice, but the long-term effects are unclear. Systematic review suggests that strategies for increased placental transfusion after birth may improve neonatal mortality rate in very premature babies,15 but it is uncertain whether any benefits would be negated by delayed resuscitation,16 and the current European Resuscitation Council recommendation therefore states that “for babies requiring resuscitation, resuscitative intervention remains the priority.”17

y improve neonatal mortality rate in very premature babies,15 but it is uncertain whether any benefits would be negated by delayed resuscitation,16 and the current European Resuscitation Council recommendation therefore states that “for babies requiring resuscitation, resuscitative intervention remains the priority.”17 Very preterm babies are those most likely to experience these major morbidities and, potentially, might benefit from deferred clamping. It therefore seems important to develop a strategy to allow initial resuscitation with the cord intact. A survey of the UK Extended Neonatal Network in 2009 to assess views on timing of cord clamping and placental transfusion found that two-thirds of those surveyed thought that initial care with the cord intact is potentially feasible for preterm vaginal births. Just under half thought this was possible for preterm caesarean births (Duley, personal communication, 2009). The aim of this paper is to describe development and preliminary testing of a mobile trolley to enable newborn care and support at the bedside, potentially with an intact cord.

Very preterm babies are those most likely to experience these major morbidities and, potentially, might benefit from deferred clamping. It therefore seems important to develop a strategy to allow initial resuscitation with the cord intact. A survey of the UK Extended Neonatal Network in 2009 to assess views on timing of cord clamping and placental transfusion found that two-thirds of those surveyed thought that initial care with the cord intact is potentially feasible for preterm vaginal births. Just under half thought this was possible for preterm caesarean births (Duley, personal communication, 2009). The aim of this paper is to describe development and preliminary testing of a mobile trolley to enable newborn care and support at the bedside, potentially with an intact cord. Methods Initial concept Resuscitation with an intact cord has been described by several authors,18–21 but no agreed technique had emerged. In January 2010, DH convened a 1-day meeting in Worcester of eight UK clinicians and researchers (AB, SB, LD, AG, AMH, DH, DO and ADW - see Acknowledgement) with a common interest in practical methods for initiating neonatal assessment and resuscitation at birth before the cord is clamped and cut. From this came the idea of a small mobile bedside resuscitation trolley, extending the concept of the platform that had been used for a cord clamping study in Glasgow (figure 1).18 The acronym BASICS for ‘Bedside Assessment, Stabilisation and Initial Cardiorespiratory Support’ was proposed, and this established a working title of the ‘BASICS trolley’. The key concepts were that this trolley would need to keep the baby warm, allow suction and respiratory support within a 50 cm radius from the mother's uterus. ADW drew up designs during the meeting (figure 2). It was agreed that he would formalise these, take the design of the trolley forward with the medical engineering department in Liverpool and register the design rights.

the baby warm, allow suction and respiratory support within a 50 cm radius from the mother's uterus. ADW drew up designs during the meeting (figure 2). It was agreed that he would formalise these, take the design of the trolley forward with the medical engineering department in Liverpool and register the design rights. Figure 1 The Glasgow trolley used in the study by Aladangady et al18 that inspired the first designs of the Bedside Assessment, Stabilisation and Initial Cardiorespiratory Support (BASICS) trolley (photograph courtesy of AMH). Figure 2 The design drawing that came out of discussions at the first group meeting in Worcestershire showing (A) the trolley design and (B) the way in which it might be positioned over the operating table at caesarean section (drawings in February 2010 by ADW, reprinted with his permission).

Figure 1 The Glasgow trolley used in the study by Aladangady et al18 that inspired the first designs of the Bedside Assessment, Stabilisation and Initial Cardiorespiratory Support (BASICS) trolley (photograph courtesy of AMH). Figure 2 The design drawing that came out of discussions at the first group meeting in Worcestershire showing (A) the trolley design and (B) the way in which it might be positioned over the operating table at caesarean section (drawings in February 2010 by ADW, reprinted with his permission). Development of the prototypes Funding to support development of the concept into a prototype was secured from the National Institute of Health Research (NIHR; LD), with additional support from the Liverpool Women's Hospital ‘Newborn Appeal’. In collaboration with the Department of Clinical Engineering at the Royal Liverpool University Hospital, the first prototype was developed by PW, using a modified hospital overbed table (figure 3). This first early prototype was developed with support from the development team (LD, AG, DH, PW, ADW and CWY), which met in Liverpool, and included input from a service user representative. Various mock delivery scenarios were staged to assess the optimum size and reach of the trolley platform. Discussions included what equipment needed to be available on the trolley to provide blended oxygen and air and to provide positive end-expiratory pressure. The Liverpool prototype (figure 4) was awarded ‘Best Redesign in Cardiovascular Medicine’ at the Medical Futures Awards in June 2011.

each of the trolley platform. Discussions included what equipment needed to be available on the trolley to provide blended oxygen and air and to provide positive end-expiratory pressure. The Liverpool prototype (figure 4) was awarded ‘Best Redesign in Cardiovascular Medicine’ at the Medical Futures Awards in June 2011. Figure 3 Drawing of the trolley's position for normal delivery (drawing by PW in March 2011, reprinted with his permission). Figure 4 The first Bedside Assessment, Stabilisation and Initial Cardiorespiratory Support (BASICS) prototype (by PW in April 2011, reprinted with his permission). Note that the problem of how to keep the baby warm was still not resolved. A key problem for the first prototype was how to keep the baby warm. Traditional resuscitaires have an overhead heater, but this was not practical for a trolley designed to go under the legs of a woman in lithotomy position or over an operating table. A commercial company (Inditherm plc) was approached to adapt their heated mattresses for the trolley. The company then agreed to collaborate on further development, with the aim of taking the product to market. The second prototype, developed in collaboration with Inditherm and manufactured by them (figure 5), has improved mobility and flexibility that allows the height and position of the trolley platform to be adjusted. This trolley was marked with the Conformité Européenne (CE) logo in October 2012 and is now marketed by Inditherm as ‘LifeStart’. Figure 5 The LifeStart trolley manufactured by Inditherm (October 2012, with permission).

A key problem for the first prototype was how to keep the baby warm. Traditional resuscitaires have an overhead heater, but this was not practical for a trolley designed to go under the legs of a woman in lithotomy position or over an operating table. A commercial company (Inditherm plc) was approached to adapt their heated mattresses for the trolley. The company then agreed to collaborate on further development, with the aim of taking the product to market. The second prototype, developed in collaboration with Inditherm and manufactured by them (figure 5), has improved mobility and flexibility that allows the height and position of the trolley platform to be adjusted. This trolley was marked with the Conformité Européenne (CE) logo in October 2012 and is now marketed by Inditherm as ‘LifeStart’. Figure 5 The LifeStart trolley manufactured by Inditherm (October 2012, with permission). BASICS trolley design A key element in the design is flexibility to allow the baby to be placed on the trolley while the umbilical cord is still intact. This required the trolley platform to be manoeuvrable: able to be lowered under the mother's legs when in lithotomy position or raised up high to reach over the woman on an operating table. The umbilical cord can be short, so the platform has to be able to be placed close to the woman's introitus. This is achieved by having a central pillar that can be adjusted up or down, and a narrow platform that reaches out from the pillar with raised edges for safety. Even if the mother delivers the baby on the floor, the trolley can be moved to the site of birth and provide the necessary resuscitation gases and suction for the baby to remain attached and with the mother.

can be adjusted up or down, and a narrow platform that reaches out from the pillar with raised edges for safety. Even if the mother delivers the baby on the floor, the trolley can be moved to the site of birth and provide the necessary resuscitation gases and suction for the baby to remain attached and with the mother. Thermal support is provided by the CosyTherm electric heated mattress. This is adjustable through a range of temperatures and takes only a few minutes to heat up. The trolley has a timer built into the platform, with audible Apgar reminder bleeps at 1 and 5 min. Fixed around the central pillar are two universal Medirails for additional equipment. These allow each hospital to customise the trolley to their specific requirements. Our early assessments used a Tom Thumb Infant Resuscitator (Viamed, Keighley, UK), oxygen blender (Inspiration Health Care Ltd., Leicestershire, UK), a suction bottle driven by the wall-supplied air supply (Oxylitre Ltd., Manchester, UK) and the control unit for the CosyTherm heated mattress (Inditherm, Rotherham, UK). We connect it with hoses to the air and oxygen wall supply. Other equipment that could be added is a light on a flexible stalk, a saturation monitor or a storage basket (for keeping the laryngoscope, hats, saturation monitor). Results The trolley was introduced into clinical service at Liverpool Women's Hospital after CE marking. We describe our early experiences here. More formal evaluation of safety and acceptability to clinicians and to parents has been reported separately.22

Fixed around the central pillar are two universal Medirails for additional equipment. These allow each hospital to customise the trolley to their specific requirements. Our early assessments used a Tom Thumb Infant Resuscitator (Viamed, Keighley, UK), oxygen blender (Inspiration Health Care Ltd., Leicestershire, UK), a suction bottle driven by the wall-supplied air supply (Oxylitre Ltd., Manchester, UK) and the control unit for the CosyTherm heated mattress (Inditherm, Rotherham, UK). We connect it with hoses to the air and oxygen wall supply. Other equipment that could be added is a light on a flexible stalk, a saturation monitor or a storage basket (for keeping the laryngoscope, hats, saturation monitor). Results The trolley was introduced into clinical service at Liverpool Women's Hospital after CE marking. We describe our early experiences here. More formal evaluation of safety and acceptability to clinicians and to parents has been reported separately.22 The first use was for an ex utero intrapartum treatment procedure at caesarean section in October 2012. Feedback from staff and parents was positive. Use was then extended to low-risk deliveries. All parents appeared appreciative of being with their baby during the first moments of life. Clinicians have been slower to accept initial bedside care, and expressed anxieties about the parents being so close to the baby during resuscitation. There have also been practical issues to resolve, particularly in the operating theatre where tubing for air and oxygen can be a trip hazard.

the first moments of life. Clinicians have been slower to accept initial bedside care, and expressed anxieties about the parents being so close to the baby during resuscitation. There have also been practical issues to resolve, particularly in the operating theatre where tubing for air and oxygen can be a trip hazard. It has also been clear that introducing the trolley for initial care and stabilisation at birth requires teamwork. The equipment must be checked and moved into position ready for the birth. Positioning of the trolley and neonatal team close to the woman needs negotiation with midwifery and obstetric teams. Presently, therefore, the neonatal team are called earlier than for traditional care at the roomside. Once the baby is born, the midwifery and obstetric teams need to understand what is happening during care at the bedside, and will often have a role in supporting the woman and her partner, and explaining to them what is happening.

erefore, the neonatal team are called earlier than for traditional care at the roomside. Once the baby is born, the midwifery and obstetric teams need to understand what is happening during care at the bedside, and will often have a role in supporting the woman and her partner, and explaining to them what is happening. When using the trolley at a caesarean section, a sterile Mayo tray cover is draped over the platform. The neonatal team member who will provide care at birth scrubs and puts on a sterile gown and gloves so as to prevent contamination of the operating field. Once the baby is on the trolley platform, the sterility rules are relaxed, so that the scrubbed neonatal team member can touch and use non-sterile equipment such as the laryngoscope, timer and hat. It is then important that the sterile field is not re-entered or contaminated. If ongoing resuscitation is required, then, after the cord is cut, the trolley can be moved away from the operating table to allow care to continue without interrupting the surgeon. Discussion Main findings We have developed a mobile neonatal resuscitation trolley to allow newborn babies to be assessed and resuscitated alongside their mothers with, or without, an intact cord.

When using the trolley at a caesarean section, a sterile Mayo tray cover is draped over the platform. The neonatal team member who will provide care at birth scrubs and puts on a sterile gown and gloves so as to prevent contamination of the operating field. Once the baby is on the trolley platform, the sterility rules are relaxed, so that the scrubbed neonatal team member can touch and use non-sterile equipment such as the laryngoscope, timer and hat. It is then important that the sterile field is not re-entered or contaminated. If ongoing resuscitation is required, then, after the cord is cut, the trolley can be moved away from the operating table to allow care to continue without interrupting the surgeon. Discussion Main findings We have developed a mobile neonatal resuscitation trolley to allow newborn babies to be assessed and resuscitated alongside their mothers with, or without, an intact cord. Strengths and limitations Clinicians and researchers who wish to implement and evaluate deferred cord clamping for all births have previously faced a problem: How do you achieve this in those babies who require immediate resuscitation? This trolley has demonstrated the feasibility of providing initial neonatal care and stabilisation at the bedside, and to do this while the cord is intact. The trolley now requires evaluation to assess its safety in routine clinical service, and its acceptability to clinicians and parents. Although parents of premature babies describe their first physical contact as very significant, it is frequently delayed until some time after birth on the special care baby unit.8 With the BASICS trolley, the mother can touch and speak to her newborn baby while it is undergoing assessment, and our initial experience is that parents value this opportunity.22

cribe their first physical contact as very significant, it is frequently delayed until some time after birth on the special care baby unit.8 With the BASICS trolley, the mother can touch and speak to her newborn baby while it is undergoing assessment, and our initial experience is that parents value this opportunity.22 Introducing neonatal care at the bedside requires a multidisciplinary approach. The attending team need to prepare for, and be confident in, providing neonatal resuscitation under more direct intense parental scrutiny. Everyone present at the birth need to understand their role in supporting this care strategy, and supporting the woman and her partner. Currently ‘deferred’ cord clamping for very premature babies occurs after 30–45 s, even though placental transfusion is likely to continue for longer. This reflects the desire not to delay initial neonatal care and support. The BASICS trolley should allow evaluation of a more physiological approach to cord clamping for all births, including sick and very premature babies. An ongoing pilot randomised trial compares cord clamping within 20 s with clamping after at least 2 min for births below 32 weeks’ gestation (ISRCTN21456601). In this study, some units are using the BASICS trolley while others are using their normal resuscitaire moved to the bedside.19 21 The pilot is to assess feasibility of a large UK randomised trial.

mpares cord clamping within 20 s with clamping after at least 2 min for births below 32 weeks’ gestation (ISRCTN21456601). In this study, some units are using the BASICS trolley while others are using their normal resuscitaire moved to the bedside.19 21 The pilot is to assess feasibility of a large UK randomised trial. Conclusion This is the first mobile resuscitation unit designed specifically to facilitate newborn resuscitation at the bedside, with an intact cord. Further evaluation will assess safety and acceptability, and whether resuscitation with an intact cord to allow longer for the neonatal circulation to be established improves outcome.

is is the first mobile resuscitation unit designed specifically to facilitate newborn resuscitation at the bedside, with an intact cord. Further evaluation will assess safety and acceptability, and whether resuscitation with an intact cord to allow longer for the neonatal circulation to be established improves outcome. The authors thank Worcestershire Royal Hospital who hosted the first development meeting. They thank Inditherm and their managing director Nick Bettles (NB) for their work in developing the fully functional CE marked mobile resuscitaire from the initial prototype. They thank David Odd (DO) and Ms Rabia Imtiaz who were part of the initial design meeting in Worcestershire Royal Hospital. Gill Gyte (GG) attended the Liverpool design meetings and provided important input into the trolley design. Tony Fisher, Head of Department of Clinical Engineering at the Royal Liverpool University Teaching Hospital, provided important advice about intellectual property as well as agreeing to develop the BASICS prototype in his department at a discounted cost. David Hutchon (DH) convened the original meeting at Worcestershire Royal Hospital to explore practical methods for neonatal resuscitation with an intact cord, and was named as one of the inventors on the original registration. He was also a member of the team for the Medical Futures award in 2010, provided input into the design of the BASICS trolley prototype, and first identified Inditherm as a commercial partner. Twitter: Follow Andrew Weeks at @adweeks and Amanda Burleigh at @delayedclamping

The authors thank Worcestershire Royal Hospital who hosted the first development meeting. They thank Inditherm and their managing director Nick Bettles (NB) for their work in developing the fully functional CE marked mobile resuscitaire from the initial prototype. They thank David Odd (DO) and Ms Rabia Imtiaz who were part of the initial design meeting in Worcestershire Royal Hospital. Gill Gyte (GG) attended the Liverpool design meetings and provided important input into the trolley design. Tony Fisher, Head of Department of Clinical Engineering at the Royal Liverpool University Teaching Hospital, provided important advice about intellectual property as well as agreeing to develop the BASICS prototype in his department at a discounted cost. David Hutchon (DH) convened the original meeting at Worcestershire Royal Hospital to explore practical methods for neonatal resuscitation with an intact cord, and was named as one of the inventors on the original registration. He was also a member of the team for the Medical Futures award in 2010, provided input into the design of the BASICS trolley prototype, and first identified Inditherm as a commercial partner. Twitter: Follow Andrew Weeks at @adweeks and Amanda Burleigh at @delayedclamping Contributors: The initial trolley concept and design was developed by AB, SB, LD, AG, AMH, DH, DO and ADW at the meeting organised by DH in Worcester. Subsequent meetings, attended by AB, NB, LD, AG, GG, DH, ADW, PW and CWY led to the development of the prototype. ADW wrote the first draft of the manuscript, which was subsequently revised by AB, SB, LD, AG, AMH, ADW, PW and CWY. All authors approved the final version of the manuscript.

g organised by DH in Worcester. Subsequent meetings, attended by AB, NB, LD, AG, GG, DH, ADW, PW and CWY led to the development of the prototype. ADW wrote the first draft of the manuscript, which was subsequently revised by AB, SB, LD, AG, AMH, ADW, PW and CWY. All authors approved the final version of the manuscript. Funding: This paper presents independent research funded by the National Institute for Health Research (NIHR) under its Programme Grants for Applied Research funding scheme (RP–PG–0609-10107). Additional support came from the Department of Clinical Engineering who offered their services at reduced cost, and the Liverpool Women's ‘Newborn Appeal’. AB and DH provided personal funding to attend all meetings and the Medical Futures awards. Inditherm Ltd manufactured the final trolley and obtained its CE marking. Competing interests: It has been agreed that any financial remuneration to the original development team will be donated to charity. LD is the chief investigator for a trial comparing alternative strategies for timing of cord clamping, for which this trolley is one strategy for providing care at the bedside. Ethics approval: The commercially produced trolley (with CE mark) was subsequently introduced into clinical service at Liverpool Women's Hospital following approval from the Medical Devices Committee, the head of Research and Development, and the Clinical Director for the Neonatology Directorate. Provenance and peer review: Not commissioned; externally peer reviewed.

Ethics approval: The commercially produced trolley (with CE mark) was subsequently introduced into clinical service at Liverpool Women's Hospital following approval from the Medical Devices Committee, the head of Research and Development, and the Clinical Director for the Neonatology Directorate. Provenance and peer review: Not commissioned; externally peer reviewed. Data sharing statement: Unpublished technical files on the development of the trolley are available on application from the authors.

Background Stroke rehabilitation is a process targeted towards restoration or maintenance of the physical, mental, intellectual and social abilities of an individual affected by stroke.1 Stroke rehabilitation enables the stroke survivor to perform his/her daily activities at an optimal functional level and helps the stroke survivor to participate in his/her social roles as independently as possible.2 The stroke survivor relearns the skills that are lost or impaired due to brain damage following stroke through rehabilitation.3

ables the stroke survivor to perform his/her daily activities at an optimal functional level and helps the stroke survivor to participate in his/her social roles as independently as possible.2 The stroke survivor relearns the skills that are lost or impaired due to brain damage following stroke through rehabilitation.3 An insult to the human brain due to stroke might have various effects on the stroke survivor, and hence healthcare professionals from various disciplines have to provide the stroke survivor with a patient-centred, comprehensive, multidisciplinary rehabilitation.4 Unlike high-income countries (HICs), the resources for rehabilitation, especially the rehabilitation workforce and infrastructure, are very limited in many low and middle-income countries (LMICs).5 If we take India as an example, rehabilitation services are often unidisciplinary, driven predominantly by physiotherapists, with lack of support from occupational therapists, speech therapists and so on. Many government-run district rehabilitation centres are non-functional and the private hospitals are staffed with only a physiotherapist in their rehabilitation centres.6 Given the scarce resources, the rehabilitation needs of the stroke survivors, especially in the LMICs, remain largely unmet.7 Provision of cost-effective, post-stroke multidisciplinary rehabilitation services for the stroke survivors therefore becomes crucial to address the unmet needs and growing magnitude of disability experienced by the stroke survivors in LMICs.

eeds of the stroke survivors, especially in the LMICs, remain largely unmet.7 Provision of cost-effective, post-stroke multidisciplinary rehabilitation services for the stroke survivors therefore becomes crucial to address the unmet needs and growing magnitude of disability experienced by the stroke survivors in LMICs. The past few years have seen a tremendous increase in the use of Smartphones by health professionals and also by the general public.8 Evidence from a recent systematic review suggests that Smartphones could be an extremely useful tool to educate patients to manage their health problems.9 Another systematic review on the use of Smartphone applications for stroke rehabilitation also demonstrates the advantages of Smartphone applications for provision of stroke-related information.10 These Smartphone applications are regarded as important by health professionals providing stroke rehabilitation themselves.10 In order to meet the growing need for post-stroke rehabilitation services in India, we developed a web-based Smartphone-enabled educational intervention for management of physical disabilities following stroke. This paper provides a detailed description of the intervention and the processes involved in its development. The paper also discusses the importance of such rehabilitation interventions for meeting the unmet needs of the stroke survivors.

ucational intervention for management of physical disabilities following stroke. This paper provides a detailed description of the intervention and the processes involved in its development. The paper also discusses the importance of such rehabilitation interventions for meeting the unmet needs of the stroke survivors. Development of the content for the intervention Systematic review of the available interventions Evidence from systematic reviews in relation to stroke rehabilitation and information provision for stroke survivors and caregivers was extensively used to develop the intervention. We also conducted a comprehensive and a global systematic review on educational interventions for reducing disabilities in acquired brain injury to investigate the evidence that was available to develop this intervention.

on provision for stroke survivors and caregivers was extensively used to develop the intervention. We also conducted a comprehensive and a global systematic review on educational interventions for reducing disabilities in acquired brain injury to investigate the evidence that was available to develop this intervention. Rehabilitation needs assessment study The content of the intervention was developed primarily based on the needs expressed by the stroke survivors and caregivers who participated in a rehabilitation needs assessment study carried out exclusively to develop this intervention. The rehabilitation needs assessment study was carried out to guide the development of a need-based rehabilitation intervention and had two components in it. One was a structured survey with 50 stroke survivors and their caregivers to identify the various kinds of rehabilitation needs that they experience. The other was a detailed in-depth interview with a subsample of the stroke survivors and caregivers selected for the survey. The purpose of the in-depth interviews was to gain a detailed understanding of the experiences of the stroke survivors in relation to accessing stroke rehabilitation services and their rehabilitation needs following a stroke. In-depth interviews with health professionals involved in the provision of stroke rehabilitation services were also carried out to understand the perspective of the health professionals about provision of stroke rehabilitation services, their knowledge about the existing Smartphone-based health interventions and their attitudes and opinions about the use of a Smartphone-enabled, care-supported education programme for domiciliary stroke rehabilitation.

rstand the perspective of the health professionals about provision of stroke rehabilitation services, their knowledge about the existing Smartphone-based health interventions and their attitudes and opinions about the use of a Smartphone-enabled, care-supported education programme for domiciliary stroke rehabilitation. Expert group for content development In addition to the needs assessment, expert guidance was obtained from a team of eight highly qualified and experienced health professionals from various neurorehabilitation disciplines (physical medicine and rehabilitation, neuropsychiatry, clinical psychology, occupational therapy, physiotherapy, social sciences, information technology, public health and m-health) with both national and global expertise in the field of neuropsychiatric rehabilitation. The expert team also included three stroke survivors and their primary caregivers. All the team members were from Tamil Nadu and they were Tamil-speaking. The key characteristics of the expert group, such as their experience, expertise, global exposure and language, facilitated the development of a culturally specific, patient-centred intervention for management of physical disabilities following a stroke.

e team members were from Tamil Nadu and they were Tamil-speaking. The key characteristics of the expert group, such as their experience, expertise, global exposure and language, facilitated the development of a culturally specific, patient-centred intervention for management of physical disabilities following a stroke. Framework of the intervention content On the basis of the findings from the rehabilitation needs assessment study, guidance from the expert group, and available evidence from systematic reviews, the framework of the intervention content was designed. The content framework included five important sections related to post-stroke rehabilitation. The sections were: Information about stroke (know more about stroke) Exercises (home-based exercises) Functional skills training (preparing oneself for daily living) Activities of daily living (engaging in activities of daily living) Assistive devices (devices to assist daily living). Content of the intervention sections Know more about stroke As the section title suggests, this section enables the stroke survivors and caregivers to know more about stroke, the impact of stroke on an individual experiencing it and advice from experts on the way forward (life after a stroke). The important subsections/topics and videos that this section includes are provided in table 1. Table 1 Content of ‘know more about stroke’ section Content of the intervention Main sections Subsections Videos Information about stroke What is a stroke? What is a transient ischaemic attack (TIA)? How does a stroke happen? Warning signs of a stroke What are the common symptoms of a stroke?

Content of the intervention sections Know more about stroke As the section title suggests, this section enables the stroke survivors and caregivers to know more about stroke, the impact of stroke on an individual experiencing it and advice from experts on the way forward (life after a stroke). The important subsections/topics and videos that this section includes are provided in table 1. Table 1 Content of ‘know more about stroke’ section Content of the intervention Main sections Subsections Videos Information about stroke What is a stroke? What is a transient ischaemic attack (TIA)? How does a stroke happen? Warning signs of a stroke What are the common symptoms of a stroke? How does a stroke affect your body? Risk factors for a stroke Common effects of a stroke Recovering from a stroke What is a stroke? How does a stroke happen? What is a transient ischaemic attack Symptoms of a stroke Effects of a stroke Modifiable and non-modifiable risk factors for a stroke Effects of a stroke on Balance Bowel and Bladder Thinking Pain Physical problems Sleep and fatigue Sensation Sleep and fatigue Speech and language Swallowing Recovery from a stroke by Public health experts a neuropsychiatrist a neurologist a physiotherapist an occupational therapist a clinical psychologist a disability rights expert The primary objective of having this section is to create awareness and enable the stroke survivors and their caregivers to gain more knowledge about stroke, because this would assist them in preventing recurrent stroke, modifying their lifestyle, making treatment decisions and planning for life after a stroke.

a disability rights expert The primary objective of having this section is to create awareness and enable the stroke survivors and their caregivers to gain more knowledge about stroke, because this would assist them in preventing recurrent stroke, modifying their lifestyle, making treatment decisions and planning for life after a stroke. Home-based exercises This section includes home-based, task-oriented exercises that the stroke survivors can practise in their home in order to maintain or improve their body fitness for functional activities. These exercises are based on eclectic treatment approaches to stroke rehabilitation (motor relearning, functional, neurodevelopmental frame of references for therapy) that enable the stroke survivors to use their affected parts of the body and engage in functional activities. These home-based exercises include the use of equipment like a chair or bed and table that are commonly available in most homes in India. They do not require the purchase of any sophisticated exercise equipment. Principles of safety and risk/hazard prevention have been thoroughly considered while developing this section. Some of the important subsections/topics and videos that this section comprises are listed in table 2. Table 2 Content of ‘Exercises’ section Content of the intervention Main sections Subsections Videos Exercises Upper limb exercises Lower limb exercises Balance exercises Active exercises Exercises to improve upper limb function Passive upper limb exercises Passive lower limb exercises Active-assisted exercises for the lower limb Active exercises for the upper limb

Table 2 Content of ‘Exercises’ section Content of the intervention Main sections Subsections Videos Exercises Upper limb exercises Lower limb exercises Balance exercises Active exercises Exercises to improve upper limb function Passive upper limb exercises Passive lower limb exercises Active-assisted exercises for the lower limb Active exercises for the upper limb Exercises for the trunk Exercises for balance Improving awareness and function of the affected hand The objective of developing this section content is to enable the stroke survivors to understand the relevance of the conscious use of the affected parts of the body following a stroke and also the importance of exercises for engaging in functional activities rather than just exercising and improving the flexibility, strength and movement of the affected body part. Preparing oneself for everyday living Functional skills are a prerequisite to participate in everyday living. One should know how to get up from a lying down position. In order to sit properly and feed or groom, one should know how to transfer from a bed to a chair or a commode for bathing or toileting. These are very important to the stroke survivor who cannot or finds it difficult to move the affected part of his/her body. This section highlights functionally oriented tasks that the stroke survivors can learn in order to participate in their day-to-day activities.

ed to a chair or a commode for bathing or toileting. These are very important to the stroke survivor who cannot or finds it difficult to move the affected part of his/her body. This section highlights functionally oriented tasks that the stroke survivors can learn in order to participate in their day-to-day activities. Exercise training provided by a physiotherapist to the stroke survivors is directly related to the development of functional skills of the individual affected by stroke. Hence, this section stresses the importance of functional skills to participate in everyday living and preparing oneself for everyday living by acquiring functional skills. Some of the important subsections/topics and videos of this section are depicted in table 3. Table 3 Content of ‘Exercises’ section Content of the intervention Main sections Subsections Videos Functional skills training Positioning the stroke survivor in bed and in a chair Bed mobility Transfers Standing up from a sitting Mobility/ambulation training Positioning on the Chair – the Bed –affected side the Bed –unaffected side the Bed—Lying on the back Bed Mobility Rolling on the bed Scooting on the bed Coming up to a sitting Sit to Stand (moderate support) Transfers Independent transfers (bed to chair/wheelchair) Transfers with maximum support Walking

Standing up from a sitting Mobility/ambulation training Positioning on the Chair – the Bed –affected side the Bed –unaffected side the Bed—Lying on the back Bed Mobility Rolling on the bed Scooting on the bed Coming up to a sitting Sit to Stand (moderate support) Transfers Independent transfers (bed to chair/wheelchair) Transfers with maximum support Walking Engaging in activities of daily living This section comprises adaptive methods and techniques to engage in activities of daily living like grooming, bathing, dressing and eating. The stroke survivors can watch, learn and practise these adaptive techniques to independently perform their activities of everyday living. The content of this section is depicted in table 4. Table 4 Content of ‘Activities of daily living’ section Content of the intervention Main sections Subsections Videos Activities of daily living Brushing Feeding Bathing Grooming Dressing Brushing Feeding Bathing Grooming Washing face Wearing a T-shirt Wearing a Shirt Wearing a dhoti/lungi Wearing a pant Wearing a saree Wearing a blouse Undressing

Engaging in activities of daily living This section comprises adaptive methods and techniques to engage in activities of daily living like grooming, bathing, dressing and eating. The stroke survivors can watch, learn and practise these adaptive techniques to independently perform their activities of everyday living. The content of this section is depicted in table 4. Table 4 Content of ‘Activities of daily living’ section Content of the intervention Main sections Subsections Videos Activities of daily living Brushing Feeding Bathing Grooming Dressing Brushing Feeding Bathing Grooming Washing face Wearing a T-shirt Wearing a Shirt Wearing a dhoti/lungi Wearing a pant Wearing a saree Wearing a blouse Undressing This section is very important from the viewpoint of both the stroke survivors and their caregivers. This is because learning to purposefully engage in one's own everyday living seems to be an important need and crucial task for the stroke survivors to independently participate in his/her personal, family and societal roles. Although the stroke survivors learn to do exercises and acquire knowledge to manage their problems post-stroke, the overall objective behind the acquisition of these skills and knowledge is to live a functionally independent life and perform their various roles at home and society actively (table 5). Table 5 Content of ‘Devices to assist daily living’ section Content of the intervention Main sections Subsections Videos Assistive devices Personal care aids Mobility aids Orthoses and supports Personal care aids Mobility aids Orthoses and supports

This section is very important from the viewpoint of both the stroke survivors and their caregivers. This is because learning to purposefully engage in one's own everyday living seems to be an important need and crucial task for the stroke survivors to independently participate in his/her personal, family and societal roles. Although the stroke survivors learn to do exercises and acquire knowledge to manage their problems post-stroke, the overall objective behind the acquisition of these skills and knowledge is to live a functionally independent life and perform their various roles at home and society actively (table 5). Table 5 Content of ‘Devices to assist daily living’ section Content of the intervention Main sections Subsections Videos Assistive devices Personal care aids Mobility aids Orthoses and supports Personal care aids Mobility aids Orthoses and supports Devices to assist daily living This is a unique section that enables the stroke survivors and their caregivers to understand the importance of using assistive devices that are readily available in India and that can assist the stroke survivor to engage in their day-to-day activities independently and also with confidence. This section also includes devices that are tailor-made to the needs of the stroke survivors living in the southern part of India like an adapted saree, Velcro-based blouse, adapted dhoti and lungi, etc. This section also has devices that are not available in India but can be designed and fabricated by the stroke survivors themselves and their caregivers, for example, the universal cuff that can assist the stroke survivor to use their affected hand for feeding, brushing, writing and grooming. The key topics covered under this section is provided in table 5.

proactive, innovative and smart planning for therapy and rehabilitation services that the stroke survivor and their caregivers should execute, outside the hospital environment. It also encourages the stroke survivors to take control of their problems following stroke and work towards an independent life after a stroke. Design of the web-based application This intervention is designed as a web-based application that uses a website as an interface (the front end). The introductory web page of the application is shown in figure 2 Users can access the application not just from Smartphones but also from a computer, PDA, Tablet and even digital television that is connected to the internet using any standard web browser. Some of the key design features of this application are: User interface, content format, language. Figure 2 Introductory page of the ‘care for stroke’ application. User interface An interface enables a user to interact with a system (Smartphone in this instance) to perform a task. For example: Navigating to different web pages in this website enables a user to find the video content that he/she prefers to watch. The users can watch the videos by navigating through user-friendly interfaces such as the touch and slide option which requires the users to either touch or slide the icons (ie, pictures and symbols) and pages in the application to watch the videos they want.

user to find the video content that he/she prefers to watch. The users can watch the videos by navigating through user-friendly interfaces such as the touch and slide option which requires the users to either touch or slide the icons (ie, pictures and symbols) and pages in the application to watch the videos they want. Content format This application is exclusively designed to support digitised audio–visual content. More than 75% of the content of this application is in the form of videos. The users can interact with the images related to the main sections and watch the videos about stroke and the management of physical disability post-stroke through this application. There is very minimal requirement for the users to read written information in this application. Language This application is built with multilingual functionality and it currently supports English and Tamil, the native language of the State of Tamil Nadu in India where it was piloted. Technical description of the application The application is built using a LAMP (Linux, Appache, MySQL, PHP) environment. The user interface of the application was designed using HTML5, CSS3, Bootstrap, Java Script, JQuery, Ajax, Google font API and Touch Swipe. It is to ensure that the user interface acts as a responsive and interactive design. Designing the application with these technologies supports the application to be installed and run on multiple devices like desktop, laptop, IPhone, IPad, Android devices andWindows devices.

uery, Ajax, Google font API and Touch Swipe. It is to ensure that the user interface acts as a responsive and interactive design. Designing the application with these technologies supports the application to be installed and run on multiple devices like desktop, laptop, IPhone, IPad, Android devices andWindows devices. The back end of the application was built in PHP5 (PHP—Hypertext Processor) language. This is to facilitate the user to interact with the database (MySQL) and view the requested information without any difficulty. Given the issues with video streaming in a country like India (ie, very slow internet connectivity and streaming), this application uses Cloud Flare CDN (Content Delivery Network) that enhances the quality and speed of the video streaming while the user is accessing the videos from the application. This application also has an administrator module, where the administrator can monitor all the activities of the users who have logged into the application. It can also generate different types of reports of the user interaction with the application. Some of the key information that could be monitored are: The title of the sections and videos viewed, Duration of the logged in session Date and time of viewing Number of sections and videos watched during a logged in session. Device used for logging in Time spent on application, Geo-location information. Structure and functionality of the application Registered website This web-based application can be accessed from the registered website name http://www.careforstroke.com

Date and time of viewing Number of sections and videos watched during a logged in session. Device used for logging in Time spent on application, Geo-location information. Structure and functionality of the application Registered website This web-based application can be accessed from the registered website name http://www.careforstroke.com Home page The application has a home page that briefly describes stroke and stroke-related disability in the vernacular (Tamil). First-time users cannot access the intervention without registering themselves. This is to ensure proactive engagement of the users, observe their utilisation pattern and to generate utilisation reports for future evaluations. The home page provides details of registration with an icon to register the first-time users. Users who have already registered to access the intervention can use the same icon to access the sign-in page (figure 2). There is a drop-down icon in the home page to change the language of the application if required. Currently, the application pages have the descriptions in English and Tamil. Sign-in page and registration This page contains an icon for first-time users to register and the sign in boxes with user name and password sections to be filled by the user to sign into the application.

Home page The application has a home page that briefly describes stroke and stroke-related disability in the vernacular (Tamil). First-time users cannot access the intervention without registering themselves. This is to ensure proactive engagement of the users, observe their utilisation pattern and to generate utilisation reports for future evaluations. The home page provides details of registration with an icon to register the first-time users. Users who have already registered to access the intervention can use the same icon to access the sign-in page (figure 2). There is a drop-down icon in the home page to change the language of the application if required. Currently, the application pages have the descriptions in English and Tamil. Sign-in page and registration This page contains an icon for first-time users to register and the sign in boxes with user name and password sections to be filled by the user to sign into the application. Registration page This page contains a drop-down box, where the user can identify and register themselves as a stroke survivor or caregiver of the stroke survivor. This helps the investigator or administrator to monitor the engagement and usage of the application by the stroke survivors and caregivers separately. On the basis of the options chosen, the user will be redirected to the specific registration page with drop-down options and text boxes to fill in the user details requested and register onto the application. After completing the registration, users will be redirected back to the sign-in page. Registration requires the users to have a username and password to ensure identity and privacy (figure 3).

egistration page with drop-down options and text boxes to fill in the user details requested and register onto the application. After completing the registration, users will be redirected back to the sign-in page. Registration requires the users to have a username and password to ensure identity and privacy (figure 3). Figure 3 Registration page. Intervention page After the user signs into the application successfully, the application is redirected to the main intervention page. This page contains brief written information about the intervention and five important sections that contain the content of the ‘care for stroke’ intervention (figure 4). Figure 4 Intervention page. Sections: There are five main sections displayed as photographic icons on the intervention page which can be touched and explored further (figure 4). These five sections contain digitised information (videos) about stroke and the various aspects that a stroke survivor can view and understand about the management of the physical disabilities following a stroke (figure 5). Figure 5 Section page.

Sections: There are five main sections displayed as photographic icons on the intervention page which can be touched and explored further (figure 4). These five sections contain digitised information (videos) about stroke and the various aspects that a stroke survivor can view and understand about the management of the physical disabilities following a stroke (figure 5). Figure 5 Section page. Subsections: When the user touches an icon on the section page, it is redirected to the corresponding subsection page that comprises topics (subsections) that the respective section contains. For example, the main intervention page will contain a photograph of the stroke survivor performing his Activities of Daily Living—ADL (intervention page); if the user touches this icon, it will take him or her to the ADL section (figure 5). If the user touches this ADL section icon, the web page will be redirected to the ADL subsection page that contains topics with video icons (images) related to ADL, in this instance, stroke survivors performing brushing, feeding, dressing, etc. Please find the section web page in figure 6. Figure 6 Subsection page.

Subsections: When the user touches an icon on the section page, it is redirected to the corresponding subsection page that comprises topics (subsections) that the respective section contains. For example, the main intervention page will contain a photograph of the stroke survivor performing his Activities of Daily Living—ADL (intervention page); if the user touches this icon, it will take him or her to the ADL section (figure 5). If the user touches this ADL section icon, the web page will be redirected to the ADL subsection page that contains topics with video icons (images) related to ADL, in this instance, stroke survivors performing brushing, feeding, dressing, etc. Please find the section web page in figure 6. Figure 6 Subsection page. Content digitised videos: When the user touches a topic in the subsection, the web page will be redirected to a page that contains detailed information about that topic in the form of 3–5 min video clips. For example, if the user touches the topic ‘Wearing a blouse’, the web page will be redirected to a video clip related to that topic. These videos are streamed online through internet or mobile internet networks and can be watched by touching the play button on the video clip. Please find the video section of the application in figure 7 below. Figure 7 Video section of the application. Shuffling between the web pages Users can shuffle between the pages by either Pressing the back button on the Smartphone Pressing the back icon on the web page Sliding the web pages back and forth using the touch screen option on the Smartphone.

Content digitised videos: When the user touches a topic in the subsection, the web page will be redirected to a page that contains detailed information about that topic in the form of 3–5 min video clips. For example, if the user touches the topic ‘Wearing a blouse’, the web page will be redirected to a video clip related to that topic. These videos are streamed online through internet or mobile internet networks and can be watched by touching the play button on the video clip. Please find the video section of the application in figure 7 below. Figure 7 Video section of the application. Shuffling between the web pages Users can shuffle between the pages by either Pressing the back button on the Smartphone Pressing the back icon on the web page Sliding the web pages back and forth using the touch screen option on the Smartphone. In addition to this, the user can return to the main intervention page at any time by touching the logo which is located on top of every web page of the application.

Shuffling between the web pages Users can shuffle between the pages by either Pressing the back button on the Smartphone Pressing the back icon on the web page Sliding the web pages back and forth using the touch screen option on the Smartphone. In addition to this, the user can return to the main intervention page at any time by touching the logo which is located on top of every web page of the application. Administrator module This Smartphone-enabled intervention is built with an administrator module, where the usage and utilisation patterns of this application by the users can be tracked continuously and reports can be generated to inform the feasibility of this intervention and also to monitor the progress of any programmes/research projects related to this intervention when scaled up to a larger community of stroke survivors. The administrator can also add videos onto (or remove videos from) the application as and when required, thereby customising or improvising the content of the intervention according to the needs of the users. The module is protected and strictly secured through a username and password to ensure privacy and confidentiality of the user information.

dd videos onto (or remove videos from) the application as and when required, thereby customising or improvising the content of the intervention according to the needs of the users. The module is protected and strictly secured through a username and password to ensure privacy and confidentiality of the user information. Discussion Stroke is one of the leading causes of death and disability worldwide.11 Globally, nearly six million people die from stroke each year, and much of this stroke burden is borne by LMICs.12 Though the primary focus of many LMICs, including India, is to prevent stroke by reducing the prevalence of its risk factors, similar attention should also be given to those who survive a stroke and are disabled post-stroke.13 14 Unlike HICs, organised multidisciplinary rehabilitation services for stroke survivors are not available in many LMICs.15–17 Given the context of many LMICs with a scarce rehabilitation workforce and resources for rehabilitation, it is critically important to develop innovative post-stroke rehabilitation interventions that could address the growing magnitude of post-stroke disability and meet the rising need for rehabilitation services in these countries.

ext of many LMICs with a scarce rehabilitation workforce and resources for rehabilitation, it is critically important to develop innovative post-stroke rehabilitation interventions that could address the growing magnitude of post-stroke disability and meet the rising need for rehabilitation services in these countries. The international telecommunication union estimated that six billion people were mobile phone users during 2011 globally, which is equivalent to 87% of the world's population. This report has also documented that India is one of the top markets for Smartphone sales globally.18 The management of chronic diseases using Smartphone technology has been described in a recent systematic review.19 This review identified 15 Smartphone applications for management of chronic conditions. Out of these 15 applications, there was only one application called Mayo clinic meditation that was similar to the ‘Careforstroke’ application. The Mayo clinic application helped patients practise meditation through a 15 min training video on meditation.

15 Smartphone applications for management of chronic conditions. Out of these 15 applications, there was only one application called Mayo clinic meditation that was similar to the ‘Careforstroke’ application. The Mayo clinic application helped patients practise meditation through a 15 min training video on meditation. Some of the Smartphone applications used in stroke rehabilitation in HICs include the Dr Droid application that helps therapists to administer and track upper limb exercises for stroke rehabilitation,20 the Think-FAST application that features stroke prevention information and a list of stroke unit locations in Australia21 and PTX, a physiotherapy exercise application for individuals with any kind of neurological conditions that includes a pictorial description of the exercises for stroke survivors.22 The National Institute of Clinical Excellence (NICE) guidelines for long-term stroke rehabilitation also recommend the use of Smartphones for communication problems in patients with stroke.23

ividuals with any kind of neurological conditions that includes a pictorial description of the exercises for stroke survivors.22 The National Institute of Clinical Excellence (NICE) guidelines for long-term stroke rehabilitation also recommend the use of Smartphones for communication problems in patients with stroke.23 A chronic condition like stroke requires uninterrupted therapeutic care and constant monitoring during the entire continuum of recovery.24 In the absence of any organised stroke care services and with the limited resources for rehabilitation, a Smartphone-enabled educational intervention for management of disabilities could be a strategy to meet the substantial rehabilitation needs of stroke survivors in India. The evidence concerning the use of Smartphones in chronic disease care in India is just emerging and the use of Smartphones in health interventions to combat diseases like diabetes, hypertension and cardiovascular diseases is progressively being investigated.25 Adoption of this strategy could possibly reduce the barriers to access and availability of stroke rehabilitation services. It could also aid in efficient and sustained monitoring of patient progress throughout the continuum of care.

abetes, hypertension and cardiovascular diseases is progressively being investigated.25 Adoption of this strategy could possibly reduce the barriers to access and availability of stroke rehabilitation services. It could also aid in efficient and sustained monitoring of patient progress throughout the continuum of care. ‘Care for stroke’ is a Smartphone-enabled educational intervention for management of physical disabilities following a stroke. The content of the intervention was developed systematically and primarily based on the needs of the stroke survivors and informed by existing global evidence. It includes inputs from highly qualified and experienced multidisciplinary stroke rehabilitation professionals in a digitised audio–visual format that is more entertaining to watch and learn compared to the other methods of patient education such as an educational workbook and group teaching or lectures. This intervention is culture-specific and language–specific, and therefore the users can easily understand and adapt the techniques to manage their post-stroke-related disabilities. Since the intervention is loaded onto a Smartphone, the user can access the intervention as and when they need. Unlike television and DVD players, Smartphones are portable and handheld and hence it might aid the user to access the intervention conveniently (without having to plug wires, operate a remote to watch videos or depend on electricity).

oaded onto a Smartphone, the user can access the intervention as and when they need. Unlike television and DVD players, Smartphones are portable and handheld and hence it might aid the user to access the intervention conveniently (without having to plug wires, operate a remote to watch videos or depend on electricity). This Smartphone-based, technology-driven intervention can be less demanding in terms of the physical abilities required by the users to learn, when compared with other kinds of educational interventions like attending group sessions, using a stroke workbook or watching a DVD educational material about stroke. The application for accessing the intervention is web-based, and hence the users can also access the content through their laptops, desktops and tablets if required. From the point of view of programme managers and evaluators, this kind of web-based educational intervention can continuously monitor the usage and utilisation pattern of the intervention by each user, and it can be helpful to generate reports to monitor the efficiency and effectiveness of this intervention while scaling up, without having to contact the users. Since the intervention is Smartphone-enabled and web-based, the user can contact the service provider directly by dialling the contact numbers on the Smartphone or by making a skype call using the mobile internet services.

ncy and effectiveness of this intervention while scaling up, without having to contact the users. Since the intervention is Smartphone-enabled and web-based, the user can contact the service provider directly by dialling the contact numbers on the Smartphone or by making a skype call using the mobile internet services. This Smartphone-enabled intervention might also motivate the caregivers and family members to comprehend the importance of stroke rehabilitation and support the stroke survivors in utilising the key aspects of the intervention in their everyday life. From a financial perspective, the cost of using this Smartphone-enabled intervention might be less costly compared to the other ways of accessing information about stroke and the ways to manage post-stroke physical disability from rehabilitation experts or hospitals.

aspects of the intervention in their everyday life. From a financial perspective, the cost of using this Smartphone-enabled intervention might be less costly compared to the other ways of accessing information about stroke and the ways to manage post-stroke physical disability from rehabilitation experts or hospitals. The ‘Care for stroke’ application is currently under pilot testing for its feasibility and acceptability with a small group of stroke survivors and their caregivers in Chennai, India. If this application is found feasible and acceptable, the investigators intend to look at the clinical and cost-effectiveness of this intervention. To date and to the best of our knowledge, there has not been a web-based, Smartphone-enabled educational application and intervention for stroke survivors with a primary focus on the rehabilitation aspect of the stroke. In a global context and from a public health perspective, ‘Care for stroke’ is one such kind of innovative intervention which could be tested not just for its feasibility and acceptability but also for its clinical and cost-effectiveness through rigorously designed, randomised clinical trials. It is very important to test this intervention in LMICs where the rehabilitation and information needs of the stroke survivors seem to be substantial and largely unmet.

ot just for its feasibility and acceptability but also for its clinical and cost-effectiveness through rigorously designed, randomised clinical trials. It is very important to test this intervention in LMICs where the rehabilitation and information needs of the stroke survivors seem to be substantial and largely unmet. The authors thank the student for undertaking the research study as a part of his doctoral study at the London School of Hygiene and Tropical Medicine. The authors thank the ethics committee of the London School of Hygiene and Tropical Medicine, PHFI-Indian Institute of Public Health—Hyderabad and The Voluntary Health Services Hospital for granting scientific and ethics approval to conduct this research study. The authors also thank the professional software and technical consultants at Suchirsoftech (India) who developed the web-based application. The authors thank the expert committee members for assisting in the development of the content for the intervention. The authors thank the media professionals at Selva photography (India) for digitising the content for the intervention. The authors also thank the models who acted for content digitisation. The authors also thank all the stroke survivors and their caregivers for their participation and inputs in developing this intervention.

thors thank the media professionals at Selva photography (India) for digitising the content for the intervention. The authors also thank the models who acted for content digitisation. The authors also thank all the stroke survivors and their caregivers for their participation and inputs in developing this intervention. Contributors: SK (Doctoral student at LSHTM) conceived the study, designed and developed the intervention and drafted the manuscript. GVSM and HK played a pivotal role in conceptualising the stages involved in the development of the intervention and reviewed the manuscript. SM provided substantial guidance for the technical development of the web-based application. SG provided guidance in the concept of the research study. Funding: This work was supported by a Wellcome Trust Capacity Strengthening Strategic Award to the Public 542 Health Foundation of India and a consortium of UK universities. The authors thank the Wellcome-trust and Public Health Foundation of India for funding the research study. Competing interests: None declared. Patient consent: Obtained. Ethics approval: London School of hygiene and Tropical Medicine, VHS Hospital and Public health Foundation of India. Provenance and peer review: Not commissioned; internally peer reviewed.

Background Stroke is a major global public health problem. According to the Global Burden of Disease (GBD) study in 2010, stroke is the second leading cause of death worldwide.1 A person experiences a stroke when a blood clot blocks a blood vessel in the brain or a vessel that supplies it, or when there is bleeding in the brain. The interruption of blood supply to the brain reduces the supply of oxygen and nutrients to it, causing injury and death of brain tissue.2 This brain damage may subsequently result in long-term disability or death of the affected individual.2 Stroke is associated with a wide variety of sensory-motor, cognitive-perceptual and behavioural impairments.3 The effects of stroke will depend on the site of the brain lesion and severity of brain damage.4 In addition to the primary impairments following a stroke, secondary complications of stroke can also hamper the recovery process.5 The prognosis in stroke depends on the degree of primary impairments and secondary complications.4

ects of stroke will depend on the site of the brain lesion and severity of brain damage.4 In addition to the primary impairments following a stroke, secondary complications of stroke can also hamper the recovery process.5 The prognosis in stroke depends on the degree of primary impairments and secondary complications.4 Disability is an umbrella term, covering impairments, activity limitations and participation restrictions.6 Impairment following stroke may present as physical, mental or cognitive. Stroke impairment might limit the ability of the stroke survivor to independently perform his or her daily living activities (eg, difficulties in walking or communicating).6 Consequently, it might also restrict effective participation of the stroke survivor in his/her family and social roles.7 Disability following stroke depends on the degree of impairment (physical, mental, cognitive) as well as the personal and contextual environment of the affected individual. Most often, stroke survivors become disabled with profound effect on their quality of life.8

e stroke survivor in his/her family and social roles.7 Disability following stroke depends on the degree of impairment (physical, mental, cognitive) as well as the personal and contextual environment of the affected individual. Most often, stroke survivors become disabled with profound effect on their quality of life.8 The impact of disability following a stroke also affects the family of the stroke survivor.9 Adapting to the new role of a carer and adjusting to the sudden impact of stroke can be highly stressful for family members.10 The demand on caregivers increases tremendously, especially if the stroke survivor experiences severe disability.11 The rehabilitation needs of stroke survivors and their family will vary extensively based on the degree of impairment and the context in which they experience a stroke (eg, accessibility to stroke services, family support, etc).7 9–12

ases tremendously, especially if the stroke survivor experiences severe disability.11 The rehabilitation needs of stroke survivors and their family will vary extensively based on the degree of impairment and the context in which they experience a stroke (eg, accessibility to stroke services, family support, etc).7 9–12 Rehabilitation needs of stroke survivors in India Evidence from a literature review suggests that India is experiencing a silent epidemic of stroke.13 Prevalence rate of stroke in India is estimated to range from 84 to 262/100 000 in rural areas and 334 to 424/100 000 in urban areas. The incidence rate is 119–145/100 000 based on recent population based studies.14 The incidence and prevalence of stroke was observed to be higher in India, compared to the incidence and prevalence of stroke in high-income countries (HICs).13 Unlike HICs, there is a dearth of information about the rehabilitation needs of people with disabilities following stroke in India.14 People with disabilities in general encounter tremendous environmental barriers in accessing rehabilitation services in India.15 Lack of policy initiatives for rehabilitation, inadequate rehabilitation resources and health professionals, lack of an accessible environment and stigma are some of the major barriers that persons with disability experience in India.16 Taking into account the disability after stroke and the existing environmental barriers to rehabilitation, the needs of stroke survivors in India are expected to be substantial and diverse.

s, lack of an accessible environment and stigma are some of the major barriers that persons with disability experience in India.16 Taking into account the disability after stroke and the existing environmental barriers to rehabilitation, the needs of stroke survivors in India are expected to be substantial and diverse. Rehabilitation services in India Rehabilitation services in India are usually hospital-based and driven predominately by physiotherapists. Therapy inputs from other health professionals, such as occupational therapists and speech therapists, are hardly available to patients with stroke.17 A recent study undertaken by the Public Health Foundation of India (PHFI) for the Ministry of Health and Family Welfare (MoHFW) has indicated a supply-demand gap of about 6 500 000 allied health professionals in India.18 Even the information needs of patients with stroke and their primary caregivers to self-manage their problems following stroke remain largely unfulfilled.7 Provision of rehabilitation services in India are usually limited to specialised hospitals located in urban areas, and many people, especially those who are poor or who live in rural areas cannot afford to pay for, or have limited access to, such services.18 When patients and caregivers travel long distances to obtain rehabilitation services, there is a huge financial implication and opportunity cost involved in accessing these services.7 Although the number of private rehabilitation facilities in India has increased, these are only accessed by a minuscule proportion of the country's vast population.19 Owing to these reasons, most people with disabilities following stroke do not have access to rehabilitation services in India.20 The existing barriers to rehabilitation suggest that the rehabilitation needs of the stroke survivors in India remain largely unmet.

a minuscule proportion of the country's vast population.19 Owing to these reasons, most people with disabilities following stroke do not have access to rehabilitation services in India.20 The existing barriers to rehabilitation suggest that the rehabilitation needs of the stroke survivors in India remain largely unmet. Given the context, it is imperative that stroke survivors and their caregivers are educated about stroke and the ways to manage post-stroke disability on their own. Educational intervention could assist stroke survivors and their families to access support services and to make informed decisions regarding their care.21 22 Educational interventions were found to improve patients’ and carers’ knowledge on the self-management of stroke.23

o manage post-stroke disability on their own. Educational intervention could assist stroke survivors and their families to access support services and to make informed decisions regarding their care.21 22 Educational interventions were found to improve patients’ and carers’ knowledge on the self-management of stroke.23 A chronic condition such as stroke requires uninterrupted therapeutic care and constant monitoring during the entire continuum of recovery.23 In the absence of any organised stroke care services and with the limited resources for rehabilitation, a Smartphone-enabled educational intervention for management of disability could be a strategy to meet the substantial rehabilitation needs of stroke survivors in India.24 The evidence concerning the use of Smartphones in chronic disease care in India is finally emerging and the use of Smartphones in interventions to combat diseases such as diabetes, hypertension and cardiovascular diseases, is progressively being investigated.25 In some HICs, Smartphones are used to create awareness about the warning signs of stroke26 and also to aid rehabilitation of language and communication impairments following stroke.27 Adoption of this strategy could possibly reduce the barriers to access and availability of stroke rehabilitation services. It could also aid in efficient and sustained monitoring of patient progress throughout the continuum of care. Thus, this study seeks to develop and evaluate a Smartphone-enabled carer-supported educational intervention for management of physical disabilities following stroke in India.

of stroke rehabilitation services. It could also aid in efficient and sustained monitoring of patient progress throughout the continuum of care. Thus, this study seeks to develop and evaluate a Smartphone-enabled carer-supported educational intervention for management of physical disabilities following stroke in India. Overall aim and strategy of the study The aim of this study is to develop and evaluate a Smartphone-enabled, carer-supported education programme for stroke survivors in India. The purpose of this formative research is to systematically (conforming to the MRC framework28) develop an educational intervention for management of post-stroke disability for stroke survivors in India, and evaluate the feasibility and acceptability of delivering the intervention using Smartphones and with caregiver support. This study will be conducted in three phases: (1) development of the intervention; (2) pre-testing of the intervention and stakeholder consultation; and (3) piloting of the intervention, and assessment of feasibility and acceptability. Processes and activities involved in each phase of the research study are explained using a flow chart in figure 1, and are described in detail below. Figure 1 Flow chart of the processes and activities of the research phases.

This study will be conducted in three phases: (1) development of the intervention; (2) pre-testing of the intervention and stakeholder consultation; and (3) piloting of the intervention, and assessment of feasibility and acceptability. Processes and activities involved in each phase of the research study are explained using a flow chart in figure 1, and are described in detail below. Figure 1 Flow chart of the processes and activities of the research phases. This study will apply mixed research methods in order to collect more comprehensive evidence and have a deeper understanding of the research problem. Mixed methods research encourages the use of multiple worldviews, and is a practical and natural approach to research pertaining to development of a complex intervention.29 It is premised on the idea that the use of quantitative and qualitative approaches in combination provides a better understanding of research problems than either approach alone.30

the use of multiple worldviews, and is a practical and natural approach to research pertaining to development of a complex intervention.29 It is premised on the idea that the use of quantitative and qualitative approaches in combination provides a better understanding of research problems than either approach alone.30 Proposed study design Methods: phase 1: Development of the intervention The overall objective of this phase is to develop a Smartphone-enabled, carer-supported education intervention for stroke survivors to manage their post-stroke disability. Research questions that will be answered in this phase along with the methods are detailed in table 1. At the end of phase I, a provisional Smartphone-enabled educational intervention for management of stroke-related disability will be developed for field testing and refinement. This provisional intervention will encompass the rehabilitation needs of patients with stroke and their caregivers identified in this formative phase. Table 1 Objectives, research questions and methods for phase I Objectives Questions Methods Explore the experience of the stroke survivors and their caregivers in accessing stroke rehabilitation services What impact does a stroke cause in the life of a stroke survivor, and his/her family and caregiver? How does an individual affected by stroke and his/her family organise themselves to manage the effects of stroke? What measures will stroke survivors and their caregivers take in order to manage disability following stroke?

Objectives Questions Methods Explore the experience of the stroke survivors and their caregivers in accessing stroke rehabilitation services What impact does a stroke cause in the life of a stroke survivor, and his/her family and caregiver? How does an individual affected by stroke and his/her family organise themselves to manage the effects of stroke? What measures will stroke survivors and their caregivers take in order to manage disability following stroke? What is the general understanding of the stroke survivor and his/her caregiver about stroke rehabilitation? What kind of stroke rehabilitation services are generally available to stroke survivors, from where and from whom? At what phase of recovery are these rehabilitation services available to stroke survivors? How do stroke survivors usually access stroke rehabilitation services? What is the general perception of stroke survivors and their caregivers on the quality of available rehabilitation services? What are the difficulties faced by stroke survivors in accessing rehabilitation services? What is the cost of obtaining stroke rehabilitation services and what is the impact of this increased expenditure on the family? Survey using structured questionnaire and in-depth interviews with stroke survivors and their caregivers Assess the information needs of patients with stroke and their carers What information do stroke survivors and the caregivers need to manage disability following a stroke? What kind of information is available to them, from where and from whom?

Survey using structured questionnaire and in-depth interviews with stroke survivors and their caregivers Assess the information needs of patients with stroke and their carers What information do stroke survivors and the caregivers need to manage disability following a stroke? What kind of information is available to them, from where and from whom? At what point of time or phase of recovery (acute, post-acute, rehab, long-term care), is this information available to them? What is the quality of available information (regarding correctness, relevance, reliability and understandability)? Survey using structured questionnaire and in-depth interviews with stroke survivors and their caregivers Assess the rehabilitation needs of patients with stroke What are the various kinds of disability experienced by stroke survivors following stroke? What are the various rehabilitation needs (physical, psychological, social) of stroke survivors and their caregivers? What kinds of rehabilitation services are required to address the needs of stroke survivors and their caregivers? What kind of rehabilitation services would enable stroke survivors to be functionally independent following stroke? Survey using structured questionnaire and in-depth interviews with stroke survivors and their caregivers Assess the barriers and facilitators for rehabilitation of stroke survivors What are the present policies and programmes for rehabilitation of persons with disabilities especially following stroke?

What kind of rehabilitation services would enable stroke survivors to be functionally independent following stroke? Survey using structured questionnaire and in-depth interviews with stroke survivors and their caregivers Assess the barriers and facilitators for rehabilitation of stroke survivors What are the present policies and programmes for rehabilitation of persons with disabilities especially following stroke? How are these rehabilitation policies and programmes implemented? Are there adequate resources (human resources, material and financial resources) for rehabilitation of stroke survivors? How are these rehabilitation services delivered to stroke survivors? What are the constraints in providing rehabilitation services to stroke survivors? What can be done to address these constraints? What are the facilitating factors for provision of rehabilitation services for stroke survivors? In-depth interviews with health professionals providing stroke rehabilitation services in hospitals Assess the experience and comfort of the stroke survivors and their caregivers in using Smartphone technology Has the stroke survivor and/or caregiver ever used a Smartphone? What was their experience in using Smartphone technology? How long and for what purpose were they using the Smartphone? What abilities are essential/required to use a Smartphone comfortably? Will a Smartphone-enabled carer-supported education programme be useful for stroke survivors? What are the advantages and disadvantages of such interventions?

What was their experience in using Smartphone technology? How long and for what purpose were they using the Smartphone? What abilities are essential/required to use a Smartphone comfortably? Will a Smartphone-enabled carer-supported education programme be useful for stroke survivors? What are the advantages and disadvantages of such interventions? In-depth interviews with stroke survivors and their caregivers Recommendations for action What are the potential rehabilitation strategies to facilitate functioning, participation and independent living among stroke survivors? What are the resources required to implement the rehabilitation strategy? Is the strategy feasible and sustainable? How can the proposed rehabilitation strategy be implemented and made sustainable? What could be the potential barriers/problems for implementation? What are the possible solutions to address the implementation barriers? In-depth interviews with health professionals providing stroke rehabilitation services in hospitals Detailed methods for phase 1 Study setting Participants will be selected from hospitals within Chennai that provide treatment and rehabilitation services for stroke survivors, and that are willing to recruit participants for this phase. Hospitals that could be potential recruitment sites for this phase will be identified and contacted, and permission will be obtained. Chennai, with a population of over 9 million, is the capital city of the Indian state of Tamil Nadu. It is the biggest industrial and commercial centre in South India, and a major cultural, economic and educational centre in the country.

for this phase will be identified and contacted, and permission will be obtained. Chennai, with a population of over 9 million, is the capital city of the Indian state of Tamil Nadu. It is the biggest industrial and commercial centre in South India, and a major cultural, economic and educational centre in the country. Participant inclusion criteria: Participants with a recently diagnosed stroke (within the previous 6 weeks) as defined by the WHO;31 Aged ≥18 years; Presenting with minor and moderate stroke (ie, scoring 1–15, according to the National Institute of Health (NIH) stroke scale);32–34 Discharged from the hospitals (recruitment sites); Residing at home with a primary caregiver. Exclusion criteria: Stroke survivors with severe communication problems identified using the NIH stroke scale;32–34 Stroke survivors who cannot provide consent autonomously; Those presenting with severe stroke (ie, scoring >15, according to the NIH stroke scale).32–34 Survey of study participants using structured questionnaires From the participants who meet the inclusion criteria, a purposive sample will be selected for the questionnaire survey. Stroke survivors: 50 participants admitted to hospital and then discharged within the previous 6 weeks. Primary caregivers of the stroke survivors: 50 participants.

Survey of study participants using structured questionnaires From the participants who meet the inclusion criteria, a purposive sample will be selected for the questionnaire survey. Stroke survivors: 50 participants admitted to hospital and then discharged within the previous 6 weeks. Primary caregivers of the stroke survivors: 50 participants. This phase will have a pragmatic approach to participant recruitment. The initial recruitment will include all eligible participants. In the later stage, recruitment will be more focused on the potential subgroups of participants stratified by their age, gender and severity, for gaining a better understanding of their specific experiences and rehabilitation needs. The purpose of this survey is to identify the various kinds of rehabilitation needs of stroke survivors, and the barriers and facilitators encountered by stroke survivors to access stroke rehabilitation services. Study participants—in-depth interviews From the participants selected for the survey, a subsample will be selected for in-depth interviews, including: 10–15 stroke survivors; 10–15 primary caregivers of stroke survivors.

The purpose of this survey is to identify the various kinds of rehabilitation needs of stroke survivors, and the barriers and facilitators encountered by stroke survivors to access stroke rehabilitation services. Study participants—in-depth interviews From the participants selected for the survey, a subsample will be selected for in-depth interviews, including: 10–15 stroke survivors; 10–15 primary caregivers of stroke survivors. The purpose of the in-depth interviews is to gain detailed understanding of the experiences of the stroke survivors in relation to accessing stroke rehabilitation services and their rehabilitation needs following stroke. Participants will be asked about their experiences of accessing stroke rehabilitation services, their rehabilitation needs, and about the barriers and facilitators to rehabilitation in various domains of their daily life, such as self-care, mobility and home-management; and leisure, social and vocational activities.

e. Participants will be asked about their experiences of accessing stroke rehabilitation services, their rehabilitation needs, and about the barriers and facilitators to rehabilitation in various domains of their daily life, such as self-care, mobility and home-management; and leisure, social and vocational activities. In addition, a purposefully selected sample of 8–10 health professionals from different rehabilitation disciplines (eg, rehabilitation medicine, neurology, physiotherapy, occupational therapy, speech therapy) will be selected and interviewed in depth. The purpose is to understand the perspective of the health professionals about provision of stroke rehabilitation services. This will include their understanding about the barriers and facilitators to accessing stroke rehabilitation services, their knowledge about the existing Smartphone-based health interventions, and their attitudes and opinions about the use of a Smartphone enabled, care-supported education programme for domiciliary stroke rehabilitation. The in-depth interview process will end when the collection of new qualitative information does not shed any further light on the issue under investigation (saturation point). If the interviews with the proposed number of participants do not reach a saturation point, additional interviews will be conducted until saturation.

n-depth interview process will end when the collection of new qualitative information does not shed any further light on the issue under investigation (saturation point). If the interviews with the proposed number of participants do not reach a saturation point, additional interviews will be conducted until saturation. An investigator will administer the questionnaire verbally and will be conducting the interviews in English or Tamil, whichever is suitable for the respondents. The interviews for stroke survivors and their primary caregivers will take place at their homes. For the health professionals, the in-depth interviews will take place at their respective hospitals. Interviews and discussions will be conducted in a secluded area so that participants’ privacy and confidentiality is assured. All the interviews will be tape recorded. Study tools Separate questionnaires and topic guides will be developed for stroke survivors, their primary caregivers and health professionals, and pilot-tested before starting the study. The tools will be revised accordingly after the pilot-testing. The questionnaire will predominantly include close-ended questions with scaled responses. The questionnaire will be developed based on the WHO-Disability Assessment Schedule (WHODAS),35 and also from tools used in previous studies.36 The in-depth interviews will have specific topic guides with open-ended questions and prompts.

tionnaire will predominantly include close-ended questions with scaled responses. The questionnaire will be developed based on the WHO-Disability Assessment Schedule (WHODAS),35 and also from tools used in previous studies.36 The in-depth interviews will have specific topic guides with open-ended questions and prompts. Informed written consent All eligible participants will be informed about the study, and written consent will be obtained from those who are willing to participate. Stroke survivors who are discharged from the hospital (within a 6-week window prior to the study) and their caregivers will be identified using the hospital discharge records and contacted over the phone. The purpose and processes of the study will be explained to the participants and consent will be obtained from potential participants in person. Analysis Quantitative analysis of questionnaire survey data Investigators will use STATA V.13.0 (StataCorp 2013. Stata Statistical Software: Release 13. College Station, Texas: StataCorp LP, USA) for analysis of data from the questionnaire schedule. Data will be double entered and compared, to detect and correct any errors that might have occurred during the data entry. The questionnaire schedule will have specific domains of interest. The descriptive frequencies, and 95% CI for each of these domains will be calculated.

sis of data from the questionnaire schedule. Data will be double entered and compared, to detect and correct any errors that might have occurred during the data entry. The questionnaire schedule will have specific domains of interest. The descriptive frequencies, and 95% CI for each of these domains will be calculated. Qualitative analysis of in-depth interviews Transcribing qualitative data Transcribing will be carried out to produce a written version of the interview. It is a full ‘script’ of the interview.30 Hand written interview notes will be documented in detail on the same day in order to avoid losing information. Tape-recorded information from the in-depth interviews of study participants will be fully transcribed verbatim within 3–5 days after the interviews. Consideration will be given to how certain things were communicated, and to the context, feelings and meanings, while transcribing. Punctuation marks and techniques such as underlining, marking with symbols, using upper case lettering, underlining and emboldening during the transcribing process, will be used during analysis. Each transcribed interview will be reviewed as soon as possible and before the next interview in order to incorporate any interesting findings into the next interview, and to explore them further. Investigators will use the framework approach to carry out the qualitative analysis. The transcribed data will be analysed using the following steps: Familiarisation with the data; Identifying a thematic framework; Indexing; Charting; Mapping and interpretation.

Qualitative analysis of in-depth interviews Transcribing qualitative data Transcribing will be carried out to produce a written version of the interview. It is a full ‘script’ of the interview.30 Hand written interview notes will be documented in detail on the same day in order to avoid losing information. Tape-recorded information from the in-depth interviews of study participants will be fully transcribed verbatim within 3–5 days after the interviews. Consideration will be given to how certain things were communicated, and to the context, feelings and meanings, while transcribing. Punctuation marks and techniques such as underlining, marking with symbols, using upper case lettering, underlining and emboldening during the transcribing process, will be used during analysis. Each transcribed interview will be reviewed as soon as possible and before the next interview in order to incorporate any interesting findings into the next interview, and to explore them further. Investigators will use the framework approach to carry out the qualitative analysis. The transcribed data will be analysed using the following steps: Familiarisation with the data; Identifying a thematic framework; Indexing; Charting; Mapping and interpretation. Results of the analysis will inform the design and development of the intervention. The overall framework of the intervention package will be finalised after the formative phase.

Qualitative analysis of in-depth interviews Transcribing qualitative data Transcribing will be carried out to produce a written version of the interview. It is a full ‘script’ of the interview.30 Hand written interview notes will be documented in detail on the same day in order to avoid losing information. Tape-recorded information from the in-depth interviews of study participants will be fully transcribed verbatim within 3–5 days after the interviews. Consideration will be given to how certain things were communicated, and to the context, feelings and meanings, while transcribing. Punctuation marks and techniques such as underlining, marking with symbols, using upper case lettering, underlining and emboldening during the transcribing process, will be used during analysis. Each transcribed interview will be reviewed as soon as possible and before the next interview in order to incorporate any interesting findings into the next interview, and to explore them further. Investigators will use the framework approach to carry out the qualitative analysis. The transcribed data will be analysed using the following steps: Familiarisation with the data; Identifying a thematic framework; Indexing; Charting; Mapping and interpretation. Results of the analysis will inform the design and development of the intervention. The overall framework of the intervention package will be finalised after the formative phase. Triangulation of information—stakeholder participatory workshops Findings from the systematic reviews and the formative work (phase I) will be shared during a participatory consultation workshop with 8–10 key stakeholders (stroke survivors, primary caregivers, health professionals, and disability and rehabilitation experts) who will be selected for the study. Such workshops will be conducted at the end of each phase of the research (3-workshops in total). The purpose of the workshop is to facilitate triangulation of the information obtained from each phase and to reach a decision on the best content for the intervention. The participatory consultation workshops will be organised to bring in the key stakeholders together to seek their opinions, extract their knowledge and to decide on the best content for the intervention in a collaborative and creative environment.

phase and to reach a decision on the best content for the intervention. The participatory consultation workshops will be organised to bring in the key stakeholders together to seek their opinions, extract their knowledge and to decide on the best content for the intervention in a collaborative and creative environment. Integrating the content of the education intervention with the Smartphone The educational intervention developed during the formative phase of the research will be transformed into a Smartphone-enabled intervention. This process will enable stroke survivors and their caregivers to use a Smartphone to access the intervention. The steps involved in developing the Smartphone enabled education programme are as follows The content of the stroke education intervention (eg, positioning techniques, pressure relief procedures, self-care tasks, functional ambulation and exercises) will be converted into an animated or an illustrated (using patient demonstration) video version. The digitised animated/video version of the stroke education intervention will be uploaded onto a Smartphone using an appropriate (Android/Windows) application platform. This will enable the participants to access the educational intervention package using the Smartphone. If there are any operation problems in uploading or technical issues with the Smartphone application, the digitised video clips will be transferred onto a specific folder that contains videos in the Smartphone.

tform. This will enable the participants to access the educational intervention package using the Smartphone. If there are any operation problems in uploading or technical issues with the Smartphone application, the digitised video clips will be transferred onto a specific folder that contains videos in the Smartphone. Once the stroke education intervention is uploaded onto the Smartphone, the Smartphone-enabled educational intervention will be ready for use by the participants. Phase II methods Phase II: Pre-testing of the intervention and stakeholder consultation Field testing of the intervention The provisional Smartphone-enabled intervention package will be field tested with a subsample of 30 adult stroke survivors and their caregivers. For this, a subsample of stroke survivors and their caregivers will be purposively identified from phase 1 survey respondents, excluding those who were part of the in-depth interview process. The Smartphone loaded with the intervention will be provided to the participants to be used at home for 2 weeks. Primary caregivers of stroke survivors selected for this phase will be asked to support the stroke survivors in accessing the intervention from the Smartphone.

ho were part of the in-depth interview process. The Smartphone loaded with the intervention will be provided to the participants to be used at home for 2 weeks. Primary caregivers of stroke survivors selected for this phase will be asked to support the stroke survivors in accessing the intervention from the Smartphone. Direct observation during field-testing Utilisation of the Smartphone-enabled intervention and the support provided by the caregivers to the stroke survivors will be assessed by an Occupational Therapist (SK) using direct observation techniques during this phase. The main purpose of using a direct observation technique in this phase is to triangulate and affirm the information provided by the participants during phase 1. Some of the key issues that will be assessed during the direct observation include: Relevance and comprehensibility of the intervention; Operational difficulties of the participants in using the Smartphone; User-friendliness of the intervention; Technical issues in the Smartphone; Training needs in order to access the intervention from the Smartphone. An observation checklist will be developed and used to assess these key issues during field-testing. The outcome of the field-testing will inform the development of a finalised version of a completely illustrated, pictorial training manual (user-friendly even for participants with low literacy level) explaining the operation of the Smartphone to access the intervention.

to assess these key issues during field-testing. The outcome of the field-testing will inform the development of a finalised version of a completely illustrated, pictorial training manual (user-friendly even for participants with low literacy level) explaining the operation of the Smartphone to access the intervention. Stakeholder consultation: refinement of the educational intervention The outcomes of the field testing phase will be shared with the key stakeholders for their feedback and recommendations during the second participatory consultation workshop. The objective of this workshop is to consult with stakeholders about the feasibility of the intervention, receive feedback and refine the intervention, as recommended by the stakeholders. The consultation workshop will be a participatory process as described above. Recommendations from the stakeholder consultations will be used to refine the intervention package for the pilot phase. Phase III: Piloting of the intervention and assessment of feasibility and acceptability The objective of this pilot phase is to implement the intervention, and evaluate the feasibility and acceptability of the intervention. This phase will be carried out as a pilot study, which will provide useful information to plan a large scale RCT of the intervention in the future. Phase III: Methods Participants and eligibility criteria Participants with a primary diagnosis of stroke will be recruited from VHS hospital in Chennai, India. The inclusion criteria will be: Adults ≥18 years;

Phase III: Piloting of the intervention and assessment of feasibility and acceptability The objective of this pilot phase is to implement the intervention, and evaluate the feasibility and acceptability of the intervention. This phase will be carried out as a pilot study, which will provide useful information to plan a large scale RCT of the intervention in the future. Phase III: Methods Participants and eligibility criteria Participants with a primary diagnosis of stroke will be recruited from VHS hospital in Chennai, India. The inclusion criteria will be: Adults ≥18 years; Recent diagnosis of first ever stroke—as defined by the WHO31 within 3–6 weeks prior to the recruitment; Severity of stroke, mild and moderate (score 1–15, according to the NIH stroke scale);32–34 Stroke survivor medically stable (reaching a point in medical treatment where life-threatening problems following stroke have been brought under control); Post-stroke functional status of the stroke survivor: requiring assistance of one person to perform basic activities of daily living including transfers, self-care and mobility; Stroke survivor residing with a primary caregiver (family member) at home. The exclusion criteria will be Participants with NIH score >15; Severe cognitive difficulties (NIH stroke scale components for cognition);34 Severe communication problem; Severe comorbidities (severe psychiatric illness, hearing loss, vision loss); Stroke survivor functionally dependent due to pre-existing conditions; Stroke survivor who does not have a primary caregiver; Stroke survivors who are unwilling/unable to adhere to the study protocol;

Severe cognitive difficulties (NIH stroke scale components for cognition);34 Severe communication problem; Severe comorbidities (severe psychiatric illness, hearing loss, vision loss); Stroke survivor functionally dependent due to pre-existing conditions; Stroke survivor who does not have a primary caregiver; Stroke survivors who are unwilling/unable to adhere to the study protocol; Participants who do not qualify the training requirements (operation of Smartphone). Eligibility assessment using NIH stroke scale will be conducted by the investigator to identify participants to be recruited for this pilot study. Participant recruitment for the pilot study Participants for the pilot study will be recruited after their hospital discharge. Information about participants who are discharged from the hospital (in the past 3–6 weeks) will be retrieved from hospital records. An eligibility assessment will be completed within 2 weeks after identification of the participant from the hospital records. Participants identified for the piloting will be identified and contacted by phone. They will be informed about the purpose and processes of the study. If a participant is interested, written informed consent from the participant will be obtained in person. Consent procedures will be completed at the participant's home.

articipants identified for the piloting will be identified and contacted by phone. They will be informed about the purpose and processes of the study. If a participant is interested, written informed consent from the participant will be obtained in person. Consent procedures will be completed at the participant's home. Participants for the pilot study A total of 30–40 participants will be recruited from VHS hospital in Chennai for the pilot phase. The admission rate of stroke survivors in this hospital is 3–4/weeks. Given the hospital admission rate, it will take 4–5 months to recruit 30–40 participants who will meet the eligibility criteria for this phase. Intervention procedure The Smartphone uploaded with the intervention will be provided to the participants and we will show the stroke survivors and their caregivers how to use the Smartphone-enabled intervention. If the stroke survivor requires assistance, their caregivers will be encouraged to support them in using the intervention. The participants will be introduced to the intervention during initial home-visit. A structured training session for the stroke survivors and their caregivers on using the Smartphone-enabled intervention will be provided. The structured training will include: Introduction to the Smartphone-enabled intervention. Accessing the educational intervention package using the smart-phone application.

Intervention procedure The Smartphone uploaded with the intervention will be provided to the participants and we will show the stroke survivors and their caregivers how to use the Smartphone-enabled intervention. If the stroke survivor requires assistance, their caregivers will be encouraged to support them in using the intervention. The participants will be introduced to the intervention during initial home-visit. A structured training session for the stroke survivors and their caregivers on using the Smartphone-enabled intervention will be provided. The structured training will include: Introduction to the Smartphone-enabled intervention. Accessing the educational intervention package using the smart-phone application. The finalised manual for Smartphone operation to access the intervention will be used during this training. A copy of the finalised Smartphone operation manual will be provided to the study participants for use at home. An occupational therapist (SK) will also assess whether the participants are able to use the Smartphone application (hands-on) appropriately during the training. An errorless attempt to retrieve the required part of the intervention from the Smartphone for more than three attempts will be considered to be successful training. After successful training, the Smartphone enabled educational intervention package will be provided to the participants for use at their home for the next 4 weeks. Participants will also be encouraged to contact the study leader (SK) if they have any concerns regarding Smartphone operation during these 4 weeks.

The finalised manual for Smartphone operation to access the intervention will be used during this training. A copy of the finalised Smartphone operation manual will be provided to the study participants for use at home. An occupational therapist (SK) will also assess whether the participants are able to use the Smartphone application (hands-on) appropriately during the training. An errorless attempt to retrieve the required part of the intervention from the Smartphone for more than three attempts will be considered to be successful training. After successful training, the Smartphone enabled educational intervention package will be provided to the participants for use at their home for the next 4 weeks. Participants will also be encouraged to contact the study leader (SK) if they have any concerns regarding Smartphone operation during these 4 weeks. Assessment of outcomes The primary outcomes of the pilot phase will be the feasibility and acceptability of the intervention. In addition to these outcomes, assessment of outcomes relating to the extent of disability and independence in activities of daily living will be carried out using the Modified Rankin scale37 and Barthel Index,38 respectively. Assessment of these outcomes will inform the feasibility of using these outcome assessment tools for future trials of the intervention. Details of the outcome assessment of phase 3 are explained in table 2. Table 2 Details of the outcome assessment for the pilot testing phase

Assessment of outcomes The primary outcomes of the pilot phase will be the feasibility and acceptability of the intervention. In addition to these outcomes, assessment of outcomes relating to the extent of disability and independence in activities of daily living will be carried out using the Modified Rankin scale37 and Barthel Index,38 respectively. Assessment of these outcomes will inform the feasibility of using these outcome assessment tools for future trials of the intervention. Details of the outcome assessment of phase 3 are explained in table 2. Table 2 Details of the outcome assessment for the pilot testing phase Outcomes Description Feasibility A list of indicators will be developed during the pilot phase of the research study to assess feasibility of the intervention. This will include Feasibility of recruitment: Time taken to recruit the proposed number of participants Proportion of eligible participants identified Proportion of participants who consented in relation to participants who are eligible Reasons for exclusion Training: Number of participants successfully trained from the number of participants recruited for training Time taken for training by participants in different age-group, stroke severity and other factors (eg, experience of using a Smartphone) Training needs of participants in different age-groups, stroke severity and other factors Study processes: An in-built mechanism will be configured onto the Smartphone application to monitor the use of the intervention by participants. These indicators include: Proportion of participants ever using the application

Time taken for training by participants in different age-group, stroke severity and other factors (eg, experience of using a Smartphone) Training needs of participants in different age-groups, stroke severity and other factors Study processes: An in-built mechanism will be configured onto the Smartphone application to monitor the use of the intervention by participants. These indicators include: Proportion of participants ever using the application Proportion of participants using the application every week Proportion of participants using the application every day Proportion of participants using it for more than 1 h Proportion of participants requiring carer support Proportion of participants and carers successfully trained in using the application Proportion of participants accessing specific contents from the intervention Proportion of participants contacting the trainer/investigator for support Proportion of participants adhering to study protocol Reasons for non-adherence Follow-up: Number of drop-outs Reasons for dropping out Acceptability During the follow-up at the end of 4 weeks, a patient experience assessment will be conducted in order to understand the reasons for adherence or non-adherence, using a small questionnaire that will be developed for this purpose, soon after the end of phase 2 (once the intervention is ready) Functional outcomes Extent of disability—Modified Rankin Scale Activities of daily living—Barthel Index

Acceptability During the follow-up at the end of 4 weeks, a patient experience assessment will be conducted in order to understand the reasons for adherence or non-adherence, using a small questionnaire that will be developed for this purpose, soon after the end of phase 2 (once the intervention is ready) Functional outcomes Extent of disability—Modified Rankin Scale Activities of daily living—Barthel Index Analysis plan for the pilot phase STATA will be used for analysis of the data in the pilot phase. Outcomes measuring the difference in proportions will be analysed using the χ2 test or Fisher's exact test. Outcomes measuring the difference in means will be analysed using the paired student t test or Wilcoxon matched pairs signed rank-sum test. Multivariate analysis using logistic regression techniques will also be conducted. To adjust for the imbalances in baseline characteristics, stratified analysis will be conducted or Mantel-Haenzel method will be used in the analysis. Stakeholder workshop The study findings will be shared during the final stakeholder workshop at the end of the pilot testing (phase 3).

Analysis plan for the pilot phase STATA will be used for analysis of the data in the pilot phase. Outcomes measuring the difference in proportions will be analysed using the χ2 test or Fisher's exact test. Outcomes measuring the difference in means will be analysed using the paired student t test or Wilcoxon matched pairs signed rank-sum test. Multivariate analysis using logistic regression techniques will also be conducted. To adjust for the imbalances in baseline characteristics, stratified analysis will be conducted or Mantel-Haenzel method will be used in the analysis. Stakeholder workshop The study findings will be shared during the final stakeholder workshop at the end of the pilot testing (phase 3). Expected outcomes of this research study This study seeks to develop and evaluate a Smartphone-enabled carer-supported educational intervention for management of post-stroke disability in India. Empirical exploration of this strategy will provide information on pragmatic solutions required to address the growing needs due to stroke disability in India and in other resource constrained settings. This research will provide an opportunity to develop a patient-centred educational intervention for management of post-stroke disability that is relevant to the context of low- and middle-income countries. Findings from the research will also provide valuable information about the resources required to deliver such interventions in resource-constrained settings.

develop a patient-centred educational intervention for management of post-stroke disability that is relevant to the context of low- and middle-income countries. Findings from the research will also provide valuable information about the resources required to deliver such interventions in resource-constrained settings. This work was supported by a Wellcome Trust Capacity Strengthening Strategic Award to the Public Health Foundation of India and a consortium of UK universities. The authors thank Wellcome Trust and Public Health Foundation of India for funding the research study and also the student who undertook the research study as a part of his doctoral study at the London School of Hygiene and Tropical Medicine. The authors thank the ethics committee of the London School of Hygiene and Tropical Medicine, PHFI-Indian Institute of Public Health—Hyderabad, and The Voluntary Health Services Hospital for granting scientific and ethics approval to conduct this research study. The authors thank Professor John Porter and Dr Solomon Salve and Professor Pat Doyle from LSHTM, Dr Audrey Prost from UCL, Dr Henry Prakash from Christian Medical College, Vellore, and Dr BR Shamanna from School of Medical Sciences, University of Hyderabad, for their valuable inputs in refining the research study design and methods.

ank Professor John Porter and Dr Solomon Salve and Professor Pat Doyle from LSHTM, Dr Audrey Prost from UCL, Dr Henry Prakash from Christian Medical College, Vellore, and Dr BR Shamanna from School of Medical Sciences, University of Hyderabad, for their valuable inputs in refining the research study design and methods. Contributors: SK conceived and designed the study, and drafted the manuscript. GVSM and HK played a crucial role in conception of the research study and provided substantial guidance in designing the research study. SK and SG provided scientific and methodological advice, related to conception and designing of the research study. Funding: Wellcome Trust—Public Health Foundation of India Capacity strengthening grant. Competing interests: None declared. Ethics approval: London School of Hygiene and Tropical Medicine, Public Health Foundation of India and Hospital Ethics Committee (VHS Hospital). Provenance and peer review: Not commissioned; internally peer reviewed. Data sharing statement: This is only a protocol of a formative research study and we are in the process of data collection at present.

Introduction A wide range of open-source hardware that includes healthcare sensors and low-cost single-board computers are readily available on the consumer market. The Arduino, arguably one of the most widespread and successful open-source hardware platforms, contains a microcontroller with an easy-to-use development environment that makes it ideal to build devices. It has a large community of supporters, and there are open software libraries available to simplify many tasks. The radical shift in approach is that these new designs are open. In principle, anybody can build an Arduino, because its full design specification is available for anybody to use. This openness has stimulated many people to build a huge range of sensors and other devices that are compatible with it. Ironically, being open rather than protecting their designs (eg, with patents) has vastly increased their sales and lowered prices through stimulating a worldwide marketplace.

ble for anybody to use. This openness has stimulated many people to build a huge range of sensors and other devices that are compatible with it. Ironically, being open rather than protecting their designs (eg, with patents) has vastly increased their sales and lowered prices through stimulating a worldwide marketplace. Why is open-source hardware a good idea for medical devices? Making the hardware design available under an open-source license allows anyone to improve and contribute to the device design, leading to very rapid innovation compared to traditional methods.1 2 It also enables the design to be modified for very specific uses, and makes the devices easy to repair, factors which also reduce the impact these devices have on the environment. An open-source medical device increases safety, security and robustness by allowing more people to inspect and improve its designs.3 In fact, it is so difficult to develop good computer security that the best encryption algorithms are always developed in the open to allow others to inspect and improve them.4 Furthermore, if the software to connect devices is open-source and the physical interfaces standardised, vendor lock-in can be prevented. Open-source also allows many manufacturers to use the same design but differentiate in other ways in order to compete, for example on usability, support or wider interoperability with other devices. The same approach can improve medical devices.

and the physical interfaces standardised, vendor lock-in can be prevented. Open-source also allows many manufacturers to use the same design but differentiate in other ways in order to compete, for example on usability, support or wider interoperability with other devices. The same approach can improve medical devices. Potential of open-source medical devices for the developing world According to the WHO, 70–90% of all medical devices donated to the developing world never function as intended.5 6 Very simple faults, like a broken fuse or dead batteries, account for 15% of these failures. Twenty per cent of all donated equipment are not used because there are no manuals available or because of poor user training. Even when training is provided, it is rare for technical staff to be provided with technical training. How would an open-source medical device like a syringe pump solve some of these problems? A new syringe pump currently on the market can cost between $500 and $10k, depending on the number of features and its application. Open-source syringe pumps can be built for 5–10% of the cost of a pump of similar performance, which has great potential for making medical devices more accessible in the developing world,7 where devices can also be designed as open-source and built for specific use cases, instead of having to depend on donated equipment from first world countries. Arguably, under-resourced ‘developed’ healthcare has exactly the same problems though on a different scale: open-source should therefore have similar benefits in the developed healthcare systems.

How would an open-source medical device like a syringe pump solve some of these problems? A new syringe pump currently on the market can cost between $500 and $10k, depending on the number of features and its application. Open-source syringe pumps can be built for 5–10% of the cost of a pump of similar performance, which has great potential for making medical devices more accessible in the developing world,7 where devices can also be designed as open-source and built for specific use cases, instead of having to depend on donated equipment from first world countries. Arguably, under-resourced ‘developed’ healthcare has exactly the same problems though on a different scale: open-source should therefore have similar benefits in the developed healthcare systems. Quantifying the value of open-source hardware development Pearce1 developed a set of formulas to quantify the value of open-source hardware design. One way is to compare distributed manufacturing to traditional manufacturing, based on the number of downloads of a design that results in a manufactured product. Another way is to calculate the costs saved by not having to replicate a product design, based on the number of design hours and hourly wage of the designers. A third approach is to calculate the market size of a distributed manufacturing approach based on the number of products and their manufacturing cost.

d product. Another way is to calculate the costs saved by not having to replicate a product design, based on the number of design hours and hourly wage of the designers. A third approach is to calculate the market size of a distributed manufacturing approach based on the number of products and their manufacturing cost. Using the first approach above, the value of a basic syringe pump design by Wijnen et al8 was shown to be between $778 000 and $12.4 million over 1 year. This is called the downloaded substitution valuation, which calculates the annual savings by comparing the cost of purchasing a traditionally manufactured product to the marginal cost of producing the open-source hardware version using distributed manufacturing. The value of the design is further quantified using the number of times that the open-source hardware design is downloaded. Regulatory issues Clinical trials for medical devices may take anything from 2 to 5 years. While 5 years is a long time for a small business, performing the design and development, and the clinical trials, as part of an open-source project means that a small business can share the load. In principle the business would need to be in operation for a much shorter time before it could successfully launch a new product.

years is a long time for a small business, performing the design and development, and the clinical trials, as part of an open-source project means that a small business can share the load. In principle the business would need to be in operation for a much shorter time before it could successfully launch a new product. Methods We selected 10 open-source hardware medical devices and compared them in terms of how easy it is to obtain the components and build the device independently. A systematic review is not possible at this stage, as many of the projects are still in early phases and are not yet reported in peer-reviewed literature. As this is a rapidly developing area, then, we do not claim our list of projects is exhaustive or fully representative. We defined the scope of the review based on the following inclusion criteria: (1) availability of source files, and (2) can potentially be used as an active medical device, that is, it relies on software and a source of electrical energy for functioning. All the design files, for example, two-dimensional drawings, CAD (computer-aided design) files, circuit schematics and layouts, need to be available and modifiable. If the files are only available in a non-editable format, the project was excluded. For example, the Robohand9 three-dimensional (3D)-printed prosthetics project does not share source files, only STL (STereoLithography) files, and then only in a non-editable format. We distinguish between: Passive medical devices, like hose clamps Medical devices relying on software, like most syringe drivers

All the design files, for example, two-dimensional drawings, CAD (computer-aided design) files, circuit schematics and layouts, need to be available and modifiable. If the files are only available in a non-editable format, the project was excluded. For example, the Robohand9 three-dimensional (3D)-printed prosthetics project does not share source files, only STL (STereoLithography) files, and then only in a non-editable format. We distinguish between: Passive medical devices, like hose clamps Medical devices relying on software, like most syringe drivers Devices for laboratory research, like centrifuges Project comparison Myoelectric prostheses Myoelectric prostheses are controlled by tiny voltages generated on the surface of the skin by the activity of residual muscles, called surface electromyography (EMG). These voltages are amplified and processed so that when the user flexes a muscle, motors on the prosthesis move in a predictable way. A myoelectric prosthesis does away with the harness and cables of a body-powered device and can be made to look very natural, although at the cost of a more restrictive and less comfortable socket. Another downside to these devices is their lack of responsiveness: most prosthetic hands move slowly to increase grip force or battery life at the expense of fast action. However, the major barrier to adoption is cost: an arm costs at least $30 000.

lthough at the cost of a more restrictive and less comfortable socket. Another downside to these devices is their lack of responsiveness: most prosthetic hands move slowly to increase grip force or battery life at the expense of fast action. However, the major barrier to adoption is cost: an arm costs at least $30 000. MyOpen MyOpen,10 part of the Open Prosthetics Project,11 creates open-source hardware for myoelectric control. A circuit board design for a digital signal processing unit and related software are available.12 The hope is that it can eventually be used to control mechatronic prostheses for amputees.13 MyOpen wants to bring the cost of a prosthetic arm as low as $250. Silva et al14 have made and tested a self-contained mechanomyogram controlled prosthesis. Their device contains three sensors, each consisting of a microphone to pick up muscle sounds and an accelerometer to detect external interference. Their signals are interpreted by a microcontroller and converted into simulated EMG signals to control an Otto Bock hand.15 Their results show that the devices perform with over 70% control accuracy.

sensors, each consisting of a microphone to pick up muscle sounds and an accelerometer to detect external interference. Their signals are interpreted by a microcontroller and converted into simulated EMG signals to control an Otto Bock hand.15 Their results show that the devices perform with over 70% control accuracy. In an attempt to replicate the research conducted by Silva et al, the Open Prosthetics project11 built a data collection system for recording MMG signals. It consists of a 3.5 mm audio plug, a microphone cartridge, some wire and a laptop. Not including the laptop, the components cost less than $5. The microphone is an electret condenser type, and is powered by the laptop's sound card. They have used the free and open-source software (FOSS) Audacity to record and process the sound. They have taped the microphone directly onto the skin with masking tape, plugged it into the laptop, and recorded the audio. No special skills or tools are required to build this system. e-Nable Another effort to develop a myoelectric prosthetic is the Limbitless Arm,16 created by a team at University of Central Florida. This prosthetic is also part of e-Nable,17 an online community that connects amputees with people who have 3D printers. It was estimated to cost approximately $350 in materials to produce a Limbitless Arm.

t to develop a myoelectric prosthetic is the Limbitless Arm,16 created by a team at University of Central Florida. This prosthetic is also part of e-Nable,17 an online community that connects amputees with people who have 3D printers. It was estimated to cost approximately $350 in materials to produce a Limbitless Arm. To produce a 3D printed mechanical hand, 3DUniverse18 is selling a ‘Hand Materials Kit’ for $25. This kit is for assembling the printed hand only and is also developed by the e-Nable community. The models are available to download freely and print.17 However, the users still need to access a 3D printer to print the parts. Note that inexpensive 3D printers are available in developing world contexts with some open-source printers even designed to be mobile.19 CT scanner Jansen20 built an open-source desktop CT scanner (see figure 1) for small objects. The complete design files are available online,21 including the mechanical design, circuit schematics and firmware. The CT scanner uses a very low intensity radioisotope X-ray source, which means that even low resolution images take hours of measurement.22 It is designed for academic and educational purposes, but the hope is that it could eventually be used as a medical scanner in developing countries. Figure 1 Open-source CT scanner (Photo by Peter Jansen, image licensed under the Creative Commons BY-SA license).

CT scanner Jansen20 built an open-source desktop CT scanner (see figure 1) for small objects. The complete design files are available online,21 including the mechanical design, circuit schematics and firmware. The CT scanner uses a very low intensity radioisotope X-ray source, which means that even low resolution images take hours of measurement.22 It is designed for academic and educational purposes, but the hope is that it could eventually be used as a medical scanner in developing countries. Figure 1 Open-source CT scanner (Photo by Peter Jansen, image licensed under the Creative Commons BY-SA license). Jansen designed the almost entirely laser-cuttable CT scanner with four axes of motion, one being a large rotary gantry. The stepper motor to rotate the gantry is a National Electrical Manufacturers Association (NEMA)17 stepper from open-source hardware distributor Adafruit ($14), which transfers motion to the drive shaft using a belt and timing pulleys. Each linear axis has a small carriage that contains mounts for either the source or detector. Each axis has an inexpensive NEMA14 stepper also from Adafruit (under $14) and an idler pulley.23 Jansen has used a very small solid state high-energy particle detector called the type-5 from radiation watch, which can be connected to an external microcontroller. The price of this particle detector is just under $80. To test out the motion and detector, Jansen put together an Arduino shield ($45) with three Pololu stepper controllers24 and a connector for the detector. An SD card slot can store the image data for large scans.

Jansen designed the almost entirely laser-cuttable CT scanner with four axes of motion, one being a large rotary gantry. The stepper motor to rotate the gantry is a National Electrical Manufacturers Association (NEMA)17 stepper from open-source hardware distributor Adafruit ($14), which transfers motion to the drive shaft using a belt and timing pulleys. Each linear axis has a small carriage that contains mounts for either the source or detector. Each axis has an inexpensive NEMA14 stepper also from Adafruit (under $14) and an idler pulley.23 Jansen has used a very small solid state high-energy particle detector called the type-5 from radiation watch, which can be connected to an external microcontroller. The price of this particle detector is just under $80. To test out the motion and detector, Jansen put together an Arduino shield ($45) with three Pololu stepper controllers24 and a connector for the detector. An SD card slot can store the image data for large scans. The source of radiation used in Jansen's design is Barium-133 (cost between $80 and $125.00) with 80–383 KeV energy. This radioisotope check source is sealed in epoxy, and is of such low intensity that it not licensed25 and considered safe unless the material is digested or taped to the body for long periods, and it can be disposed of as general trash. The total cost of Jansen's open-source CT scanner (excluding the laser cutter) is about $300. Open-source laser cutters can be built for around $1000. All required components are readily available and only basic knowledge of electronics and laser cutting is required.

periods, and it can be disposed of as general trash. The total cost of Jansen's open-source CT scanner (excluding the laser cutter) is about $300. Open-source laser cutters can be built for around $1000. All required components are readily available and only basic knowledge of electronics and laser cutting is required. Infusion pumps Infusion pumps dispense fluid over a set amount of time. In hospitals they are used to deliver medication for treatments like chemotherapy and pain management, while laboratories use them for everything from microfluidics to bioprinting. Fechko's peristaltic pump Fechko26 built a peristaltic pump (see figure 2) using a Raspberry Pi single-board computer and a standardised T-slot profile construction set called OpenBeam. The pump was designed as a fluid reward system for laboratory animals, but can be adapted for other uses as well. Figure 2 Raspberry Pi Peristaltic Pump (Photo by Amber Fechko, image licensed under the Creative Commons BY-SA license). The electronics consist of modular components that do not require soldering. The electronics used are an Arduino or Raspberry Pi and components easily obtained from open-source hardware distributors. Source code is available for Arduino, using a motor shield, and Raspberry Pi, using either a Gertboard or L293D motor control chip. All components used are available off-the-shelf and no special skills are required to build the pump.

or Raspberry Pi and components easily obtained from open-source hardware distributors. Source code is available for Arduino, using a motor shield, and Raspberry Pi, using either a Gertboard or L293D motor control chip. All components used are available off-the-shelf and no special skills are required to build the pump. Wijnen et al's syringe pump Wijnen et al8 published a complete open-source syringe pump library, consisting of a series of 3D-printable parts to build a syringe pump. Performance of the syringe pump was assessed and the methods used for assessment are detailed. The design, bill of materials and assembly instructions are publicly available. The 3D-printable parts can be modified using an open-source parametric 3D design tool, called OpenSCAD27 and derivatives like OpenPump28 29 have already been created. To build the device a 3D printer, or 3D printing service, can be used to print the necessary parts. The electronics are based on a Raspberry Pi, stepper motor and driver components that can be obtained from open-source hardware distributors. Most mechanical components can be obtained from a local hardware store. Some specialist components like ball bearings, linear bearings and flexible couplings can be sourced online. Source code is available for the Raspberry Pi.

per motor and driver components that can be obtained from open-source hardware distributors. Most mechanical components can be obtained from a local hardware store. Some specialist components like ball bearings, linear bearings and flexible couplings can be sourced online. Source code is available for the Raspberry Pi. Physiological monitoring e-Health sensor platform The e-Health Sensor Platform30 by Spanish electronics manufacturer Libelium is an open physiological monitoring platform using nine different sensors: pulse, oxygen in blood (SpO2), airflow (breathing), body temperature, ECG, glucometer, galvanic skin response, blood pressure (sphygmomanometer) and patient position (accelerometer). The platform, shown in figures 3 and 4, consists of a board or shield that plugs into an Arduino or Raspberry Pi and interfaces with the various sensors, allowing for experimenting directly with the signals coming from these sensors. Figure 3 Arduino (bottom, showing USB connector) and e-Health shield. UK £1 coin (2.25 cm diameter) for scale. (Photo by Patrick Oladimeji, image licensed under the Creative Commons BY-SA license). Figure 4 Blood glucose meter connected to e-Health shield (Photo by Patrick Oladimeji, image licensed under the Creative Commons BY-SA license).

Figure 3 Arduino (bottom, showing USB connector) and e-Health shield. UK £1 coin (2.25 cm diameter) for scale. (Photo by Patrick Oladimeji, image licensed under the Creative Commons BY-SA license). Figure 4 Blood glucose meter connected to e-Health shield (Photo by Patrick Oladimeji, image licensed under the Creative Commons BY-SA license). To build the device a Raspberry Pi or Arduino is required. No special skills are required, as it comes as a modular kit where the various components are just plugged in. The components provided are off-the-shelf sensors and products that have been modified to work with the kit. Source code is available for Raspberry Pi and Arduino. OpenBCI OpenBCI is an eight-channel EEG signal capture platform (see figure 5) of which the hardware, software and mechanical design files are available online.31 It is intended for human brain-computer interface technologies, but the hardware can also be used to perform other types of biosensing, like EMG and heart rate (EKG). Figure 5 OpenBCI brain-computer interface (Photo by Wikimedia Commons, image licensed under the Creative Commons BY-SA license). All hardware and electrodes are included with the kit. It comes as a modular kit, with the hardware design files available. A 3D-printable EEG headset, called Spiderclaw, is in development. Source code is available for Arduino (8-bit), ChipKit (32-bit), Processing and Python.

Figure 5 OpenBCI brain-computer interface (Photo by Wikimedia Commons, image licensed under the Creative Commons BY-SA license). All hardware and electrodes are included with the kit. It comes as a modular kit, with the hardware design files available. A 3D-printable EEG headset, called Spiderclaw, is in development. Source code is available for Arduino (8-bit), ChipKit (32-bit), Processing and Python. Do-it-yourself blood pressure monitor Russell et al32 published instructions to build a do-it-yourself (DIY) blood pressure monitor. The instructions include a bill of materials and circuit diagrams. The blood pressure monitor relies on an electronic pressure sensor connected to an aneroid sphygmomanometer (a mechanical type with a dial). The monitor runs on batteries that can be recharged using a hand crank generator. They do mention that it is an experimental prototype and that the measurements should not be relied on for clinical use. The device can be built using basic tools from a home hardware store. The various electronic components that can be sourced from distributors, as well as an off-the-shelf sphygmomanometer. An off-the-shelf project box is used to contain the electronics. Source code is available for a Microchip PIC microcontroller, and only basic knowledge of electronics and soldering is required.

store. The various electronic components that can be sourced from distributors, as well as an off-the-shelf sphygmomanometer. An off-the-shelf project box is used to contain the electronics. Source code is available for a Microchip PIC microcontroller, and only basic knowledge of electronics and soldering is required. Diabetes Diabeto Diabeto is a small device that plugs into a glucometer and transfers blood glucose readings to a smartphone app. The data can then be displayed and analysed on the Diabeto web app, the mobile app as well as the Pebble smartwatch. The hardware design files, bill of materials and user instructions have been made available online.33 The Diabeto is available as a commercial off-the-shelf device, but the hardware design files are available online. The various electronic components can be sourced from distributors. The bill of materials includes manufacturer names and product codes. The enclosure and software do not seem to be open-source. Nightscout xDrip The Nightscout project34 is an open-source project that allows real-time access to a continuous glucose monitor (CGM) connected to a mobile phone, sending the data to a web app, smartphone app and smartwatch app. xDrip35 is an open-source hardware device that reads the wireless signals transmitted by the CGM sensor and sends data to a smartphone via Bluetooth Low-Energy (BLE). From the smartphone app the data can also be transmitted to the Nightscout database.

e, sending the data to a web app, smartphone app and smartwatch app. xDrip35 is an open-source hardware device that reads the wireless signals transmitted by the CGM sensor and sends data to a smartphone via Bluetooth Low-Energy (BLE). From the smartphone app the data can also be transmitted to the Nightscout database. To build the device, a soldering iron and a Dexcom CGM is needed. There are four electronic components that can be sourced from open-source hardware distributors: BLE module, USB wireless module, battery and charger, and wires. Source code for an Android mobile app is available. Conclusions Open-source medical devices are already available, even if they are not marketed as such. Table 1 shows a summary of the 10 devices that were compared in this paper. Evidently, open-source offers a solution to the high costs and slow pace of innovation of medical devices currently. Off-the-shelf healthcare sensors are available that can be connected to cheap embedded computers, and it is easy to construct practical systems that can be used for inventing new clinical applications that work. Rapidly achieving sophisticated results that would have been impossible only a year or two ago transforms one's attitude to medical devices: it is interesting, rewarding, and stimulates creative innovation. Table 1 Summary of project comparison

Conclusions Open-source medical devices are already available, even if they are not marketed as such. Table 1 shows a summary of the 10 devices that were compared in this paper. Evidently, open-source offers a solution to the high costs and slow pace of innovation of medical devices currently. Off-the-shelf healthcare sensors are available that can be connected to cheap embedded computers, and it is easy to construct practical systems that can be used for inventing new clinical applications that work. Rapidly achieving sophisticated results that would have been impossible only a year or two ago transforms one's attitude to medical devices: it is interesting, rewarding, and stimulates creative innovation. Table 1 Summary of project comparison Project Estimated cost to build OSHW license used Microcontroller/platform used MyOpen10–13 $250 GNU public license10 Blackfin DSP e-Nable16–18 $350 Creative commons attribution non-commercial Arduino CT scanner20–22 $300 GNU public license V.3 Arduino Fechko's peristaltic pump26 $170 N/A Raspberry Pi or Arduino Wijnen et al's syringe pump8 $971 Creative commons attribution share-alike Raspberry Pi e-Health sensor platform30 $500 Unknown Raspberry Pi or Arduino OpenBCI31 $450–$800 Creative commons share-alike Arduino or ChipKit DIY blood pressure monitor32 $50 N/A Microchip PIC Diabeto33 $50 Unknown Atmel ATTiny Nightscout xDrip34 35 $60 N/A Pololu Wixel DIY, do-it-yourself; DSP, digital signal processing; N/A, not applicable; OSHW, open-source hardware.

spberry Pi or Arduino OpenBCI31 $450–$800 Creative commons share-alike Arduino or ChipKit DIY blood pressure monitor32 $50 N/A Microchip PIC Diabeto33 $50 Unknown Atmel ATTiny Nightscout xDrip34 35 $60 N/A Pololu Wixel DIY, do-it-yourself; DSP, digital signal processing; N/A, not applicable; OSHW, open-source hardware. There remain many challenges for open-source hardware in healthcare. Medical devices need to be regulated, and this can be time consuming and expensive, but still quite manageable for an open-source project. Crowdfunding is one way of getting the funds together to apply for certification. There are successful open business models for open-source hardware, making it possible to invest in them as commercial entities. Finally, note that while open-source hardware does not use patents for intellectual property protection, open-source hardware companies generally do use trademarks to protect their brands and to assure a level of quality in the products. Twitter: Follow Gerrit Niezen at @gendor Contributors: GN and PE conceived the work. GN, PE and HT wrote this manuscript. Funding: This work was supported by the UK Engineering and Physical Sciences Research Council (grant number EP/G059063). Competing interests: GN reports personal fees from Tidepool, a non-profit developing open-source software for type 1 diabetes, outside the submitted work. PE and HT have nothing to disclose. Provenance and peer review: Not commissioned; externally peer reviewed.

Introduction The neonatal intensive care unit (NICU) in a hospital attends to the needs of critically ill newborn babies, some of whom are born prematurely. The babies may weigh as little as 500 g compared with 3.5 kg, the normal birth weight of a term baby. The care provided is among the most intensive, specialised and high dependency within the hospital environment. Uniquely, the unit combines this clinical care with attending to each baby's individual developmental needs, as well as supporting the parents as they bond with their child. Various technologies assist the clinical teams as shown in figure 1. Babies often lie within individual incubators with some form of respiratory support via invasive or non-invasive ventilation, as well as intravenous infusion lines to administer nutrition and drugs, as shown in figure 2. In addition to these interventions, multiple monitors are used to provide continuous quantification of vital signs. The sensors used in this environment include electrocardiogram (ECG) probes (heart activity), thermistors (skin temperature) and pulse oximeters (blood oxygen saturation, SpO2). These sensors are all connected by wires to a monitoring unit outside of the incubator. There is also a pressure sensor connected to a separate sleep apnoea alarm. The large number of wires in and around the incubator might be reduced by introducing wireless systems, but these do not currently exist in the NICU and limited research1 has been conducted in implementing such a system.

tside of the incubator. There is also a pressure sensor connected to a separate sleep apnoea alarm. The large number of wires in and around the incubator might be reduced by introducing wireless systems, but these do not currently exist in the NICU and limited research1 has been conducted in implementing such a system. Figure 1 The neonatal care environment. The display in the top left visualises the vital sign signals from the ECG probes, pulse oximeter cuff and the temperature sensor. The unit in the bottom right is a non-invasive ventilator. The incubator provides a thermally (and, if required, humidity) regulated environment for the baby. Figure 2 A premature baby inside an incubator in the neonatal intensive care unit. The sticker attached to the chest is one of three ECG probes. A pulse oximeter cuff is secured to the foot, and a temperature probe is attached to the baby's back. The nasal tubing provides non-invasive ventilation.

Figure 1 The neonatal care environment. The display in the top left visualises the vital sign signals from the ECG probes, pulse oximeter cuff and the temperature sensor. The unit in the bottom right is a non-invasive ventilator. The incubator provides a thermally (and, if required, humidity) regulated environment for the baby. Figure 2 A premature baby inside an incubator in the neonatal intensive care unit. The sticker attached to the chest is one of three ECG probes. A pulse oximeter cuff is secured to the foot, and a temperature probe is attached to the baby's back. The nasal tubing provides non-invasive ventilation. Reducing the number of wires that are attached to the babies could make the delivery of clinical care easier and might encourage more physical interaction between parents and their babies. Physical interaction is an important part of parental care during the neonatal period, often taking the form of ‘kangaroo care’, where, for example, a mother places the baby against her chest for skin-to-skin contact. For those in intensive care, the babies normally would still be connected to life-supporting equipment and have limited mobility; therefore, this is always undertaken in a chair next to the incubator or cot. This practice is encouraged as part of the widely used Newborn Individualized Developmental Care and Assessment Program (NIDCAP) to enhance parent–baby bonding and has been shown to improve clinical outcomes.2 Kangaroo care is particularly critical for preterm neonates.3–6 The clinical impact of this bonding experience is pronounced, with respect to short-term physiological stability and long-term health outcomes. Benefits for the baby include less pain,6 better sleeping,3 improved weight gain3 4 and earlier discharge.3 Conversely, poor parent–baby bonding ‘impairs hormonal, epigenetic and neuronal development in preterm infants’.7

e is pronounced, with respect to short-term physiological stability and long-term health outcomes. Benefits for the baby include less pain,6 better sleeping,3 improved weight gain3 4 and earlier discharge.3 Conversely, poor parent–baby bonding ‘impairs hormonal, epigenetic and neuronal development in preterm infants’.7 A few studies provide a general view of how technology in the NICU affects parents and their interactions with their baby; however, there is no research exclusively focused on how sensor wires impact care in the NICU. An interview study into the effects of technology on parents in the NICU reported that they found the environment to be ‘oppressive’ and that it ‘delayed the development of their ability to participate in the care of the child’.8 A separate questionnaire survey identified those pieces of equipment which parents saw as a barrier to interacting with their baby9: the monitoring equipment was identified as a lower obstruction compared with more invasive interventions such as respiratory support or infusion lines. While these prior studies provide an important context for understanding the parents' experience of the NICU and its equipment, there are limitations. To enable the effective development of wireless solutions, we require a better understanding of how the present wired technologies are experienced and how a future wireless system might be received. The study reported here addresses this and forms part of the initial scoping exercise for the technological development of a wireless system for use in intensive care.

s solutions, we require a better understanding of how the present wired technologies are experienced and how a future wireless system might be received. The study reported here addresses this and forms part of the initial scoping exercise for the technological development of a wireless system for use in intensive care. Methods Cohort selection and recruitment The study involved semistructured interviews10 with parents and nursing staff from a single NICU. The interviews were conducted at the University of Cambridge Addenbrooke's Hospital Trust in February 2016 with nurses and parents on the unit. The sample size was determined by the number of suitable parents on the unit, their emotional state and the consequent imposition on the nurses' time. Nurses were recruited into the study through distribution of information sheets; suitable parents were identified by the nurse in charge of the unit. Six parents and seven nurses were interviewed (see tables 1 and 2). For comparison, this is in line with other studies with similar participant groups.11 Five of the nurses were based in the NICU and two worked in the acute neonatal transfer service (ANTS), which is based at the same location. Each of the parents had one baby on the unit. Five of the babies were born at extremely low birth weight (<1 kg); three were born extremely prematurely (<28 weeks), two others very premature (28–32 weeks) and one moderately premature (32–37 weeks).12 13 Table 1 Parent participants Participant code Time on unit P1 4 weeks P2 12 weeks P3 6 weeks P4 16 weeks P5 6 weeks P6 4 weeks Table 2 Nurse participants

Methods Cohort selection and recruitment The study involved semistructured interviews10 with parents and nursing staff from a single NICU. The interviews were conducted at the University of Cambridge Addenbrooke's Hospital Trust in February 2016 with nurses and parents on the unit. The sample size was determined by the number of suitable parents on the unit, their emotional state and the consequent imposition on the nurses' time. Nurses were recruited into the study through distribution of information sheets; suitable parents were identified by the nurse in charge of the unit. Six parents and seven nurses were interviewed (see tables 1 and 2). For comparison, this is in line with other studies with similar participant groups.11 Five of the nurses were based in the NICU and two worked in the acute neonatal transfer service (ANTS), which is based at the same location. Each of the parents had one baby on the unit. Five of the babies were born at extremely low birth weight (<1 kg); three were born extremely prematurely (<28 weeks), two others very premature (28–32 weeks) and one moderately premature (32–37 weeks).12 13 Table 1 Parent participants Participant code Time on unit P1 4 weeks P2 12 weeks P3 6 weeks P4 16 weeks P5 6 weeks P6 4 weeks Table 2 Nurse participants Participant code Position N1 Senior sister N2 Senior sister N3 Advanced neonatal practitioner N4 Junior sister N5 Senior staff nurse N6 Junior sister (ANTS) N7 Junior sister (ANTS) ANTS, acute neonatal transfer service.

Participant code Time on unit P1 4 weeks P2 12 weeks P3 6 weeks P4 16 weeks P5 6 weeks P6 4 weeks Table 2 Nurse participants Participant code Position N1 Senior sister N2 Senior sister N3 Advanced neonatal practitioner N4 Junior sister N5 Senior staff nurse N6 Junior sister (ANTS) N7 Junior sister (ANTS) ANTS, acute neonatal transfer service. Interview development All participants were interviewed individually except the nurses from the ANTS who requested they be interviewed together. The interviews followed an established procedure for performing semistructured interviews,10 14 and were conducted by the first author who is an electronics engineer undertaking research into developing wireless monitoring for neonates. The interview dialogue was initiated with predefined questions which were intended to elicit responses on the subject of the two main research themes: current perceptions of the existing patient monitoring systems and their peripheral attachments; future perceptions of a system which removes the wires between the sensor probe and the patient monitoring device.

Interview development All participants were interviewed individually except the nurses from the ANTS who requested they be interviewed together. The interviews followed an established procedure for performing semistructured interviews,10 14 and were conducted by the first author who is an electronics engineer undertaking research into developing wireless monitoring for neonates. The interview dialogue was initiated with predefined questions which were intended to elicit responses on the subject of the two main research themes: current perceptions of the existing patient monitoring systems and their peripheral attachments; future perceptions of a system which removes the wires between the sensor probe and the patient monitoring device. The first part of each interview focused on eliciting the participant's views and experiences of the current systems. The second part involved introducing the concept of a wireless system and exploring the participant's views on its potential impact. The potential wireless system was described to the participants as functionally equivalent to the current setup, but with two physical changes: the connecting wires would be removed; and a small electronics module would be added to the sensor for transmitting the data to be displayed on the monitor.

s potential impact. The potential wireless system was described to the participants as functionally equivalent to the current setup, but with two physical changes: the connecting wires would be removed; and a small electronics module would be added to the sensor for transmitting the data to be displayed on the monitor. Interviews with the participants were conducted in a private room near to the unit. Furthermore, a senior nurse, familiar to the parent, was present in the room to provide support or answer clinical questions. Interviews were recorded using a digital voice recorder to allow later analysis. The interviews lasted around 10 min each; a debrief was conducted post interview. Data analysis Any specifically identifiable data was anonymised during interview transcription. Analysis of the interviews was performed using an established procedure for a general inductive method using thematic analysis.14–17 The interview transcripts were processed using ATLAS.ti (Scientific Software Development GmbH) which permits the iterative assignment of themes to segments of transcript. Results The analysis led to the identification of three overall themes that are used to structure the report below, with quotations used to illustrate key points. The quotations have been edited for clarity and brevity (with edits in square brackets). Participant codes from tables 1 and 2 are included after the quotation (Px for parents and Nx for nurses).

fication of three overall themes that are used to structure the report below, with quotations used to illustrate key points. The quotations have been edited for clarity and brevity (with edits in square brackets). Participant codes from tables 1 and 2 are included after the quotation (Px for parents and Nx for nurses). Parental care: physical and psychological barriers The nurses and parents discussed what they perceived to be the main obstructions to parental care in the NICU. These comprised physical and psychological barriers. For the parents, the experience of having their baby in the NICU is clearly a very emotional one. Specific emotions that were expressed as a direct result of seeing their baby attached to the wires were intimidation, sadness, shock and fear. The parents identified their first encounter with the unit as the toughest, but as time progressed, they became more used to it.Crickey. There was a lot of wires, a heck of a lot of wires—there were more wires than him. So yeah, it was quite overwhelming. [P1] It took ages, about a month at least, [for the anxiety about the wires to settle]. [P5]

For the parents, the experience of having their baby in the NICU is clearly a very emotional one. Specific emotions that were expressed as a direct result of seeing their baby attached to the wires were intimidation, sadness, shock and fear. The parents identified their first encounter with the unit as the toughest, but as time progressed, they became more used to it.Crickey. There was a lot of wires, a heck of a lot of wires—there were more wires than him. So yeah, it was quite overwhelming. [P1] It took ages, about a month at least, [for the anxiety about the wires to settle]. [P5] The monitoring equipment and other support systems had a significant impact on the parents' physical interaction with their babies. Most of the parents indicated a reluctance to touch their baby as a result of the wires and monitoring systems, because of the fear it induced.It made me frightened with all the wires that were there. It made me feel very uncomfortable and I didn't sort of want to…I wanted to hold him but I felt like I couldn't because there was a lot of equipment there. [P1] The first thing the parents always say is “look at the wires” and that's the first barrier. [N5]

The monitoring equipment and other support systems had a significant impact on the parents' physical interaction with their babies. Most of the parents indicated a reluctance to touch their baby as a result of the wires and monitoring systems, because of the fear it induced.It made me frightened with all the wires that were there. It made me feel very uncomfortable and I didn't sort of want to…I wanted to hold him but I felt like I couldn't because there was a lot of equipment there. [P1] The first thing the parents always say is “look at the wires” and that's the first barrier. [N5] The majority of the parents said that while handling their baby, they were worried about disturbing the equipment, and all of them commented that skin-to-skin contact was negatively affected by the wires (a view that was supported by the nurses). This was because of the time it took to take the baby out (and then put them back) and because the parents felt the wires were uncomfortable against the skin.Taking the baby out for kangaroo care just takes ages. I have to detangle her, then the nurses have to take all the wires off, then they take her out, then they connect her all back up. It takes a good half an hour just to get her out for a cuddle, especially as she wriggles a lot—she does tend to tangle all her wires into one, or she pulls them out. She's a bit feisty! [P5] Both participant groups reported that the wires cause problems for the parents and the way they interact with their baby. These problems affect their emotional state and are a barrier to physical interactions including kangaroo care.

The majority of the parents said that while handling their baby, they were worried about disturbing the equipment, and all of them commented that skin-to-skin contact was negatively affected by the wires (a view that was supported by the nurses). This was because of the time it took to take the baby out (and then put them back) and because the parents felt the wires were uncomfortable against the skin.Taking the baby out for kangaroo care just takes ages. I have to detangle her, then the nurses have to take all the wires off, then they take her out, then they connect her all back up. It takes a good half an hour just to get her out for a cuddle, especially as she wriggles a lot—she does tend to tangle all her wires into one, or she pulls them out. She's a bit feisty! [P5] Both participant groups reported that the wires cause problems for the parents and the way they interact with their baby. These problems affect their emotional state and are a barrier to physical interactions including kangaroo care. Delivering medical care: practical issues with the current system The participants felt that some aspects of the existing systems presented challenges to the delivery of medical care.

Both participant groups reported that the wires cause problems for the parents and the way they interact with their baby. These problems affect their emotional state and are a barrier to physical interactions including kangaroo care. Delivering medical care: practical issues with the current system The participants felt that some aspects of the existing systems presented challenges to the delivery of medical care. The nurses generally have a positive view of the current system, as it is what they have been trained to use, and have experience of using over many years. They understand the subtle challenges that some of the systems pose—for instance the poor adherence of the ECG stickers in the humid incubators—and mitigate this by changing the ECG probes more frequently.If you have a small baby in humidity, some of the monitoring [probes] fall off because of the humidity. So you might lose a trace of the ECG. [N2] The temperature probe sticker was identified as adhering better, as the placement underneath the baby kept it securely attached and a stronger adhesive is used. However, one parent noted that on removal, the area of attachment could be sore. The pulse oximeter probes used on this NICU were a soft-touch, velcro-secured wrap which required changing every six hours to prevent pressure sores. One nurse commented that, owing to the fragility of the babies' skin, the probe needed to be moved frequently to avoid burns caused by heat generated from the pulse oximeter.

pulse oximeter probes used on this NICU were a soft-touch, velcro-secured wrap which required changing every six hours to prevent pressure sores. One nurse commented that, owing to the fragility of the babies' skin, the probe needed to be moved frequently to avoid burns caused by heat generated from the pulse oximeter. Achieving good contact between the sensor probe and the skin of the neonate is difficult because of the fragility of their poorly developed skin. This can lead to complications when changing the probes as the sticky pad can remove some of the baby's skin.Very premature babies, [that is], less than 25/26 weeks, their skin is obviously more friable and they're more likely to get skin breakdown. [N1] One of the nurses identified that excessive handling of very premature babies can have a negative impact on their development, saying that anything that would ease the transfer of the babies would greatly reduce the stress they experience. Trying to untangle wires can increase the need to move and handle the babies.The less you can handle the very fragile babies the better. It has a massive clinical impact—stress—on very sick and premature babies. So from a developmental care point of view, even just the small differences can help, if the move from one place to another is smoother. [N6]

ease the need to move and handle the babies.The less you can handle the very fragile babies the better. It has a massive clinical impact—stress—on very sick and premature babies. So from a developmental care point of view, even just the small differences can help, if the move from one place to another is smoother. [N6] Half of the parents and the majority of the nurses reported that there could be a great deal of ‘clutter’ of tangled wires in the incubator. For the nurses, this caused problems when repositioning the babies; for the parents, it complicated the process of taking the babies out for kangaroo care. The ‘clutter’ is caused by the many different medical systems that are located outside of the incubator but which must be connected to the baby. Furthermore, nurses are aware of the risk of pressure sores from babies lying on wires in the incubator.Quite often if babies have got quite a lot of monitoring on, the wires are in a bit of a knot at the bottom of the bed. [N2] Yeah, it is awkward, having to untangle everything—untangle his leg. [P4] One nurse highlighted baby positioning as being extremely important for their care, which she felt was hindered by the number of wires in the incubator.Positioning is really important with premature babies—really really important, and the wires can make that [trickier]. [N6] The participants identified these inadequacies in the current, wired system for vital sign monitoring. Wireless technologies might help to mitigate these problems if such systems were accepted in the NICU.

One nurse highlighted baby positioning as being extremely important for their care, which she felt was hindered by the number of wires in the incubator.Positioning is really important with premature babies—really really important, and the wires can make that [trickier]. [N6] The participants identified these inadequacies in the current, wired system for vital sign monitoring. Wireless technologies might help to mitigate these problems if such systems were accepted in the NICU. Proposal for a wireless system: perceived benefits and risks All of the participants responded positively to the idea of a wireless system when it was first introduced to them.I think it could be a great idea actually, because if you don't have anything between the baby and the monitor, I think it's going to be easier to move the babies. [N4] It would just be easier; I wouldn't be so scared to touch her. I think you'd be a lot more likely to touch the baby without the wires there. [P5]

Proposal for a wireless system: perceived benefits and risks All of the participants responded positively to the idea of a wireless system when it was first introduced to them.I think it could be a great idea actually, because if you don't have anything between the baby and the monitor, I think it's going to be easier to move the babies. [N4] It would just be easier; I wouldn't be so scared to touch her. I think you'd be a lot more likely to touch the baby without the wires there. [P5] Perceived benefits of the wireless system included improved comfort for the baby in the incubator because there would be no wires to lie on; a less worrying visual appearance of the baby which would reduce anxiety for the parents; and better physical interaction with the baby, in particular easier and more comfortable kangaroo care.I think it would be great to have it be wireless, if it can be as effective [as the present system]. As a parent you become a little bit obsessed by them [the wires] at first, because they take your attention away from looking [at your baby]. Certainly [having a wireless system] would improve things, [it would] make you feel more comfortable. You'd be less entangled. [P3] It would be really good when we take babies out for kangaroo care, obviously all these wires trailing about causes a lot of problems. Especially if there are lots of fluids lines as well. So that would make a huge difference. [N1]

Perceived benefits of the wireless system included improved comfort for the baby in the incubator because there would be no wires to lie on; a less worrying visual appearance of the baby which would reduce anxiety for the parents; and better physical interaction with the baby, in particular easier and more comfortable kangaroo care.I think it would be great to have it be wireless, if it can be as effective [as the present system]. As a parent you become a little bit obsessed by them [the wires] at first, because they take your attention away from looking [at your baby]. Certainly [having a wireless system] would improve things, [it would] make you feel more comfortable. You'd be less entangled. [P3] It would be really good when we take babies out for kangaroo care, obviously all these wires trailing about causes a lot of problems. Especially if there are lots of fluids lines as well. So that would make a huge difference. [N1] Three of the nurses (including the two nurses from ANTS) indicated that a wireless system would be beneficial when babies are being transported, whether within or between hospitals. The simplification of the movement process was identified as the main reason that a wireless system would help in this setting.Well I think it would be great on transport. We have very limited space as it is with the incubators—they're very small, very compact. So having a baby where you haven't got all the additional wires and all the cabling coming in [would be] a plus point. [N7]

main reason that a wireless system would help in this setting.Well I think it would be great on transport. We have very limited space as it is with the incubators—they're very small, very compact. So having a baby where you haven't got all the additional wires and all the cabling coming in [would be] a plus point. [N7] Some of the participants raised concerns regarding a wireless system, including issues related to battery life, reliability of the wireless link, ensuring the right sensor connects to the right monitor and quality of data not being compromised. Another concern was heating of the babies from the sensor unit, lest it damage their extremely fragile skin. One nurse was worried that without obvious wires to act as a reminder, a sensor might be forgotten about and left attached to the baby. However, she went on to say that this would be unlikely as the babies are checked regularly. The size and weight increase of a wireless sensor attached to the skin was also identified as a potential problem as it could lead to pressure sores or interfere with phototherapy.For a 23/24 weeker with very fragile skin, [I wonder] how much that bulk is going to impact on that baby; because obviously the leads are very light. So that would worry me, that they'd cause a pressure injury. [N3] The concept of a wireless system was well received by all the participants, with many areas of care seen as potentially benefiting.

The size and weight increase of a wireless sensor attached to the skin was also identified as a potential problem as it could lead to pressure sores or interfere with phototherapy.For a 23/24 weeker with very fragile skin, [I wonder] how much that bulk is going to impact on that baby; because obviously the leads are very light. So that would worry me, that they'd cause a pressure injury. [N3] The concept of a wireless system was well received by all the participants, with many areas of care seen as potentially benefiting. Discussion This study has built on the previous literature by seeking to understand the experiences that nurses and parents have with one particular aspect of the medical technology found in the NICU—the monitoring system, and especially its wires. Furthermore, we have explored how these stakeholders responded to the idea of a wireless system which might help to overcome some of the challenges they currently experience with the wired sensors.

ve with one particular aspect of the medical technology found in the NICU—the monitoring system, and especially its wires. Furthermore, we have explored how these stakeholders responded to the idea of a wireless system which might help to overcome some of the challenges they currently experience with the wired sensors. Similar to previous studies in this field, we found that the wires interfere with parent–baby bonding despite the nurses' best efforts.8 The parents' ability to be involved in their babies' care has previously been found to be one of the most important factors in determining parental satisfaction with the care provided.18 Although babies would still be attached to the incubator by other life-supporting systems such as infusion lines and breathing support, the findings of this present study suggest that by removing the wires connecting the sensors to the vital sign monitors, we can improve the parent–baby bonding experience and therefore improve parent satisfaction. Furthermore, because the wires present a number of challenges to the provision of clinical care by the nursing team, moving to a wireless system may also reduce the discomfort suffered by the babies.

to the vital sign monitors, we can improve the parent–baby bonding experience and therefore improve parent satisfaction. Furthermore, because the wires present a number of challenges to the provision of clinical care by the nursing team, moving to a wireless system may also reduce the discomfort suffered by the babies. Previous work investigating nurses' attitudes towards technology in an intensive care setting found that, while it could complicate their everyday practice, it was recognised as being ‘good for patients’, bringing a range of benefits.19 In line with this, all participants interviewed in the present study had an overall positive attitude towards a wireless version of the monitoring system and many expressed the view that the wireless system would improve the kangaroo care experience. When concerns were raised, they centred on expectations that the size of the device must be small and unobtrusive, the data needs to be as good as the current system and the system needs to be completely reliable. While no adverse effects have been demonstrated to result from radio frequency radiation exposure within the regulated limits,20 it might plausibly have been raised as an issue given public concerns. However, none of the participants expressed concerns over this, indicating that it might not be a significant barrier to uptake.

adverse effects have been demonstrated to result from radio frequency radiation exposure within the regulated limits,20 it might plausibly have been raised as an issue given public concerns. However, none of the participants expressed concerns over this, indicating that it might not be a significant barrier to uptake. Current standard clinical care requires many wires connecting multiple sensors and monitors to safely manage babies within the NICU. These wires pose additional challenges to the parents and nurses who work together to provide care for the babies. In particular, the wires can negatively impact the provision of kangaroo care and can complicate the repositioning and transportation of the babies. Developing a wireless system could mitigate these problems if the technology is implemented well and accepted by those in the NICU. Recent advances in wireless technology could realise this opportunity but there remain challenges in designing a device suitable for this unique environment.

transportation of the babies. Developing a wireless system could mitigate these problems if the technology is implemented well and accepted by those in the NICU. Recent advances in wireless technology could realise this opportunity but there remain challenges in designing a device suitable for this unique environment. To address these challenges, further work is planned to design and test a wireless vital sign monitoring system on the NICU. As part of the design work, user studies will be conducted which will further assess the impact the wireless system will have. By redesigning the interface between the sensor and the monitoring unit, we are influencing the relationship between various technologies and the babies, nurses, parents and others involved in their care. Prior work has shown that these relationships are complex,8 19 and so we must be mindful of the potential for negative or unintended consequences. That said, this present study has shown that wireless monitoring on the NICU has great potential and that key user groups would welcome innovation in this area. Correction notice: This article has been corrected since it was published Online First. Page 1, paragraph 2, the sentence ‘Various technologies assists the clinical teams...’ has been revised to read ‘Various technologies assist the clinical teams...’. Also, figure 2 caption has been revised to read ‘A pulse oximeter...’.

Correction notice: This article has been corrected since it was published Online First. Page 1, paragraph 2, the sentence ‘Various technologies assists the clinical teams...’ has been revised to read ‘Various technologies assist the clinical teams...’. Also, figure 2 caption has been revised to read ‘A pulse oximeter...’. We would like to thank the participants for giving their time to the study; the nursing team in the NICU at Addenbrooke's Hospital, Cambridge, for their support in conducting the interviews; and Dr Paul Robertson of the University of Cambridge Engineering Department, who provided guidance throughout the study. The raw interview data from this study cannot be made freely available because inherent to that data is sensitive information relating to the individuals and organisations involved. Contributors: OB devised and carried out the interview study with the help of KB and JL. Data was processed by OB, with guidance from NC. All authors gave final approval for publication. Funding: This project was funded by the EPSRC CDT in Sensor Technologies and Applications (grant number EP/L015889/1). Competing interests: None declared. Patient consent: Obtained. Ethics approval: The study received ethical approval from the University of Cambridge Engineering Department's Research Ethics Committee. The study was further registered with Addenbrooke's Hospital Patient Experience Department. All participants consented to being quoted anonymously in publication. Provenance and peer review: Not commissioned; externally peer reviewed.

Introduction Postpartum haemorrhage (PPH) is an obstetric emergency that can follow vaginal or caesarean delivery. It is estimated that each year PPH accounts for 27% of the 303 000 maternal deaths worldwide1 2 and a further 20 million women suffer long-term effects.3 Although the absolute risk of death is much lower in high-income countries (1 in 100 000 vs 1 in 1000 births in low-income countries),4 and in spite of marked improvements in management, PPH remains a significant contributor to maternal morbidity and mortality throughout the world.5 A low-cost, effective intervention that can be used by first-level maternity services providers could be a major advance in reducing maternal mortality from PPH, especially in low resource settings where the majority of deaths occur.

Introduction Postpartum haemorrhage (PPH) is an obstetric emergency that can follow vaginal or caesarean delivery. It is estimated that each year PPH accounts for 27% of the 303 000 maternal deaths worldwide1 2 and a further 20 million women suffer long-term effects.3 Although the absolute risk of death is much lower in high-income countries (1 in 100 000 vs 1 in 1000 births in low-income countries),4 and in spite of marked improvements in management, PPH remains a significant contributor to maternal morbidity and mortality throughout the world.5 A low-cost, effective intervention that can be used by first-level maternity services providers could be a major advance in reducing maternal mortality from PPH, especially in low resource settings where the majority of deaths occur. A vital step in the physiological prevention of PPH is the immediate contraction and retraction of myometrial muscle fibres during and after the third stage of labour. Uterine atony is a condition characterised by the inability of the uterus to contract adequately after the placenta has separated from the uterus. This condition is thought to be the most common cause of PPH6 7 and is often unpredictable.8 Its presentation is very similar to that of genital trauma and often the only way to distinguish the two is under anaesthetic in an operating theatre. Initial treatment therefore assumes that the cause of any postpartum bleeding is atony, with formal examination under anaesthetic only if this treatment fails. However, there is evidence that repeated use of uterotonics has little effect.9–11 Furthermore, on-going haemorrhage caused by delays in stopping the bleeding can lead to coagulation problems, which only serves to worsen the bleeding and sets up a vicious cycle. Given all this, it is fortunate that most women with PPH stop bleeding spontaneously.12 However, the inability to predict who will stop bleeding spontaneously forces caregivers to intervene aggressively at an early stage in the process on all women.

which only serves to worsen the bleeding and sets up a vicious cycle. Given all this, it is fortunate that most women with PPH stop bleeding spontaneously.12 However, the inability to predict who will stop bleeding spontaneously forces caregivers to intervene aggressively at an early stage in the process on all women. The theory behind the device Bimanual compression (BMC) is an old technique in which the uterus is compressed between a hand on the lower abdomen and a hand inserted into the vagina and formed into a fist. Although it is highly effective, it is a very painful manoeuvre for the mother (unless she has an epidural in place) and tiring for the practitioner. The idea of the Butterfly device is to achieve the benefits of BMC without being so invasive, thus allowing it to be more widely used. It has been designed to be a slim, easily insertable replacement to a fist in the vagina, thus increasing acceptability of uterine compression to women and clinicians. The size and shape of the compression platform was based on that of a shelf pessary, an intravaginal device used for treating uterine prolapse. The depth of insertion was based on the length of a standard Cusco's vaginal speculum (figure 1—shelf pessary and Cusco's speculum combined). This should make uterine compression available for use at a much earlier stage in the PPH treatment process and provide an effective treatment without the need for medicines or advanced diagnostic skills (figure 2).

he length of a standard Cusco's vaginal speculum (figure 1—shelf pessary and Cusco's speculum combined). This should make uterine compression available for use at a much earlier stage in the PPH treatment process and provide an effective treatment without the need for medicines or advanced diagnostic skills (figure 2). Figure 1 A composite picture combining a shelf pessary (used for the treatment of uterovaginal prolapse) and a Cusco speculum (used for vaginal and cervical examination). Figure 2 Diagram showing the normal progression of severe PPH. Use of the PPH Butterfly should provide an early diagnosis of the underlying cause of the bleeding while preventing on-going blood loss. PPH, postpartum haemorrhage. Once inserted, the uterus is compressed against the PPH Butterfly by a hand externally on the woman's abdomen. The device will be held in place by its handle, which can be wedged against the bed, or be held by the clinician or an assistant.

Figure 2 Diagram showing the normal progression of severe PPH. Use of the PPH Butterfly should provide an early diagnosis of the underlying cause of the bleeding while preventing on-going blood loss. PPH, postpartum haemorrhage. Once inserted, the uterus is compressed against the PPH Butterfly by a hand externally on the woman's abdomen. The device will be held in place by its handle, which can be wedged against the bed, or be held by the clinician or an assistant. The tiring nature of traditional BMC is an important issue which limits its effectiveness. In training sessions, we demonstrate this to junior doctors by asking them to compress a book between fist and hand—but they tire rapidly and many are unable to continue this for more than 60 s. However, in keeping with physiological principles of uterine contractions (that generally only occur every 3 min) and blood clots (that take 3–5 min to develop), a pressure period of just 1 min is grossly inadequate except as a temporising measure. To achieve lasting haemostasis, a compression for 8–10 min is needed and this is unlikely to be obtained effectively with standard techniques. To enable prolonged pressure using the Butterfly device, the handle is constructed so as to allow the clinician to stabilise it against the bed once it is inserted. Thus, the clinician's weight can be used to stabilise the device against the bed as well as putting pressure on the uterus—this is a far more ergonomically efficient technique.

ressure using the Butterfly device, the handle is constructed so as to allow the clinician to stabilise it against the bed once it is inserted. Thus, the clinician's weight can be used to stabilise the device against the bed as well as putting pressure on the uterus—this is a far more ergonomically efficient technique. The PPH Butterfly is also designed to work as a management tool that assists in the diagnosis of primary PPH. There are four well-recognised causes of PPH; uterine atony, genital tract trauma, retained products of conception and bleeding disorders. The most common causes are uterine atony followed by genital tract trauma, but the two are often difficult to differentiate clinically. However, if the bleeding stops with uterine compression, the cause is almost certainly uterine atony: if it continues then it is likely to be from vaginal lacerations (figure 3). The surface of the platform of the device resembles a grill, with multiple holes. This gives a surface against which to compress the uterus, while preventing the trapping of blood and clots above it. Figure 3 The platform of the PPH Butterfly rests just beneath the uterine cervix (shown as a red dotted line above). Bleeding from above the red line is usually due to uterine atony and will stop abruptly with uterine compression. That from below the red line is from vaginal lacerations and blood loss will continue even with uterine compression. PPH, postpartum haemorrhage.

neath the uterine cervix (shown as a red dotted line above). Bleeding from above the red line is usually due to uterine atony and will stop abruptly with uterine compression. That from below the red line is from vaginal lacerations and blood loss will continue even with uterine compression. PPH, postpartum haemorrhage. If this concept could be realised, it could transform care in PPH. In high resource settings, it would have the effect of preventing anaemia, as well as the need for anaesthetics, intensive care, blood transfusions and the progression to clotting disorders. If it could also be used in poorer settings where many of these technologies are unavailable, then it could also save many mother's lives every year.

tings, it would have the effect of preventing anaemia, as well as the need for anaesthetics, intensive care, blood transfusions and the progression to clotting disorders. If it could also be used in poorer settings where many of these technologies are unavailable, then it could also save many mother's lives every year. Methods The PPH Butterfly was invented by Professor Weeks and John Porter, with the intellectual property held by the University of Liverpool. Following a grant from NIHR i4i funding stream, prototypes were designed and developed with the Department of Medical Physics and Clinical Engineering, University of Liverpool/Royal Liverpool University and Broadgreen Hospitals NHS Trust (RLBUHT). A collaborative relationship was established with Pelican Feminine Healthcare(Cardiff, UK) early in the project. Their input involved an advisory role as part of a twice-yearly Scientific Advisory Group, and optimising the computer-aided design for injection moulding through their design team at Interplex Select Moulds (High Wycombe, UK). The final injection moulding of the model for in vivo testing was undertaken by Proto Labs (Telford, UK) with assistance from Plastribution (Ashby-de-la-Zouch, UK).

ic Advisory Group, and optimising the computer-aided design for injection moulding through their design team at Interplex Select Moulds (High Wycombe, UK). The final injection moulding of the model for in vivo testing was undertaken by Proto Labs (Telford, UK) with assistance from Plastribution (Ashby-de-la-Zouch, UK). As part of the overall project, a series of focus groups were held so as to evaluate the initial concept and design and then review subsequent developments. There were three focus groups: two comprised of mixed groups of midwives and obstetricians, and one comprised of local female members of the public. The first groups explored the groups' initial views of the PPH Butterfly—its aesthetics, purpose, functionality and perceived problems, and sought suggestions for improvements. Further focus groups were held a year later when they were updated as to the progress of the project, saw the outcome of their previous comments and reviewed the progression from the original design. Updates on the PPH Butterfly as a device, and the project overall, were also presented 6-monthly to the PPH Butterfly Advisory Group. This group comprised of four individuals from the local trust (consultant midwife, consultant obstetrician, head of risk management and a non-executive director), a commercial advisor (from Pelican Healthcare) and the Commercialisation Manager from the University of Liverpool. The group served to provide feedback and advice on a variety of aspects of the project from an outsider's perspective.

dwife, consultant obstetrician, head of risk management and a non-executive director), a commercial advisor (from Pelican Healthcare) and the Commercialisation Manager from the University of Liverpool. The group served to provide feedback and advice on a variety of aspects of the project from an outsider's perspective. Results The initial concept is loosely based on combining the shelf pessary and Cusco's vaginal speculum (figure 1). The vagina immediately after delivery is however not in its usual state as it has just been widely distended by the passage of the fetal head. The vagina is therefore relatively distended and loose immediately after childbirth, making it easier to admit a device.

e shelf pessary and Cusco's vaginal speculum (figure 1). The vagina immediately after delivery is however not in its usual state as it has just been widely distended by the passage of the fetal head. The vagina is therefore relatively distended and loose immediately after childbirth, making it easier to admit a device. To date, there have been eight major iterations of the device with subtle adjustments made at each iteration. The first model, made in the kitchen of one of the inventors (figure 4), was made from a plastic hand cream bottle with the side cut out with scissors and a paediatric asthma spacer device. A wire grill was fashioned to place over the cut-out section of the bottle, and the end of the spacer was inserted obliquely into its end. A series of coloured pin heads were inserted into the handle to represent a pressure gauge. The aim was to insert the device into the vagina in an anterioposterior aligned position before rotating it into position in the middle of the birth canal. Its primary function was diagnostic: to separate out blood coming from the uterus and the birth canal. Blood from the uterus would flow through the device and out of the end of the stalk while blood coming from vaginal lacerations would emerge from the birth canal around the sides of the stalk. Figure 4 First design using hand cream bottle and asthma spacer.

To date, there have been eight major iterations of the device with subtle adjustments made at each iteration. The first model, made in the kitchen of one of the inventors (figure 4), was made from a plastic hand cream bottle with the side cut out with scissors and a paediatric asthma spacer device. A wire grill was fashioned to place over the cut-out section of the bottle, and the end of the spacer was inserted obliquely into its end. A series of coloured pin heads were inserted into the handle to represent a pressure gauge. The aim was to insert the device into the vagina in an anterioposterior aligned position before rotating it into position in the middle of the birth canal. Its primary function was diagnostic: to separate out blood coming from the uterus and the birth canal. Blood from the uterus would flow through the device and out of the end of the stalk while blood coming from vaginal lacerations would emerge from the birth canal around the sides of the stalk. Figure 4 First design using hand cream bottle and asthma spacer. As discussions about the device progressed, it became clear that the BMC function was critical to its success. To be successful, the platform therefore needed to be much larger, similar to that of a ‘shelf pessary’, a gynaecological support device to support a prolapsed uterus in women who are unfit for surgery (figure 5). This would, however, make it much more difficult to insert and then to hold in a stable position during compression. The key design goals were to achieve: A platform that could allow atraumatic compression of the uterus while allowing drainage of blood through it.

d uterus in women who are unfit for surgery (figure 5). This would, however, make it much more difficult to insert and then to hold in a stable position during compression. The key design goals were to achieve: A platform that could allow atraumatic compression of the uterus while allowing drainage of blood through it. A system for inserting the platform into the upper vagina without trauma or discomfort and which can be reversed for its removal. A handle to allow the user to hold the platform from outside the birth canal and hold it stable despite up to 100 N pressure on any part of the platform. A safety mechanism to prevent inadvertent overinsertion of the device. A mechanism to assess how much blood is coming from above the device and how much is coming from below it. A mechanism which allows the device to be gripped securely against a surface under the woman's buttocks with minimal effort from the user. Figure 5 A shelf pessary.

A safety mechanism to prevent inadvertent overinsertion of the device. A mechanism to assess how much blood is coming from above the device and how much is coming from below it. A mechanism which allows the device to be gripped securely against a surface under the woman's buttocks with minimal effort from the user. Figure 5 A shelf pessary. In order to see the effects of each design change, numerous prototypes were made. Initially, the prototypes from hand drawn ideas were made from available materials or adapted from other instruments and machines. However, as the ideas became more developed, the Engineering department moved to computer-aided design (CAD) with 3D printing of APS prototypes. These served to give more of a realistic feel as to how the final device would look. Consequently, these prototype devices were used in a series of manikin studies that were undertaken as part of a student PhD. These studies provided information as to clinicians' opinion of the PPH Butterfly in comparison to traditional BMC. It also assessed the level of pressure that could be achieved over a 5-min period in a randomised trial of the two methods.

ed in a series of manikin studies that were undertaken as part of a student PhD. These studies provided information as to clinicians' opinion of the PPH Butterfly in comparison to traditional BMC. It also assessed the level of pressure that could be achieved over a 5-min period in a randomised trial of the two methods. In the final device, the platform shape and size is modelled on that of a large shelf pessary (figures 5 and 6). The addition of a handle enables the user to easily insert the device and align it into its correct position while the large handle limits the depth of insertion. When folded, the device can be slipped longitudinally into the vagina with minimal trauma. Once inserted, the device is unfolded by sliding the two arms of the device over each other so that the uterine platform pivots to face anteriorly towards the maternal abdomen (figure 7). Thus, the surface of the platform ends up perpendicular to the axis of the uterus. On sliding back one handle, the device will return to its original folded state, allowing the device to be removed without discomfort or trauma. Figure 6 The final design of the platform. Figure 7 The PPH Butterfly being inserted into the manikin (previous design of the device). PPH, postpartum haemorrhage.

In the final device, the platform shape and size is modelled on that of a large shelf pessary (figures 5 and 6). The addition of a handle enables the user to easily insert the device and align it into its correct position while the large handle limits the depth of insertion. When folded, the device can be slipped longitudinally into the vagina with minimal trauma. Once inserted, the device is unfolded by sliding the two arms of the device over each other so that the uterine platform pivots to face anteriorly towards the maternal abdomen (figure 7). Thus, the surface of the platform ends up perpendicular to the axis of the uterus. On sliding back one handle, the device will return to its original folded state, allowing the device to be removed without discomfort or trauma. Figure 6 The final design of the platform. Figure 7 The PPH Butterfly being inserted into the manikin (previous design of the device). PPH, postpartum haemorrhage. The PPH Butterfly is intended as a one-piece injection moulding with living hinges. It is moulded ‘open’ and folded into a stable isosceles triangle for use. When deployed, both handles come into locked apposition and are held locked with the grasp of the fingers. Plan dimensions for the PPH Butterfly triangle form are shown in figure 8. The platform height was originally developed to mimic the size and shape of a fist, although this will be in constant review during human testing. Figure 8 Design diagram showing original dimensions of the PPH Butterfly once folded for use. PPH, postpartum haemorrhage.

The PPH Butterfly is intended as a one-piece injection moulding with living hinges. It is moulded ‘open’ and folded into a stable isosceles triangle for use. When deployed, both handles come into locked apposition and are held locked with the grasp of the fingers. Plan dimensions for the PPH Butterfly triangle form are shown in figure 8. The platform height was originally developed to mimic the size and shape of a fist, although this will be in constant review during human testing. Figure 8 Design diagram showing original dimensions of the PPH Butterfly once folded for use. PPH, postpartum haemorrhage. The device was developed in accordance with the Medical Devices Directive standard BS EN ISO 14155:2011. There were several areas of the device that required particular attention and revisions. Stability of platform (locking) Initial designs of the device demonstrated difficulty with the strength and stability of the platform that is the ability of the platform to lock into place once in use. This design feature was key to the functioning and thus potential success of the device. There was also concern that, unless fully stable, the platform could suddenly dislodge/disengage and cause trauma to the vaginal, uterus or cervix. This was a key concern with the initial designs in which the platform was supported by a central handle (eg, figures 9 and 11). The answer to this problem came in the triangular design (figure 8), in which the two supporting arms are continuous with the platform and with the handles, which can be locked together.

rvix. This was a key concern with the initial designs in which the platform was supported by a central handle (eg, figures 9 and 11). The answer to this problem came in the triangular design (figure 8), in which the two supporting arms are continuous with the platform and with the handles, which can be locked together. Figure 9 The original design from which the name PPH Butterfly originated. The two wings on either side of the platform fold together for insertion, opening up once within the vagina. Channels on the surface of the wings direct blood down into the drainage channel that runs through the centre of the handle. PPH, postpartum haemorrhage. Figure 10 A variety of platform designs (in chronological order from left to right) showing the evolution of the design. Figure 11 This iteration has a compression gauge as part of the central pillar supporting the platform (note the draft pulley system, designed to pivot the platform from a longitudinal position for insertion into a horizontal position for use). Platform design The platform of the device acts as a surface against which the uterus can be compressed. It was estimated that it needed to be strong enough to support 10 kg of pressure. However, it also needed to have large perforations so as to allow blood and small clots to flow through it. If clots could not pass through the platform, then they could build up above it and give the impression that bleeding had stopped when in fact it was simply being held back inside the uterus.

t 10 kg of pressure. However, it also needed to have large perforations so as to allow blood and small clots to flow through it. If clots could not pass through the platform, then they could build up above it and give the impression that bleeding had stopped when in fact it was simply being held back inside the uterus. Initial designs had a central drainage channel that brought blood and clots through the platform and along the central handle (figure 9). There was concern however that a single channel could easily be blocked off by loose cervical tissue and so the decision was made to make numerous large perforations in a platform, large enough to allow small blood clots to pass but leaving enough residual material to support the compression. Consideration then turned to the platform design, attempting to produce a design that had holes large enough for small blood clots to pass through, but small enough to restrict the passage of the uterus or cervix itself. Any barrier also had to be atraumatic to the lower part of the uterus (ie, without a ‘cheese wire’ cutting effect) and not allow the passage of soft tissue through the perforations which could then swell and be trapped distal to the uterus. Ideally, it would also be aesthetically appealing. Several designs were put to the focus groups, including some where the perforations formed recurring patterns or pictures. An abstract design with variable sized perforations with a depth of at least 4 mm was finally settled on which remains in use in the first in vivo device (figure 6).

lso be aesthetically appealing. Several designs were put to the focus groups, including some where the perforations formed recurring patterns or pictures. An abstract design with variable sized perforations with a depth of at least 4 mm was finally settled on which remains in use in the first in vivo device (figure 6). The size of the platform was not easy to determine but was based on the dimensions of a flat hand and clenched fist as used in BMC. The birth canal immediately after childbirth has just been widely distended with the passage of a baby's head. This is typically 10 cm diameter at the time of birth, and the resulting distension of the birth canal is thought to remain for several hours after birth. It is this distension that gives space for the insertion of a clinician's flat hand into the vagina and space to clench it into a fist for BMC. These dimensions were therefore chosen for the initial in vivo version of the device. Subsequent testing of the device in five normal postnatal women revealed that the platform design was too large. For subsequent human tests, two sizes were therefore produced with reduced dimensions of 8.0×5.9 cm and 6.9×5.0 cm.

The size of the platform was not easy to determine but was based on the dimensions of a flat hand and clenched fist as used in BMC. The birth canal immediately after childbirth has just been widely distended with the passage of a baby's head. This is typically 10 cm diameter at the time of birth, and the resulting distension of the birth canal is thought to remain for several hours after birth. It is this distension that gives space for the insertion of a clinician's flat hand into the vagina and space to clench it into a fist for BMC. These dimensions were therefore chosen for the initial in vivo version of the device. Subsequent testing of the device in five normal postnatal women revealed that the platform design was too large. For subsequent human tests, two sizes were therefore produced with reduced dimensions of 8.0×5.9 cm and 6.9×5.0 cm. Living hinges To facilitate folding of the arms, the device has ‘living hinges’ at either side of the platform. This is a section of the device between the platform and the arms where the plastic suddenly narrows to 2 mm thickness, thus producing a natural point at which the device will crease. This allows the device to be moulded in a single part, reducing its complexity and price. Although a living hinge is strong, there is the potential for tearing after repeated folding. Initially, however, these hinges are strong and carry no risk of tearing during the single use that the device is made for.

will crease. This allows the device to be moulded in a single part, reducing its complexity and price. Although a living hinge is strong, there is the potential for tearing after repeated folding. Initially, however, these hinges are strong and carry no risk of tearing during the single use that the device is made for. The first injection-moulded prototype had a living hinge that extended right along both sides of the platform. Once folded, this formed a sharp ridge along the leading edge of the platform, which would be likely to cause trauma to the vaginal walls during insertion (figure 12). The platform edge was therefore redesigned to recess the living hinge behind two lateral pedicles (figure 13). This gave a smooth leading edge, but also gave the advantage of added stability once in the compression position. In the first injection mould, the total compression force was borne by the living hinges. In the redesign, however, the pedicles took the full force of the compression as soon as there was any give in the hinges. This gave greater stability and weight bearing ability. Figure 12 The initial hinge design—strong but with a sharp leading edge that could result in trauma on insertion. Figure 13 The redesigned shorter hinge. The sharp edges of the folded living hinge are now recessed behind extensions to the platform. Once folded for use, each arm has two pedicles lateral to the hinge that rest over the outer edges of the platform. As soon as the hinge starts to ‘give’ with pressure, these pedicles push tight against the platform thus ensuring strength and stability.

d living hinge are now recessed behind extensions to the platform. Once folded for use, each arm has two pedicles lateral to the hinge that rest over the outer edges of the platform. As soon as the hinge starts to ‘give’ with pressure, these pedicles push tight against the platform thus ensuring strength and stability. Minimising trauma In order to aid the ease of insertion and minimise any potential trauma to women, the device edges have been rounded as much as possible. A plastic was also selected for manufacture that had a smooth surface and could be lubricated with one of the water-soluble gels commonly used in hospitals. In the injection-moulded prototypes, flashes and joint ridges were smoothed out by hand prior to sterilisation and clinical use. The change in hinge design (see above) also minimised trauma.

for manufacture that had a smooth surface and could be lubricated with one of the water-soluble gels commonly used in hospitals. In the injection-moulded prototypes, flashes and joint ridges were smoothed out by hand prior to sterilisation and clinical use. The change in hinge design (see above) also minimised trauma. Preventing overinsertion During formal risk assessment, it was identified that there was a danger of inexperienced users inserting the device too far into the birth canal and so damaging the top of the vagina. It was therefore planned that a guard or shield would be added halfway up the handle of the PPH Butterfly to act as a safety mechanism by indicating to clinicians how far the device could be inserted safely. Its distance from the platform was based on the length of an adult Cusco's speculum, an instrument for inspection of the cervix which is in common hospital use. It was subsequently decided that a guard could cause confusion and that it would be better to simply increase the size of the handle of the device. This not only clearly indicates the limit of insertion, but it also prevents overinsertion through its large size (figures 7 and 9).

cervix which is in common hospital use. It was subsequently decided that a guard could cause confusion and that it would be better to simply increase the size of the handle of the device. This not only clearly indicates the limit of insertion, but it also prevents overinsertion through its large size (figures 7 and 9). Handle—inclusion of finger holes The Butterfly handle varied in material, shape and length over the course of the design process. Concerns were raised as to the difficulty and discomfort that would be experienced when having to grip the handles for the length of time that would be necessary to compress the uterus effectively. The solution came with the inclusion of five finger holes in a wide handle for the user's fingers to go through (figure 14). Placing the user's fingers through the holes not only held the two sides firmly together even in the presence of blood or other body fluids, but also gave a stable grasp for the user. The inclusion of the finger holes allows for variation in the way that the device can be held by the user: the device can either be held ‘in-line’ with the Butterfly stem in continuity with the forearm when in use or the hand can hold the handle from above. This allows the user to push the handle into the bed beneath the woman to hold it stable. This should provide a less tiring position for the user as his/her body weight can be used to simultaneously compress the uterus and hold the device in place.

ith the forearm when in use or the hand can hold the handle from above. This allows the user to push the handle into the bed beneath the woman to hold it stable. This should provide a less tiring position for the user as his/her body weight can be used to simultaneously compress the uterus and hold the device in place. Figure 14 (A–D) The final iteration of the device. The device is folded flat for insertion (A) and then the handles brought together so as to create a stable triangular structure to support the weigh-bearing platform (B). The handles are held together by complimentary protrusion and recesses in the handle as well as by the finger grips. The device can be held in a longitudinal fashion (C) to mimick standard bimanual compression. For women lying on a bed, the handle can be stabilised from above (D) so as to wedge the handle onto the bed. This mechanism should be less tiring for the clinician as he/she can use their body weight to put pressure on the uterus as well as to hold the device in position. Material choice A material had to be chosen that is suitable for internal use in women, while being robust, smooth and economical. Polypropylene was an obvious choice for this. The TOTAL product PPM H250 was chosen as it is ‘a narrow molecular weight distribution homopolymer with a Melt Flow Index of 25 g/10 min, recommended for the manufacture of medical packaging and medical devices, to the exclusion of implants’ (TOTAL, 2013).

h and economical. Polypropylene was an obvious choice for this. The TOTAL product PPM H250 was chosen as it is ‘a narrow molecular weight distribution homopolymer with a Melt Flow Index of 25 g/10 min, recommended for the manufacture of medical packaging and medical devices, to the exclusion of implants’ (TOTAL, 2013). Naming the device The ‘PPH Butterfly’ name was initially chosen because the earlier incarnations of the device had operated in such a way that when opened, the platform of the device resembled a pair of wings, like that of a butterfly. When the design changed to a platform with folding arms, it was felt this resemblance was lost, and a change in name to ‘PPH Shelf’ (after the shelf pessary on which the device is based) was considered. However, this option was not at all popular with the focus groups, and it was decided to retain the original name of ‘PPH Butterfly’.

changed to a platform with folding arms, it was felt this resemblance was lost, and a change in name to ‘PPH Shelf’ (after the shelf pessary on which the device is based) was considered. However, this option was not at all popular with the focus groups, and it was decided to retain the original name of ‘PPH Butterfly’. Blood collection The original concept of the device was that it would divide the genital tract into two and separate any blood coming from above the device from that coming from below. This can be seen in the very first home-made prototype (figure 4) where blood from above the device passed through the grill and down the handle. This was later modified so that the platform had channels along which blood flowed into a central stalk (figure 9). The realisation that a central hinge in the platform would be a point of weakness led to the production of a solid platform (figure 10). This would need to be twisted to allow its longitudinal insertion into the birth canal, complicating the drainage of blood. A suggestion came from a focus group to attach a blood collection bag to the rear of the platform. The aim of the bag was to collect any blood coming through the platform and give a physical recording of the amount of blood that had been lost while the device was in use. It would also have enabled a visual assessment and quantification of the volume and source of the blood loss. While an attractive option, the bag would have added complexity to the manufacture, and there were fears that it would be compressed in the birth canal and prevent blood entering it. Furthermore, an assessment of the source of the bleeding could be made without the bag: if the bleeding stopped as the uterus was compressed, then you would know the bleeding was uterine in origin. If it continued during uterine compression, then you would know that it was from vaginal lacerations. Consequently, it was agreed that the bag provided considerable complexity for little added benefit and it was removed.

eding stopped as the uterus was compressed, then you would know the bleeding was uterine in origin. If it continued during uterine compression, then you would know that it was from vaginal lacerations. Consequently, it was agreed that the bag provided considerable complexity for little added benefit and it was removed. Compression gauge Earlier in the design process consideration was given to the inclusion of a compression gauge that would provide feedback as to the level of pressure that was being used to compress the uterus. This can be seen in figure 11. This was to be sited along the handle and would have provided biofeedback to the user as to whether the pressure was adequate. However, when measurements were taken of the pressure exerted by 10 experts on the uterus of a manikin during BMC, the pressures were found to be highly variable. It was not therefore possible to provide a standard pressure, and the compression gauge was abandoned. Device assessment and risk management During initial BMC measurements in a manikin with prototype PPH Butterfly devices, peak levels of 42 N were recorded. Subsequent stress testing of real prototypes (PPH Model (2) A, B and C), as part of the Medicines and Healthcare Products Regulatory Agency (MHRA) submission, demonstrated devices could withstand forces of up to 902, 1140 and 1122 N, respectively, without buckling.

Device assessment and risk management During initial BMC measurements in a manikin with prototype PPH Butterfly devices, peak levels of 42 N were recorded. Subsequent stress testing of real prototypes (PPH Model (2) A, B and C), as part of the Medicines and Healthcare Products Regulatory Agency (MHRA) submission, demonstrated devices could withstand forces of up to 902, 1140 and 1122 N, respectively, without buckling. Feedback was obtained throughout the design period from clinicians who compared the PPH Butterfly device and BMC in an obstetric manikin. In initial studies, a commercial potato masher was used as a substitute for the device but this was replaced by 3D printed prototypes once available. A final cross-over study compared the uterine pressures generated by the device and bimanually in 20 doctors who had previous experience of undertaking BMC and 22 midwives who had no experience of it. The study measured maximum pressure generated as well as total pressure over a period of 5 min. There was no difference in the mean or maximum uterine pressures generated using BMC or the PPH Butterfly and no difference between the two groups. Discussion Now that the design of the PPH Butterfly is complete, the next step in the project is to test the device in postnatal humans—normal and in those with bleeding. Initial tests on normal postnatal bleeding will assess the size of the platform and handles as well as the mechanism for insertion. Once this has been explored, it will then be necessary to determine that the PPH Butterfly can effectively manage PPH.

est the device in postnatal humans—normal and in those with bleeding. Initial tests on normal postnatal bleeding will assess the size of the platform and handles as well as the mechanism for insertion. Once this has been explored, it will then be necessary to determine that the PPH Butterfly can effectively manage PPH. If successful, the PPH Butterfly has the potential to make a significant impact on PPH management. This would be particularly true within the developing world, where there is limited access to resources. The PPH Butterfly has been designed specifically with these countries in mind, with particular attention being paid to the overall cost of the device, which has been kept as low as possible, while not compromising the design. While it is designed to be single use and disposable in the UK, there are also plans to make a version that can be sterilised and re-used for low resource settings. This will allow it to reach a much wider number of women than would be possible if it only withstood one use. The use of BMC to treat PPH is limited by its invasive nature and how tiring it is for the clinician. These factors currently limit its use to those who are bleeding very heavily where its efficacy is likely to be the lowest. Consequently, its effectiveness is not demonstrated, which in turn leads to a further reluctance from clinical staff to use it as a method of PPH management.

nd how tiring it is for the clinician. These factors currently limit its use to those who are bleeding very heavily where its efficacy is likely to be the lowest. Consequently, its effectiveness is not demonstrated, which in turn leads to a further reluctance from clinical staff to use it as a method of PPH management. Suggestions have been made by the design team and members of the focus groups for ways in which the device could be enhanced, such as the inclusion of a light on the handle, and the addition of a soft covering or local haemostatics to the platform. While such additions may enhance the device, they are not necessary in relation to the basic function of the PPH Butterfly and instead could be seen as potential future ‘add-ons’.

ould be enhanced, such as the inclusion of a light on the handle, and the addition of a soft covering or local haemostatics to the platform. While such additions may enhance the device, they are not necessary in relation to the basic function of the PPH Butterfly and instead could be seen as potential future ‘add-ons’. Discussions with focus groups during the course of the project have provided the project team with real time feedback as to the design of the device and possible ideas for its development. Currently, the device needs to achieve basic functionality, while being safe and cost-effective. Once these aspects have been proven, it may be possible in the future for there to be some ‘add-ons’ to the device, at an additional cost. The focus groups facilitated the opportunity to view the device from a different perspective. Having focus groups comprised of clinicians and members of the public allowed the project team to see past their own enthusiastic viewpoint and also gave an appreciation of the way in which the PPH Butterfly may be seen by those women it could be used on. This was useful prior to beginning any recruitment to the human studies of the project, as it allowed the project team to carefully consider the best way in which to identify and approach women, as well as how to project the PPH Butterfly and the study itself, in order to increase the likelihood of participation.

d on. This was useful prior to beginning any recruitment to the human studies of the project, as it allowed the project team to carefully consider the best way in which to identify and approach women, as well as how to project the PPH Butterfly and the study itself, in order to increase the likelihood of participation. Conclusions The PPH Butterfly is a novel device designed to treat PPH through uterine compression. It is built to be less invasive and less tiring than traditional BMC. Clinical assessments are currently being undertaken to determine its safety and effectiveness in PPH management. Correction notice: This article has been corrected since it was published Online First. Figures 9-13 were published in the wrong order. This has now been amended. The authors thank Phil Little (PL) (Plastribution), Daryl Parker (DP) (Protolabs), Richard Carter (RC) (Pelican Feminine Healthcare ), Colin Kent (CK) (Pelican Healthcare) and Simon Collins (SC) (Interplex). Tony Fisher (TF), Head of Department of Medical Physics and Clinical Engineering at the Royal Liverpool University Teaching Hospital, provided important advice about the Medicines & Healthcare Products Regulatory Agency application process and fulfilment of criteria, as well as agreeing to develop the PPH Butterfly prototype in his department at a discounted cost. The authors also thank Tristan Payne and Victoria Kearns for their assistance with the MHRA process. Twitter: Follow Andrew Weeks @adweeks

The authors thank Phil Little (PL) (Plastribution), Daryl Parker (DP) (Protolabs), Richard Carter (RC) (Pelican Feminine Healthcare ), Colin Kent (CK) (Pelican Healthcare) and Simon Collins (SC) (Interplex). Tony Fisher (TF), Head of Department of Medical Physics and Clinical Engineering at the Royal Liverpool University Teaching Hospital, provided important advice about the Medicines & Healthcare Products Regulatory Agency application process and fulfilment of criteria, as well as agreeing to develop the PPH Butterfly prototype in his department at a discounted cost. The authors also thank Tristan Payne and Victoria Kearns for their assistance with the MHRA process. Twitter: Follow Andrew Weeks @adweeks Contributors: The initial PPH Butterfly concept and design was developed by AW, PW and JP. Subsequent design meetings were attended by CC and DL. CC wrote the first draft of the manuscript, which was subsequently revised by AW. All authors approved the final version before submission. Funding: The very first designs of the device were not externally funded. Subsequent funding was received from the National Institute for Health Research (NIHR) Invention for Innovation funding stream II-LA-0712-20007. This funding was over a period of 36 months. Additional funding came from the Sir Halley Stewart Trust for the manikin and from the University of Libya for NA PhD. Additional support came from the Department of Medical Physics and Clinical Engineering who offered their services at reduced cost.

stream II-LA-0712-20007. This funding was over a period of 36 months. Additional funding came from the Sir Halley Stewart Trust for the manikin and from the University of Libya for NA PhD. Additional support came from the Department of Medical Physics and Clinical Engineering who offered their services at reduced cost. Disclaimer: The views expressed in this paper are those of the author(s) and not necessarily those of the NHS, the NIHR or the Department of Health. Competing interests: AW and JP are named as inventors of the device and would receive a proportion of any loyalties generated. Ethics approval: Development of the PPH Butterfly did not involve humans and so formal ethical approval was not required. Provenance and peer review: Not commissioned; externally peer reviewed. Data sharing statement: No additional data are available from the study, other than the device technical file which is currently held by the University of Liverpool.

Introduction The new sustainable development agenda aims to ensure that the momentum generated by the millennium development goals is continued beyond 2015. India has achieved success in maternal health but is off-track on child health. More vigorous efforts will be needed to meet the new global targets of zero preventable child deaths and a much sharper reduction in maternal deaths by 2030.1 Innovation will be central to achieving these goals in maternal and child health; it could be a product, process or healthcare system innovation.2 While the unmet health burden will definitely need action on known determinants such as shortage/maldistribution of health workers, coverage of health services, quality of care (QoC) and better monitoring of programmes, ‘disruptive innovations’ in healthcare will also be critical, especially for those at the bottom of the pyramid and traditionally left out of the healthcare net owing to political, economic or geographic reasons.3

of health workers, coverage of health services, quality of care (QoC) and better monitoring of programmes, ‘disruptive innovations’ in healthcare will also be critical, especially for those at the bottom of the pyramid and traditionally left out of the healthcare net owing to political, economic or geographic reasons.3 The disease burden among mothers and neonates is mostly caused by a handful of avoidable conditions—haemorrhage, gestational hypertension, infections and obstructed labour in mothers4; and prematurity/low birth weight, infections and birth  asphyxia/birth trauma in newborn babies.5 They occur mostly around the time of childbirth and within the first month following delivery. Though the direction of causality is unclear, neonatal hypothermia is associated with all of these three conditions. The prevalence of hypothermia has been found to range from 11% to 92%, and case fatality rates for newborn hypothermia are seen to vary from 9% to 52%.6 Currently, in most parts of the world, newborn temperature measurement and documentation are incomplete, resulting in an incomplete understanding of the epidemiology of hypothermia and its clinical consequences.6 Remote, real-time monitoring of key physiological parameters,7 such as temperature of mothers and newborns, is a promising pathway for the early detection of complications presenting as hypothermia or as fever in mothers/newborns, offering a potential opportunity to impact access as well as QoC for these vulnerable populations. In the development of health technology for surveillance, ‘frugal innovation’ using social innovation design principles such as safety and efficacy; robustness; independence of electricity mains or replaceable batteries; operational capability without needing replacement consumables and simplicity of operation provides an opportunity for the design and deployment of healthcare systems that address access and quality concerns.8 Here we describe the design, development and testing of a wearable sensor device for remote biomonitoring (RBM) of body temperatures in mothers and newborns in southern India.

d simplicity of operation provides an opportunity for the design and deployment of healthcare systems that address access and quality concerns.8 Here we describe the design, development and testing of a wearable sensor device for remote biomonitoring (RBM) of body temperatures in mothers and newborns in southern India. Methods Study setting St John’s Medical College Hospital, Bangalore, is a 1300-bedded, tertiary-care hospital with 2500 deliveries per annum. It has a Level 3 nursery and takes care of 1000 neonates (inborn:outborn ratio=2:1) in the neonatal intensive care unit (NICU) per year, with survival rates of 99% at 48 hours. About one-third of newborns are low birth weight and one-fifth are preterm. Background rates of maternal mortality ratio and infant mortality ratio were 1.33 and 31 per 1000 live-births for Karnataka state.9 Choice of site of temperature measurement in adult and newborn In the adult, the upper arm was selected for continuous monitoring. While the axillary or other sites (rectal/tympanic/forehead) are recommended for episodic measurement, abdominal skin temperature, despite it being subject to the vasomotor activity of the skin, is appropriate for continuous monitoring (figure 1A,B). In addition to being close to the liver, a metabolically active organ facilitating a measurement close to the core temperature,10 it is also a non-invasive method that is steady, continuous, easy-to-use and comfortable for the infant.11 The temperature ranges of normal and abnormal (hypothermia or hyperthermia) are shown in figure 2.

being close to the liver, a metabolically active organ facilitating a measurement close to the core temperature,10 it is also a non-invasive method that is steady, continuous, easy-to-use and comfortable for the infant.11 The temperature ranges of normal and abnormal (hypothermia or hyperthermia) are shown in figure 2. Figure 1 Positioning of the wearable remote biomonitoring sensor for temperature surveillance in (A) mother (upper arm) and (B) newborn (epigastrium). Figure 2 Temperature cut-off for neonates (with alerts beyond the normal range of 36.5–37.5°C). Innovation pathway We adopted the USAID 2006 Innovation Pathway model that incorporated four steps from research to field implementation: (1) priority-setting phase and product design, (2) product development and proof of concept, (3) product introduction and (4) field implementation.12 As part of step 1, we reviewed the substeps of (1) problem identification (epidemiological/technological/social/financial), (2) critical review of temperature measurement devices and other issues (temperature sensors or thermistors, core body temperature measurement issues) and (3) determination of niche of this device vis-à-vis other devices through horizon scanning.

problem identification (epidemiological/technological/social/financial), (2) critical review of temperature measurement devices and other issues (temperature sensors or thermistors, core body temperature measurement issues) and (3) determination of niche of this device vis-à-vis other devices through horizon scanning. Epidemiological burden of chief causes of maternal and neonatal mortality and morbidity were reviewed as a first step.4 5 A technological review revealed that the device to be strapped onto the mother or newborn for vital sign detection had to be a medical-grade device with extremely high safety profile. Further, it had to remain functional in situations such as soiling or wetting by the newborn, and also continue to work in different geographic areas with varying population densities as well as buildings. Several devices such as the ThermoSpot, a temperature sensor tag and remote infrared-based instruments were evaluated. Social review revealed that a sensor being continually strapped on to the body for long periods was not dissimilar to that of talismans strapped around the arms of adults or around the necks/waists of children and might therefore not be too disagreeable to families. A financial review revealed that for the device to be used widely, especially by low-income populations, it would have to be a low-cost device. Implementation review focused on possible uptake and utilisation of such a device in rural, urban and slum communities. The focus of this paper is to report on the clinical evaluation of the product as part of step 2 of the innovation pathway.

especially by low-income populations, it would have to be a low-cost device. Implementation review focused on possible uptake and utilisation of such a device in rural, urban and slum communities. The focus of this paper is to report on the clinical evaluation of the product as part of step 2 of the innovation pathway. Conceptual framework The conceptual framework for remote monitoring is illustrated in figure 3. The innovation framework consisted of five components: (1) a low-cost, wearable sensor tag; (2) a gateway device acting as an ultra-low-power ‘real-time’ communication link; (3) piggy-backing on a commercial cellular communication network; (4) smart data analytics system and (5) feedback loop to the caregiver or front-line healthworker. This framework was to be ideated, designed, prototyped and developed into a Class A medical device with lowest risk level as per Medical Device Regulatory Authority of India13 bearing in mind that the end product should enable tenets of good QoC, namely effectiveness, timeliness, safety, people-centredness and equity.14 Here we describe the functioning of the first two of five components. Figure 3 Conceptual framework for remote biomonitoring for temperature surveillance.

Conceptual framework The conceptual framework for remote monitoring is illustrated in figure 3. The innovation framework consisted of five components: (1) a low-cost, wearable sensor tag; (2) a gateway device acting as an ultra-low-power ‘real-time’ communication link; (3) piggy-backing on a commercial cellular communication network; (4) smart data analytics system and (5) feedback loop to the caregiver or front-line healthworker. This framework was to be ideated, designed, prototyped and developed into a Class A medical device with lowest risk level as per Medical Device Regulatory Authority of India13 bearing in mind that the end product should enable tenets of good QoC, namely effectiveness, timeliness, safety, people-centredness and equity.14 Here we describe the functioning of the first two of five components. Figure 3 Conceptual framework for remote biomonitoring for temperature surveillance. Requirements for components Two key requirements for the ‘on-body’ sensor were safety and accuracy. Given the fragility of the newborn skin,15 the device had to be hypoallergenic,16 burst/leak proof,17 cause minimal infections and dissipate minimal heat or ultra-low-power, non-ionising electromagnetic radiation.18 The adhesive used to secure the device to the skin should similarly cause minimal ‘medical adhesive-related skin injury’,19 allergy or infections. Device accuracy was targeted to be ±0.2°C in in vitro conditions and ±0.5°C in actual clinical practice. Other mechanical requirements for the device were long battery life up to 28 days (with sampling frequency of 5 min); robustness (without any malfunction on coming into contact with body fluids like sweat, blood, urine, faeces; at least ‘ingress protection class 67’ (dustproof and waterproof); human-centric and aesthetic design for non-intrusiveness over prolonged use; and ability to withstand mild shock or vibration; and that the device should not get reset accidentally (online supplementary figure 1). There was also a requirement for the device to store data locally and communicate with a gateway device for onward transmission of data via the configured cellular network.

onged use; and ability to withstand mild shock or vibration; and that the device should not get reset accidentally (online supplementary figure 1). There was also a requirement for the device to store data locally and communicate with a gateway device for onward transmission of data via the configured cellular network. 10.1136/bmjinnov-2016-000153.supp1Supplementary file 1 Product design and engineering Based on the clinical requirements, a preliminary design was constituted and subsequent design choices underwent multiple iterations driven by technological capabilities and user reviews. The wearable sensor enclosures were pebble-shaped or coin-stack-shaped with all the electronics embedded inside (online supplementary figure 1; figure 4A,B). A battery of tests (both mechanical and electrical) were undertaken to confirm device performance, robustness and reliability. After several rounds of pre-testing, design optimisation was achieved. Figure 4 (A) Disassembled prototype; (B) vertical cross-sectional view of enclosure.

Product design and engineering Based on the clinical requirements, a preliminary design was constituted and subsequent design choices underwent multiple iterations driven by technological capabilities and user reviews. The wearable sensor enclosures were pebble-shaped or coin-stack-shaped with all the electronics embedded inside (online supplementary figure 1; figure 4A,B). A battery of tests (both mechanical and electrical) were undertaken to confirm device performance, robustness and reliability. After several rounds of pre-testing, design optimisation was achieved. Figure 4 (A) Disassembled prototype; (B) vertical cross-sectional view of enclosure. Prototype and implementation The details of the prototype sensor device are given elsewhere.20 Briefly, the sensor hardware platform, developed using a multilayer printed circuit board, consisted of a microcontroller with integrated Bluetooth V.4.0 Low Energy, a 12-bit ADC (CC2540 from Texas Instruments), a NICU-grade temperature sensor with its analog front-end circuit, status LEDs, power supply and RF balun filter and antenna for wireless communication over 2.4 GHz ISM band. High-precision MF51E NTC thermistors (Cantherm, 2009)i were used for extremely accurate temperature measurements. An embedded Inverted F-antenna with higher efficiency, longer range and a wider bandwidth than a chip antenna was used. It also had a very low-tolerance resistor enabling a low-power consumption during both active (150 μA) and sleep (1 μA) modes. The sensor hardware was programmable as per requirements. A 3 V coin battery powered the device. A baby-friendly enclosure was made from medical-grade hypoallergenic plastics (figure 4A,B; online supplementary figures 1 and 2).

esistor enabling a low-power consumption during both active (150 μA) and sleep (1 μA) modes. The sensor hardware was programmable as per requirements. A 3 V coin battery powered the device. A baby-friendly enclosure was made from medical-grade hypoallergenic plastics (figure 4A,B; online supplementary figures 1 and 2). The sensor communicated with a gateway device (a smartphone or a Raspberry pi) that could subsequently relay the temperature data over a secured internet backbone provided by GPRS/Wi-Fi onto a centralised database for storage (online supplementary figure 3). Phases of device testing The planned phases of device testing were pre-clinical testing (in June 2013) and clinical trial phases I and IIa for the evaluation of safety and efficacy in the following sequence: healthy adult volunteers (May 2014); healthy mothers (January to February 2015); healthy babies (February 2015); stable babies in the NICU (February to March 2015) and babies with morbidities such as hypoxic ischaemic encephalopathy (HIE) (March 2015). The results of the pre-clinical testing in the laboratory setting were published earlier.20 Briefly, the response-time of the sensor device to attain thermal equilibrium with the surroundings was 4 min compared with 3 min observed with a precision-grade digital thermometer used as a reference standard. In terms of accuracy, the error was calculated to be within 0.1°C of the reference standard while using waterbaths in the temperature range of 25–40°C (online supplementary figure 4). The details of the clinical phase of testing are outlined below.

ith a precision-grade digital thermometer used as a reference standard. In terms of accuracy, the error was calculated to be within 0.1°C of the reference standard while using waterbaths in the temperature range of 25–40°C (online supplementary figure 4). The details of the clinical phase of testing are outlined below. Seven free-living healthy adult volunteers (males=2; females=5), with no known morbidities, were the first-phase participants over a 7-day period. All of them had the devices strapped with an armband secured with Velcro tape onto their left upper arms (figure 1A) and were invited to contribute at least a minimum of 24 hours of observations accumulated from over one or more days and report any adverse events or side effects they experienced. In parallel, they also noted down timed axillary temperature readings (at least five times in a 24-hour window) using a digital thermometer for paired comparisons. Testing in healthy postnatal mothers (n=11) was first carried out among those with well babies in the postnatal ward and then among mothers (n=7) with neonates admitted in the NICU. The devices were secured onto their upper arms with armband secured with Velcro tape. Paired readings taken every 5 min over a 1-hour period with an axillary digital thermometer were compared against the readings of the sensor with its time stamp.

al ward and then among mothers (n=7) with neonates admitted in the NICU. The devices were secured onto their upper arms with armband secured with Velcro tape. Paired readings taken every 5 min over a 1-hour period with an axillary digital thermometer were compared against the readings of the sensor with its time stamp. Testing among neonates was carried out in three different phases in all of which the devices were secured with cotton and micropore adhesive to the upper epigastrium. In the first phase, well babies (n=3) in their early neonatal periods in the postnatal ward had their axillary temperature readings taken every 5 min over a 1-hour period with a digital thermometer and compared against the sensor readings. In the second phase, sick but stable neonates (n=10) from the NICU were recruited. They were under the radiant warmer with the temperature probe fixed onto the upper abdomen beside the sensor device and so the readings taken every 5 min over a 1-hour period from the warmer panel were compared against the sensor readings. But since all these babies had their temperatures maintained within a narrow normal range under the radiant warmer, we included one sick infant with HIE due to birth asphyxia and who was on treatment with therapeutic whole-body cooling. This facilitated comparison of readings in the range of 33–34°C during the cooling and rewarming phases of the treatment. The radiant warmer probe readings taken every 5 min over a 1-hour period from the warmer panel were compared against the sensor readings.

nd who was on treatment with therapeutic whole-body cooling. This facilitated comparison of readings in the range of 33–34°C during the cooling and rewarming phases of the treatment. The radiant warmer probe readings taken every 5 min over a 1-hour period from the warmer panel were compared against the sensor readings. Definitions Safety We instituted an adverse event reporting and resolution protocol for the wearable sensor devices. This enabled capture of the number and severity of adverse events as well as the individual clinical management as also feedback for changes into device design. Adverse events to be recorded were dermatitis, infection, thermal injury, radiation injury, device leak/burst and others. Accuracy Accuracy of the device was estimated by comparing temperatures recorded by the device against other measurements routinely used in clinical practice. For mothers and well babies in postnatal wards, the comparisons were between the device temperatures versus axillary temperatures read from a digital thermometer (once readings stabilised after the beep—usually after 3 min); for newborns in NICU, the comparisons were between the time-stamped device temperatures versus skin probe temperatures obtained from the control panels of calibrated radiant warmers (Phoenix Medical Systems Private Limited, Chennai, India, or Zeal Medical, Mumbai, India). Statistical analysis Mean (±SD) was calculated for the paired sets of readings noted in mothers and newborns and the mean differences were obtained. Paired t-test was used for testing of significance between two different methods.

Accuracy Accuracy of the device was estimated by comparing temperatures recorded by the device against other measurements routinely used in clinical practice. For mothers and well babies in postnatal wards, the comparisons were between the device temperatures versus axillary temperatures read from a digital thermometer (once readings stabilised after the beep—usually after 3 min); for newborns in NICU, the comparisons were between the time-stamped device temperatures versus skin probe temperatures obtained from the control panels of calibrated radiant warmers (Phoenix Medical Systems Private Limited, Chennai, India, or Zeal Medical, Mumbai, India). Statistical analysis Mean (±SD) was calculated for the paired sets of readings noted in mothers and newborns and the mean differences were obtained. Paired t-test was used for testing of significance between two different methods. Results The seven adult volunteers, aged 20–45 years, contributed a total of 345 hours of readings (range=25–111 hours) over the 7-day period. No major adverse events were noted; four of seven participants noted a minor adverse event of sweating under the device/armband and one participant developed mild skin allergy with copper coating of the sensor.

s, aged 20–45 years, contributed a total of 345 hours of readings (range=25–111 hours) over the 7-day period. No major adverse events were noted; four of seven participants noted a minor adverse event of sweating under the device/armband and one participant developed mild skin allergy with copper coating of the sensor. The first set of 11 postnatal mothers provided a total of 312 paired readings. Ten of them had their temperatures measured during the mid-day (10:30 to 15:30) and one mother had her temperatures recorded in the early morning (between 5:00 and 6:30). The second set of seven postnatal mothers provided a total of 91 paired readings during the mid-day period (10:30 to 13:30). No adverse events were noted in any of the mothers. The mean skin temperature measured by the sensor was 2°C lower than the axillary temperature readings (sensor=34.1°C vs digital=36.1°C) in the volunteers as well as both sets of mothers and this difference was statistically significant (t-test=13.8; P<0.001) (table 1) (online supplementary figure 5). Table 1 Results of remote biomonitoring device testing of upper arm skin temperature measurement against axillary digital thermometer measurement in postnatal mothers

The first set of 11 postnatal mothers provided a total of 312 paired readings. Ten of them had their temperatures measured during the mid-day (10:30 to 15:30) and one mother had her temperatures recorded in the early morning (between 5:00 and 6:30). The second set of seven postnatal mothers provided a total of 91 paired readings during the mid-day period (10:30 to 13:30). No adverse events were noted in any of the mothers. The mean skin temperature measured by the sensor was 2°C lower than the axillary temperature readings (sensor=34.1°C vs digital=36.1°C) in the volunteers as well as both sets of mothers and this difference was statistically significant (t-test=13.8; P<0.001) (table 1) (online supplementary figure 5). Table 1 Results of remote biomonitoring device testing of upper arm skin temperature measurement against axillary digital thermometer measurement in postnatal mothers Measurements: Postnatal mothers’ categories Time of temperature recording (hh:mm) Readings (n) Temperature Mean (±SD) Temperature Mean difference (first reading) (last reading) Device (upper arm) Digital (axilla) (device-digital) Mothers of well babies 1 11:17 12:17 13 35.34 36.8 −1.4 2 11:02 12:02 13 35.15 36.7 −1.5 3 11:03 12:03 13 33.69 36.0 −2.3 4 11:18 12:18 13 34.02 36.4 −2.4 5 10:36 11:36 13 33.18 36.1 −2.9 6 11:50 12:50 13 34.21 36.2 −1.9 7 11:31 12:31 13 33.75 35.8 −2.1 8 11:38 12:38 13 33.31 36.1 −2.8 9 12:32 13:32 13 33.88 35.7 −1.8 10 14:19 15:19 13 35.69 36.4 −0.7 11 05:08 06:18 4 33.00 35.5 −2.5 Subgroup – – 312 34.1 (±0.9) 36.1 (±0.4) −2.0*

6.0 −2.3 4 11:18 12:18 13 34.02 36.4 −2.4 5 10:36 11:36 13 33.18 36.1 −2.9 6 11:50 12:50 13 34.21 36.2 −1.9 7 11:31 12:31 13 33.75 35.8 −2.1 8 11:38 12:38 13 33.31 36.1 −2.8 9 12:32 13:32 13 33.88 35.7 −1.8 10 14:19 15:19 13 35.69 36.4 −0.7 11 05:08 06:18 4 33.00 35.5 −2.5 Subgroup – – 312 34.1 (±0.9) 36.1 (±0.4) −2.0* Mothers of sick neonates 12 12:27 13:27 13 32.85 35.3 −2.4 13 11:08 12:08 13 35.29 36.5 −1.2 14 11:30 12:30 13 33.35 36.4 −3.1 15 14:18 15:18 13 34.16 36.4 −2.2 16 10:25 11:25 13 33.39 35.9 −2.5 17 11:41 12:41 13 35.16 36.4 −1.3 18 11:16 12:16 13 33.97 36.1 −2.2 Subgroup – – 91 34.0 (±0.9) 36.1 (±0.4) −2.1† Overall – – 403 34.1 (±0.9) 36.1 (±0.4) −2.0‡ All  P values < 0.001. *t-test=10.5; df=10. †t-test=8.3; df=6. ‡t-test=13.8; df=17. The first set of three neonates provided a total of 39 paired readings during the early morning time window (4:00 to 5:00). The mean difference in temperature was 0.14°C (sensor=36.87°C vs digital=36.73°C; P=0.2). The second set of 10 neonates in the NICU provided a total of 130 paired readings. Their mean skin temperature measured by the sensor was 0.6°C lower than that measured by the radiant warmer probe (sensor=35.88°C vs warmer probe=36.46°C; P<0.001). The last phase study neonate in the NICU provided a total of 25 paired readings with the mean sensor reading being not different from the radian warmer probe reading (sensor=33.53°C vs warmer probe=33.54°C; P=0.8) (table 2). No adverse events were noted in any of the babies.

r=35.88°C vs warmer probe=36.46°C; P<0.001). The last phase study neonate in the NICU provided a total of 25 paired readings with the mean sensor reading being not different from the radian warmer probe reading (sensor=33.53°C vs warmer probe=33.54°C; P=0.8) (table 2). No adverse events were noted in any of the babies. Table 2 Results of remote biomonitoring device testing of newborn baby temperature measurement in different temperature ranges by different groups of neonates Measurements: Neonates’ categories Time of temperature recording (hh:mm) Readings (n) Temperature Mean (±SD) Temperature Mean difference (first reading) (last reading) Device (abdomen) Digital (axilla) (device-digital) Well babies (in postnatal ward) 1 03:54 04: 54 13 36.61 36.6 0.01 2 04:02 05: 02 13 37.00 36.8 0.18 3 04:05 05: 05 13 37.00 36.8 0.21 Subgroup – – 39 36.87 (±0.2) 36.73 (±0.1) 0.13*

s (n) Temperature Mean (±SD) Temperature Mean difference (first reading) (last reading) Device (abdomen) Digital (axilla) (device-digital) Well babies (in postnatal ward) 1 03:54 04: 54 13 36.61 36.6 0.01 2 04:02 05: 02 13 37.00 36.8 0.18 3 04:05 05: 05 13 37.00 36.8 0.21 Subgroup – – 39 36.87 (±0.2) 36.73 (±0.1) 0.13* Device (abdomen) Probe (abdomen) (device-probe) Sick neonates in NICU under radiant warmer with temperature >35.5°C 4 15:43 16:43 13 36.62 36.57 0.05 5 11:16 12:16 13 35.78 36.56 −0.78 6 12:47 13:47 13 35.66 36.34 −0.68 7 11:56 12:56 13 35.61 36.64 −1.03 8 14:26 15:26 13 35.50 36.35 −0.85 9 11:25 12:25 13 35.41 36.32 −0.91 10 14:00 15:00 13 35.76 36.43 −0.67 11 11:04 12:04 13 36.71 36.45 0.26 12 12:17 13:17 13 35.84 36.43 −0.59 13 12:39 13:39 13 35.89 36.47 −0.58 Subgroup – – 130 35.88 (±0.4) 36.46 (±0.1) −0.6† Sick neonate in NICU Undergoing therapeutic cooling for birth asphyxia (temperature approximately 33°C) Neonate 14 02:51 03:51 13 33.10 33.15 0.05 Undergoing rewarming after therapeutic cooling for birth asphyxia (temperature approximately 34°C) Neonate 14 12 33.95 33.92 0.03 Subgroup – – 25 33.53 (±-0.6) 33.54 (±0.5) 0.01‡ Overall – – 155 35.76 (±1.1) 36.12 (±1.1) −0.4§ *t=2.2; df=2; P=0.2. †t=4.4; df=9; P<0.001. ‡t=0.25; df=1; P=0.8. §t=3.0; df=14; P<0.01. NICU, neonatal intensive care unit.

Device (abdomen) Probe (abdomen) (device-probe) Sick neonates in NICU under radiant warmer with temperature >35.5°C 4 15:43 16:43 13 36.62 36.57 0.05 5 11:16 12:16 13 35.78 36.56 −0.78 6 12:47 13:47 13 35.66 36.34 −0.68 7 11:56 12:56 13 35.61 36.64 −1.03 8 14:26 15:26 13 35.50 36.35 −0.85 9 11:25 12:25 13 35.41 36.32 −0.91 10 14:00 15:00 13 35.76 36.43 −0.67 11 11:04 12:04 13 36.71 36.45 0.26 12 12:17 13:17 13 35.84 36.43 −0.59 13 12:39 13:39 13 35.89 36.47 −0.58 Subgroup – – 130 35.88 (±0.4) 36.46 (±0.1) −0.6† Sick neonate in NICU Undergoing therapeutic cooling for birth asphyxia (temperature approximately 33°C) Neonate 14 02:51 03:51 13 33.10 33.15 0.05 Undergoing rewarming after therapeutic cooling for birth asphyxia (temperature approximately 34°C) Neonate 14 12 33.95 33.92 0.03 Subgroup – – 25 33.53 (±-0.6) 33.54 (±0.5) 0.01‡ Overall – – 155 35.76 (±1.1) 36.12 (±1.1) −0.4§ *t=2.2; df=2; P=0.2. †t=4.4; df=9; P<0.001. ‡t=0.25; df=1; P=0.8. §t=3.0; df=14; P<0.01. NICU, neonatal intensive care unit. Discussion Our RBM device, a battery-operated, wearable temperature sensor tag designed and developed based on two sets of criteria, namely client needs and technological requirements, was within an enclosure package that took into account design elements and incorporated hermetic sealing with no sharp corners/crevices for accumulation of bacteria/dust and designed to be easily sterilisable. It is a product built on a ‘needs-based biodesign innovation’21 and qualifies as a breakthrough product that could potentially offer a clinically meaningful advantage for life-threatening conditions in resource-poor settings where no alternate means of diagnosis exists for mothers and newborns. The biomonitoring sensor satisfies the definition of an innovation with characteristics of novelty and application with a clearly intended benefit.22

y offer a clinically meaningful advantage for life-threatening conditions in resource-poor settings where no alternate means of diagnosis exists for mothers and newborns. The biomonitoring sensor satisfies the definition of an innovation with characteristics of novelty and application with a clearly intended benefit.22 In our study, skin temperatures in mothers were lower than the axillary temperatures by 2°C; this expected difference can be used to calibrate the device readings in future when used for temperature surveillance in adults. In newborns, it had a precision of ±0.5°C relative to axillary measurements at 15 min intervals and was comfortable to wear. Our innovation was similar to other products such as the unobtrusive continuous temperature monitoring of infants in NICUs reported elsewhere.10 Differences in the reporting of sweating under the device/band by adult volunteers and mothers were possibly due to differences in environmental temperatures during different testing seasons. Allergy reports led to design modifications with a gold/stainless steel plating over the sensor surface.

s reported elsewhere.10 Differences in the reporting of sweating under the device/band by adult volunteers and mothers were possibly due to differences in environmental temperatures during different testing seasons. Allergy reports led to design modifications with a gold/stainless steel plating over the sensor surface. Our current round of testing is not without limitations. Our preliminary round of testing has revealed the health technology to be safe (in the short term) and accurate for short duration testing. The next step would be to test the sensor and other technologies for longer durations (days to weeks) in both mothers and newborns. Further, conversion of the sensor into a multiparameter detection device is likely to render it more useful in clinical and public health settings as also improving the cost-effectiveness of the device. Another major limitation of our current approach was the use of trial-and-error methods in testing which resulted in cost and time over-runs. In the next rounds of testing, more rigorous approaches may be employed across the phases of design, development and testing.23 24 Though our device has been found to be safe in the short term, additional in vitro and in vivo safety testing must continue on the long-term safety of early neonatal exposure to contact devices as well as low-dose radiofrequency radiation.25

hes may be employed across the phases of design, development and testing.23 24 Though our device has been found to be safe in the short term, additional in vitro and in vivo safety testing must continue on the long-term safety of early neonatal exposure to contact devices as well as low-dose radiofrequency radiation.25 The sensor device will also have to be tested for its ability to not only ‘sense’ data on neonatal temperatures but also to act as a feedback tool for thermoprotective measures. It could also be used to collect information on possible risk factors such as maternal temperature, environmental conditions and sociocultural practices and then based on back-end analytics using predefined algorithms, specific messages could be sent to hospitals, families or community health workers for preventive/corrective measures including behaviour change or referral of sick newborns (figures 2 and 3 and online supplementary figure 3).6 In summary, this proof-of-concept study shows this early-stage innovation26 to be promising in terms of safety and accuracy (with appropriate calibration) for monitoring of temperatures in adults and newborns that can potentially be carried forward to the next stages of device accuracy refinement as well as further phases of clinical research for improved maternal and neonatal health. Investing in such innovations would be critical to achieving maternal and neonatal health goals over the next decade.1 6 27 i Cantherm (2009). High precision NTC thermistors. http://www.cantherm.com/products/thermistors/mf51e.html/webfigures_20160812.pdf.

In summary, this proof-of-concept study shows this early-stage innovation26 to be promising in terms of safety and accuracy (with appropriate calibration) for monitoring of temperatures in adults and newborns that can potentially be carried forward to the next stages of device accuracy refinement as well as further phases of clinical research for improved maternal and neonatal health. Investing in such innovations would be critical to achieving maternal and neonatal health goals over the next decade.1 6 27 i Cantherm (2009). High precision NTC thermistors. http://www.cantherm.com/products/thermistors/mf51e.html/webfigures_20160812.pdf. Contributors: PKM, PT and BA conceived and designed the testing programme. PKM, PT, SB, SR, MW, SA, AT, SCN, HR and BA designed the protocol. SB, SR, AT and SCN oversaw the implementation of the programme. MW and SA managed the day-to-day operations of the programme. HR provided technological support. PKM drafted the manuscript. All the authors edited and revised the draft and approved the final version. PKM and BA are the guarantors. Funding: This work was supported by an intramural research grant from the St John’s Medical College Research Society, Bangalore, India, and a development grant from the Robert Bosch Centre for Cyber Physical Systems, Indian Institute of Science, Bangalore, India.

Contributors: PKM, PT and BA conceived and designed the testing programme. PKM, PT, SB, SR, MW, SA, AT, SCN, HR and BA designed the protocol. SB, SR, AT and SCN oversaw the implementation of the programme. MW and SA managed the day-to-day operations of the programme. HR provided technological support. PKM drafted the manuscript. All the authors edited and revised the draft and approved the final version. PKM and BA are the guarantors. Funding: This work was supported by an intramural research grant from the St John’s Medical College Research Society, Bangalore, India, and a development grant from the Robert Bosch Centre for Cyber Physical Systems, Indian Institute of Science, Bangalore, India. Competing interests: PKM, PT, SB, SR, MW, SA, AT and SCN are employed at an academic medical centre (St John’s National Academy of Health Sciences, Bangalore, India) that receives public sector research funding, as well as revenue through private-sector fee-for-service medical transactions. BA and HR are employed at a public-sector science research centre (Indian Institute of Science). Patient consent: Obtained. Ethics approval: Ethics approval for the study was obtained from the St John’s Institutional Ethics Review Board (IERB #59/2011 dated 15 March 2011 and #125/2014 dated 16 August 2014). The clinical trial was registered with the Clinical Trials Registry of India (CTRI/2015/05/005779). Provenance and peer review: Not commissioned; externally peer reviewed.

Introduction Over 285 million people worldwide suffer from significant visual impairment of which approximately 60%–70% is considered to be preventable and treatable.1 2 The majority of cases reside in low-income and middle-income countries (LMICs)1 where access to diagnostic tools is least.3 4 Early identification and treatment, particularly of those conditions affecting the retina and optic nerve, relies on tools such as the traditional direct ophthalmoscope (TDO). Although the TDO has traditionally been one of the main ophthalmic screening and diagnostic tools, it has limitations of high initial cost of purchase, requirement for regular maintenance (bulbs and batteries) and perceived difficulty of use.5 6 Consequently, the ‘functional’ availability of this important device in LMICs is limited. The Vision 2020 Right to Sight initiative is a global strategy, launched in 1999, with the aim to eliminate avoidable blindness by 2020.7 The initiative has three main aims: disease control, human resource development and infrastructure strengthening including relevant technology development for eye care delivery.8 Along with the Lancet Commission,9 recommendations have emphasised the development of ‘frugal’ and culturally appropriate technology for users in LMICs. In response to these challenges, a novel low-cost multipurpose diagnostic device has been developed by Arclight Medical10 employing a number of innovative and unique design features (figure 1). Figure 1 The Arclight with selected features highlighted.

The Vision 2020 Right to Sight initiative is a global strategy, launched in 1999, with the aim to eliminate avoidable blindness by 2020.7 The initiative has three main aims: disease control, human resource development and infrastructure strengthening including relevant technology development for eye care delivery.8 Along with the Lancet Commission,9 recommendations have emphasised the development of ‘frugal’ and culturally appropriate technology for users in LMICs. In response to these challenges, a novel low-cost multipurpose diagnostic device has been developed by Arclight Medical10 employing a number of innovative and unique design features (figure 1). Figure 1 The Arclight with selected features highlighted. The key patented feature of the device is to employ a Light Emitting Diode (LED), which can be charged by an integrated photovoltaic (solar) panel. This novel design eliminates the need for bulky, hard to source and expensive replacement filament bulbs and batteries and mains electricity charging. The LED is small enough to be placed directly below the viewing hole, facing the patient, creating a near axial light source (figure 2), avoiding using a mirror to redirect light from below as is traditionally done. These design changes create a slim (110 mm long×26 mm wide×9 mm thick) and light device (18 g). Figure 2 The Arclight illumination uses a light emitting diode placed directly below the viewing hole, facing the patient, creating a near-axial light source unlike the traditional direct ophthalmoscope, which uses a mirror to redirect light into an axial path.

The key patented feature of the device is to employ a Light Emitting Diode (LED), which can be charged by an integrated photovoltaic (solar) panel. This novel design eliminates the need for bulky, hard to source and expensive replacement filament bulbs and batteries and mains electricity charging. The LED is small enough to be placed directly below the viewing hole, facing the patient, creating a near axial light source (figure 2), avoiding using a mirror to redirect light from below as is traditionally done. These design changes create a slim (110 mm long×26 mm wide×9 mm thick) and light device (18 g). Figure 2 The Arclight illumination uses a light emitting diode placed directly below the viewing hole, facing the patient, creating a near-axial light source unlike the traditional direct ophthalmoscope, which uses a mirror to redirect light into an axial path. Additionally, the Arclight has an integrated magnifying loupe and otoscope broadening its diagnostic potential. As it is available at only £5.00 to low-income users when sold in bulk,11 it is potentially an economically and practical alternative to TDOs in LMICs. As yet, however, no studies have evaluated the effectiveness of the Arclight in diagnosing retinal disorders (including diabetic retinopathy) among eye care providers in a LMIC. The aim of this study The aim of the study is to compare how the Arclight device performs in comparison to a TDO when used by Malawian eye healthcare professionals examining the fundi of simulated eyes.

Additionally, the Arclight has an integrated magnifying loupe and otoscope broadening its diagnostic potential. As it is available at only £5.00 to low-income users when sold in bulk,11 it is potentially an economically and practical alternative to TDOs in LMICs. As yet, however, no studies have evaluated the effectiveness of the Arclight in diagnosing retinal disorders (including diabetic retinopathy) among eye care providers in a LMIC. The aim of this study The aim of the study is to compare how the Arclight device performs in comparison to a TDO when used by Malawian eye healthcare professionals examining the fundi of simulated eyes. Materials and methods Appropriate ethical approval was obtained. Twenty-five eye healthcare professionals and optometry students were recruited with fully informed consent. Participants were recruited from two hospital eye departments and one school of optometry. The participants were already familiar with using a TDO. An introductory ‘refresher’ session was provided on direct ophthalmoscopy and the examiners had a brief training session with the Arclight to familiarise themselves with the practicalities of how the device works.

Materials and methods Appropriate ethical approval was obtained. Twenty-five eye healthcare professionals and optometry students were recruited with fully informed consent. Participants were recruited from two hospital eye departments and one school of optometry. The participants were already familiar with using a TDO. An introductory ‘refresher’ session was provided on direct ophthalmoscopy and the examiners had a brief training session with the Arclight to familiarise themselves with the practicalities of how the device works. Clinical signs, diagnosis and ease of use (EOU) scores Participants were asked to examine 12 retinal pathology slides placed in an Adam-Rouilly teaching manikin head (figure 3). Slides included background retinopathy and diabetic maculopathy, preproliferative retinopathy, advanced proliferative retinopathy, diabetic maculopathy, pan laser photocoagulation, normal fundus, glaucomatous disc, papilloedema/swollen disc, toxoplasmosis scar/optic atrophy, cytomegalovirus, central retinal vein occlusion and central retinal artery occlusion. A cross-over design was utilised with 12 participants using the Arclight first (group 1) and the TDO second and 13 participants the converse (group 2). The TDO used in this study was a Keeler Professional V.2.8. The order of the slides was randomised before ‘crossing over’. Figure 3 The Arclight being used with teaching slides in a manikin head.

Clinical signs, diagnosis and ease of use (EOU) scores Participants were asked to examine 12 retinal pathology slides placed in an Adam-Rouilly teaching manikin head (figure 3). Slides included background retinopathy and diabetic maculopathy, preproliferative retinopathy, advanced proliferative retinopathy, diabetic maculopathy, pan laser photocoagulation, normal fundus, glaucomatous disc, papilloedema/swollen disc, toxoplasmosis scar/optic atrophy, cytomegalovirus, central retinal vein occlusion and central retinal artery occlusion. A cross-over design was utilised with 12 participants using the Arclight first (group 1) and the TDO second and 13 participants the converse (group 2). The TDO used in this study was a Keeler Professional V.2.8. The order of the slides was randomised before ‘crossing over’. Figure 3 The Arclight being used with teaching slides in a manikin head. The participants were asked to describe the clinical signs they saw on the slide and to provide a diagnosis. Every clinical sign correctly identified received one mark. The time taken to reach a diagnosis was also recorded. Correct diagnoses, recognition of clinical signs and time taken to diagnose were normally distributed and a paired t-test was used to assess statistical significance. Participants were additionally asked to rate the EOU as described below:Could not use this ophthalmoscope to see the red reflex. Could see the red reflex. Could not focus the fundus. Could see vessels but not optic disc. Could see the disc and retinal fields but it was not in focus.

The participants were asked to describe the clinical signs they saw on the slide and to provide a diagnosis. Every clinical sign correctly identified received one mark. The time taken to reach a diagnosis was also recorded. Correct diagnoses, recognition of clinical signs and time taken to diagnose were normally distributed and a paired t-test was used to assess statistical significance. Participants were additionally asked to rate the EOU as described below:Could not use this ophthalmoscope to see the red reflex. Could see the red reflex. Could not focus the fundus. Could see vessels but not optic disc. Could see the disc and retinal fields but it was not in focus. Could see the disc and retinal fields in focus with high level of difficulty. Could see disc and retinal fields in focus with medium level of difficulty. Could see disc and retinal fields in focus with low level of difficulty. For descriptive statistics, an EOU score greater than 7 was defined as ‘easy’ and less than 6 (</6) as ‘difficult’. The EOU score frequency was not normally distributed and a non-parametric Wilcoxon sign rank test was applied.

Could see disc and retinal fields in focus with medium level of difficulty. Could see disc and retinal fields in focus with low level of difficulty. For descriptive statistics, an EOU score greater than 7 was defined as ‘easy’ and less than 6 (</6) as ‘difficult’. The EOU score frequency was not normally distributed and a non-parametric Wilcoxon sign rank test was applied. Results Identification of clinical signs A possible 53 pathological features could be diagnosed in the 12 slides. Arclight mean score per participant was 31.6 (95% CI 28.85 to 34.35) and TDO 28.72 (95% CI 26.42 to 31.01). Participants using the Arclight were significantly better at correctly identifying clinical signs compared with the TDO; the mean difference in the total number of correctly identified features by each participant was 2.89 in favour of the Arclight, t=3.285, P=0.003, 95% CI 1.13 to 4.65. Pathology diagnosis Fourteen possible diagnoses could be made from the 12 slides; Arclight mean score was 9.8 (95% CI 8.87 to 10.73) and the TDO 9.12 (95% CI 8.23 to 10.01). Users of the Arclight were significantly better at correctly diagnosing the pathology illustrated on the slide; the mean difference in the number of slides correctly diagnosed by participants was 0.685 in favour of the Arclight, t=2.957, P=0.007, 95% CI 0.21 to 1.17.

7 to 10.73) and the TDO 9.12 (95% CI 8.23 to 10.01). Users of the Arclight were significantly better at correctly diagnosing the pathology illustrated on the slide; the mean difference in the number of slides correctly diagnosed by participants was 0.685 in favour of the Arclight, t=2.957, P=0.007, 95% CI 0.21 to 1.17. Four of the 12 slides displayed solely signs of diabetic retinopathy (DR), so a maximum of 100 DR slides could be identified by all 25 participants, using the Arclight; participants correctly identified 83 slides with DR signs and the TDO 79 slides. When analysing theses signs, the sensitivity of the Arclight was 83% (95% CI 75.63 to 90.36) and TDO 79% (95% CI 71.01 to 87.00) and specificity of the Arclight was 98% (95% CI 96.06 to 99.94) and TDO 97.5% (95% CI 95.34 to 99.66) demonstrating that the devices were comparable to each other as a screening tool. Time to diagnosis Arclight mean time per slide was 93.19 s (95% CI 80.40 to 105.98) and TDO mean 103.15 s (95% CI 85.99 to 120.31). There was a non-significant trend for participants to make faster diagnoses with the Arclight; the mean difference between the time taken by participants to diagnose each slide was −8.95 s in favour of the Arclight, t=1.978, with P=0.06, 95% CI −18 to 0.1.

0.40 to 105.98) and TDO mean 103.15 s (95% CI 85.99 to 120.31). There was a non-significant trend for participants to make faster diagnoses with the Arclight; the mean difference between the time taken by participants to diagnose each slide was −8.95 s in favour of the Arclight, t=1.978, with P=0.06, 95% CI −18 to 0.1. EOU score For every examination, the user recorded a subjective EOU score; overall 300 examinations were performed with each device. More examinations were rated as ‘easy’ (defined as a score greater than 7) when using the Arclight (68%) compared with the TDO (59%). Mean score for Arclight was 6.7 (SEM±0.16) and TDO 6.5 (SEM±0.16). Using the Wilcoxon Sign rank test, this was statistically significant (P=0.01). Discussion Health clinics in LMICs often become ‘grave yards’ of redundant non-functional devices.9 Diagnostic tools are usually designed for health systems of wealthy countries, leading to expensive and overly complex devices, which are unnecessarily challenging to use and hard to maintain.3

EOU score For every examination, the user recorded a subjective EOU score; overall 300 examinations were performed with each device. More examinations were rated as ‘easy’ (defined as a score greater than 7) when using the Arclight (68%) compared with the TDO (59%). Mean score for Arclight was 6.7 (SEM±0.16) and TDO 6.5 (SEM±0.16). Using the Wilcoxon Sign rank test, this was statistically significant (P=0.01). Discussion Health clinics in LMICs often become ‘grave yards’ of redundant non-functional devices.9 Diagnostic tools are usually designed for health systems of wealthy countries, leading to expensive and overly complex devices, which are unnecessarily challenging to use and hard to maintain.3 This study has shown that the Arclight can satisfy the recommendations of the Vision 202012 initiative and the Lancet Commission9 to develop devices for LMICs that are low cost, portable and independent of scarce and expensive consumables while not compromising the core function. In the hands of Malawian eye care workers, the Arclight, despite being many fold less expensive than TDOs, performed better. This supports Lowe et al’s5 previous study demonstrating the Arclight to be a suitable alternative to the TDO for the purpose of screening and accurately diagnosing optic nerve disease.

function. In the hands of Malawian eye care workers, the Arclight, despite being many fold less expensive than TDOs, performed better. This supports Lowe et al’s5 previous study demonstrating the Arclight to be a suitable alternative to the TDO for the purpose of screening and accurately diagnosing optic nerve disease. Our study is also consistent with this previous study demonstrating the Arclight to be easier to use than a TDO. Possible reasons for this include the simplified design of the Arclight with a minimum of switches and dials creating an intuitive and uncomplicated device. Also, as previously mentioned, the Arclight is small, light and importantly very slim (9 mm thick) allowing the user to get closer to the pupil plane of the patient potentially offering a wider more stable field of view (figure 2). These features may also help to explain why users are more accurate at identifying retinal signs and make more correct diagnoses compared with when using a TDO. There were, however, important limitations to this study. An Adam Rouilly teaching head with simulated eyes cannot replicate media opacity, variability in pupil size or the head and eye movement of a real patient and consequently further comparative studies using real patients and in particular those with diabetic retinopathy should be considered in the future.5 13

his study. An Adam Rouilly teaching head with simulated eyes cannot replicate media opacity, variability in pupil size or the head and eye movement of a real patient and consequently further comparative studies using real patients and in particular those with diabetic retinopathy should be considered in the future.5 13 An additional limitation is that participant’s previous training, experience and familiarity with a TDO may have negatively influenced their impression of the Arclight.6 Despite this, the Arclight was considered easier to use than the TDO in this study emphasising the simplified and intuitive design as a major strength of the device. Further studies should explore the other functions of the device, including the anterior segment loupe, which allows differentiation of causes of medial opacity (eg, cataract vs corneal scar) and a magnified view of the tarsal plates for diagnosis and grading trachoma.14 Additionally, no studies have evaluated the effectiveness of the otoscope function. Conclusion In conclusion, this study provides additional evidence that the Arclight is a genuine practical and economic alternative to the TDO. By overcoming barriers to acquisition, maintenance and practical use that are present in low-resource settings, the device represents an important addition to the current strategies being employed to reduce the global burden of blindness.2 In addition, the Arclight can also function as a loupe and otoscope and consequently could have a broader role to play in healthcare.

d practical use that are present in low-resource settings, the device represents an important addition to the current strategies being employed to reduce the global burden of blindness.2 In addition, the Arclight can also function as a loupe and otoscope and consequently could have a broader role to play in healthcare. A very special thanks to the three institutions that facilitated this study and the important work they do to help reduce the burden of eye disease in Malawi. Thank you also to all participants who gave up their valuable time to assist in this study. Contributors: RB, EF, CA and AB conceived the idea for the study and planned the study. RB, EF, CA, JM and TC secured ethical approval and executed the study. RB, EF and AB analysed the data. RB, DR and AB wrote the initial manuscript. AB is responsible for the overall content as guarantor. All authors edited the manuscript. RB and AB submitted the manuscript. Funding: This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors. Competing interests: AB is seconded to the University of St Andrews from NHS Fife. The University owns a social enterprise subsidiary company, for which AB acts as an unpaid adviser. The social enterprise business sells the Arclight to users in high resource countries with all profits being used to fund distribution and education exercises of the device in low-income countries via the University’s Global Health Implementation Team.

erprise subsidiary company, for which AB acts as an unpaid adviser. The social enterprise business sells the Arclight to users in high resource countries with all profits being used to fund distribution and education exercises of the device in low-income countries via the University’s Global Health Implementation Team. Ethics approval: University of Dundee & St Andrews Research & Teaching Ethics Committees (URTEC) and University of Malawi College of Medicine Research and Ethics Committee (COMREC). Provenance and peer review: Not commissioned; externally peer reviewed.

Introduction The health sector is a priority for governments and supranational institutions. WHO defines health systems as those that are working in harmony and are ‘built on having trained and motivated health workers, a well-maintained infrastructure, and a reliable supply of medicines and technologies, backed by adequate funding, strong health plans and evidence-based policies’.1 The activities that these institutions and actors develop are aimed at preventing and controlling disease, treating the sick, and conducting research and training in health. This is why the health sector has a direct impact on several of the United Nations’ Sustainable Development Goals.2 Although the health sector’s importance is more than evident at the social level, its importance at the economic level, recognised over recent years, should also be noted. According to the World Bank (2017), health spending represented 12.3% of the gross domestic product (GDP) of the countries of the Organization for Economic Cooperation and Development in 2014, and its contribution to economic growth in some countries is between 3.5% and 5.5%.3 One of the main challenges facing the health sector comes from the need to innovate efficiently. This is so that, on the supply side, health institutions can perform care processes more quickly, accurately and economically; and on the demand side, patients can receive a more responsive and higher quality service that is better adapted to their needs and circumstances, thus improving their experience throughout the entire process.

on the supply side, health institutions can perform care processes more quickly, accurately and economically; and on the demand side, patients can receive a more responsive and higher quality service that is better adapted to their needs and circumstances, thus improving their experience throughout the entire process. The intense use of information and communication technologies (ICTs) in the health field evidences a notable improvement in results obtained by institutions, health professionals and patients.4 In the Americas, the relationship between economic development and health innovation is not particularly evident.5 Specifically, low-income to middle-income economies (deemed so based on the International Monetary Fund wealth index) allocate a high percentage of their GDP to health expenditures while showing innovation indicators equal to or higher than those of high-income economies such as the USA and Canada.

rticularly evident.5 Specifically, low-income to middle-income economies (deemed so based on the International Monetary Fund wealth index) allocate a high percentage of their GDP to health expenditures while showing innovation indicators equal to or higher than those of high-income economies such as the USA and Canada. Given this situation, the present work aims to determine whether a nation’s striving towards the health and well-being of its citizens depends solely on the level of economic development, or whether it also depends on the social actions of its leaders. To do this, we have taken the region of the Americas and its experiences in the development of digital health (or eHealth) as a unit of analysis. Data from the Third Global Survey on eHealth for the region of the Americas seem to show that health and well-being depend on both countries’ economic development and the social actions of its leaders since, regardless of governments’ political leanings (which may be totally opposed as in countries such as Cuba and Honduras), the indicators related to certain health management practices are high. Furthermore, the policies developed in the healthcare arena dub universal coverage, eHealth and technology training for healthcare professionals as key in health policy.

ings (which may be totally opposed as in countries such as Cuba and Honduras), the indicators related to certain health management practices are high. Furthermore, the policies developed in the healthcare arena dub universal coverage, eHealth and technology training for healthcare professionals as key in health policy. Methodology We analysed the data from the Third Global Survey on eHealth from 2015 for the region of the Americas. Of the 35 member countries, data pertaining to 19 (54.9%) were obtained. Analysis was carried out using univariate and bivariate statistical methods. For this reason, the results are mainly presented as a percentage of the total number of members that answered a question or as the absolute number of countries that gave a particular response.

ies, data pertaining to 19 (54.9%) were obtained. Analysis was carried out using univariate and bivariate statistical methods. For this reason, the results are mainly presented as a percentage of the total number of members that answered a question or as the absolute number of countries that gave a particular response. Global Survey on eHealth The Third Global Survey on eHealth is an instrument developed by WHO’s Global Observatory for eHealth,6 based on the consultation with, and information from, some of the strategic partners of WHO and the Pan American Health Organisation: governments, collaborating centres, professional associations and international organisations. The survey is modified and updated based on exhaustive analysis of information collected from the responses of the Member States. The purpose of WHO’s digital health studies carried out worldwide is to observe and determine the reference points in the adoption and progress of eHealth at the national, regional and global levels. The first survey was conducted in 2005 and its objective was to collect information at the national level in order to establish an analysis of the existing situation. The second survey, conducted in 2009 and based on the first survey, incorporated new questions and included a new approach on the subject of eHealth. The 2015 survey examined digital health in the context of its role in supporting universal health coverage (UHC).7 The objectives of this third survey were to measure, at a global level, progress in the development of eHealth and to later compare these results with those obtained in previous studies in order to identify current barriers and explore future trends in terms of digital health.

pporting universal health coverage (UHC).7 The objectives of this third survey were to measure, at a global level, progress in the development of eHealth and to later compare these results with those obtained in previous studies in order to identify current barriers and explore future trends in terms of digital health. Results The region is characterised by a strong cultural, social, linguistic, demographic and economic diversity that exists among the 35 countries that comprise it. A result of the region’s historical tradition, this situation also influences the way of approaching and managing health. Therefore, for some countries, the level of economic development—or even the political orientation of governments—seems to have no impact on the way health is managed (in particular, the degree of social coverage offered or the level of innovation achieved). Social approach to health policy: national coverage and eHealth plans Health is a matter of great social and economic importance for the countries in the American continent. Yet, the health policies implemented by the different states show great disparity, both in terms of the level of coverage offered and in relation to the implementation of eHealth policies.

coverage and eHealth plans Health is a matter of great social and economic importance for the countries in the American continent. Yet, the health policies implemented by the different states show great disparity, both in terms of the level of coverage offered and in relation to the implementation of eHealth policies. The survey data show that 52.6% of the countries of the Americas have a national policy or strategy for UHC. However, this percentage is not consistent across the region, especially when considering income levels; contrary to what might be expected, 100% of low-income countries have a UHC policy, compared with 50% and 54.5% of countries with high and middle-to-high incomes, respectively (table 1). Table 1 Health policy indicators by countries according to their level of income, universal health coverage, and information and communication technology use

The survey data show that 52.6% of the countries of the Americas have a national policy or strategy for UHC. However, this percentage is not consistent across the region, especially when considering income levels; contrary to what might be expected, 100% of low-income countries have a UHC policy, compared with 50% and 54.5% of countries with high and middle-to-high incomes, respectively (table 1). Table 1 Health policy indicators by countries according to their level of income, universal health coverage, and information and communication technology use All countries (%) High-income countries (%) Middle-income to high-income countries (%) Low-income to middle-income countries (%) Universal health coverage 52.6 50.0 54.5 100 eHealth strategy 57.9 83.3 36.4 100 Electronic health records 52.6 50.0 54.5 50.0 Entity that oversees eHealth 26.3 33.4 9.1 100 Entity incentive on research, development and innovation 63.2 83.3 45.5 100 A similar situation was observed in the development of eHealth practices. The level of establishment of digital health and the existence of a regulatory system for this type of practice are not precisely linked to a country’s level of economic development. In general terms, 57.9% of the countries in the sample indicate that they have a national eHealth policy or strategy, 52.6% report having a national electronic health record system, but only 26.3% have an entity that supervises the quality, safety and reliability regulations for mobile health applications (table 1).

t. In general terms, 57.9% of the countries in the sample indicate that they have a national eHealth policy or strategy, 52.6% report having a national electronic health record system, but only 26.3% have an entity that supervises the quality, safety and reliability regulations for mobile health applications (table 1). As in the previous case, these percentages are higher for countries with high or low incomes. The countries have no significant differences in the existence of an electronic health record. However, in eHealth and the availability of a supervisory entity for eHealth, the differences are notable, with these entities existing in 100% of low-income to middle-income countries. Discussion Health: a multifaceted reality Beyond agreeing on a definition for health8—a fundamental yet controversial issue—it is imperative to establish the way in which governments and organisations have to design and carry out policies for their management. In 2005, all WHO Member States pledged to fight for UHC; this represented a collective expression of the belief that all people should have access to the health services they need without risk of financial ruin or impoverishment.9

Discussion Health: a multifaceted reality Beyond agreeing on a definition for health8—a fundamental yet controversial issue—it is imperative to establish the way in which governments and organisations have to design and carry out policies for their management. In 2005, all WHO Member States pledged to fight for UHC; this represented a collective expression of the belief that all people should have access to the health services they need without risk of financial ruin or impoverishment.9 From a social point of view, working towards UHC is a powerful mechanism to achieve better health and well-being outcomes, as well as to promote human development. In fact, WHO’s declaration of its intentions for UHC automatically deems health and social well-being a universal right. Nonetheless, any action involves the use of resources, so along with the social dimension, health also has an economic dimension that influences the way in which each country orients its health policies.10

t, WHO’s declaration of its intentions for UHC automatically deems health and social well-being a universal right. Nonetheless, any action involves the use of resources, so along with the social dimension, health also has an economic dimension that influences the way in which each country orients its health policies.10 Accordingly, achieving improvements in the relationship between the quantity and quality of life gained and the resources used to obtain such gains requires adequate health policies, but, above all, it is about making cost-effective clinical decisions.11 The efficiency of a health system depends essentially on health institutions and organisations, as well as healthcare professionals, having the necessary information, training and incentives. They are the ones who make most of the decisions regarding prevention, diagnosis, treatment and rehabilitation. For that reason, the availability of financial, human and technological resources seems to be necessary to develop efficient health policies and also to offer UHC.12

sary information, training and incentives. They are the ones who make most of the decisions regarding prevention, diagnosis, treatment and rehabilitation. For that reason, the availability of financial, human and technological resources seems to be necessary to develop efficient health policies and also to offer UHC.12 As Niembro et al 13 point out, the aforementioned idea that resource availability is essential for UHC shows a direct relationship between a country’s level of economic development and its health policies. Yet, the results of the Third Global Survey on eHealth call into question some of the previous statements. For instance, contrary to expectations, some countries with scarce economic resources allocate abundant technological, human and economic resources to health management, offering UHC and also investing in highly innovative, competitive and efficient healthcare management. Economic development in the Americas: the health sector One of the main characteristics of the region of the Americas is the different levels of economic development of its countries. According to data from the International Monetary Fund,14 82.2% of the countries in the region can be classified as middle-income to high-income economies, whose per capita income in US dollars is between 2500 and 25 000 per year. Only 5.3% are considered high-income countries while 11.4% are classified as low-income to middle-income countries (below US$2500 per year per capita).

% of the countries in the region can be classified as middle-income to high-income economies, whose per capita income in US dollars is between 2500 and 25 000 per year. Only 5.3% are considered high-income countries while 11.4% are classified as low-income to middle-income countries (below US$2500 per year per capita). For all the countries in the region, health is a matter of priority. However, as the World Bank indicates,15 the importance of health expenditure for the Americas, as a percentage of total GDP, is a true reflection of the level of economic development of the countries (total health expenditure is the sum of public and private health expenditure, covering the provision of preventative and curative health services, family planning activities, nutrition activities and emergency healthcare assistance, but not including water supply or sanitary services). Consequently, high-income countries show health expenditures from 11.5% in Canada to 17.1% in the USA, whereas in lower-income countries, health expenditure does not exceed 6%–9% of GDP; for example, Nicaragua and the Dominican Republic have health expenditures of 3.6% and 5.6%, respectively. However, some low-income to middle-income countries do not display this pattern of behaviour; countries such as Haiti, Honduras and Cuba show very high expenditure given their level of economic development.

f GDP; for example, Nicaragua and the Dominican Republic have health expenditures of 3.6% and 5.6%, respectively. However, some low-income to middle-income countries do not display this pattern of behaviour; countries such as Haiti, Honduras and Cuba show very high expenditure given their level of economic development. In the specific case of Haiti, World Bank data show how the percentage of expenditure has decreased from 10.41% in 2011 to 7.5% in 2015. Meanwhile, the economies of Honduras and Cuba show health expenditures reaching 8.72% and 10.40% of GDP, respectively. This strong investment is manifested in the development and implementation of a health policy oriented towards social well-being, which competes with the most developed economies in terms of innovation. Innovation indicators: training and innovative programmes in eHealth Undoubtedly, innovation is one of the main sources of productivity and efficiency in the provision of health services in the context of the knowledge society. As Hall and Rosenberg16 indicate, the main source of innovation is the intentional application of knowledge and technology to develop new or better services, processes and organisational changes. Thus, the main innovation in the field of health comes from the intensive use of ICT, through the development of eHealth programmes and the constant training in technology that health personnel receive.

onal application of knowledge and technology to develop new or better services, processes and organisational changes. Thus, the main innovation in the field of health comes from the intensive use of ICT, through the development of eHealth programmes and the constant training in technology that health personnel receive. In the Americas, we can appreciate important differences between the countries both in the disposition of an entity that offers incentives and orientation for innovation, research and evaluation and in the use of advanced practices in eHealth, or the training of health professionals in technology and social networks. As previously noted, the survey data indicate that high-income and low-income to middle-income countries show higher percentages in relation to the existence of entities that promote innovation (table 1). Likewise, these countries also exceed 60%—compared with 40% and 50% in lower-income countries—in all cases regarding the use of eHealth practices, such as mobile health, remote patient monitoring or telehealth.

countries show higher percentages in relation to the existence of entities that promote innovation (table 1). Likewise, these countries also exceed 60%—compared with 40% and 50% in lower-income countries—in all cases regarding the use of eHealth practices, such as mobile health, remote patient monitoring or telehealth. Finally, with respect to training of healthcare professionals in the use of ICT, the percentages reported by the low-income countries are higher than those shown by the high-income and middle-income to high-income countries. One hundred per cent of low-income countries report offering ICT training to healthcare professionals, compared with 83% of wealthy countries and 81% of middle-income to high-income countries. It should be noted, however, that regardless of the economic level of the country, training offered to professionals in the use of social networks is relatively low, not exceeding 18%. Conclusions UHC is a citizen’s right, an instrument of economic efficiency, and a mechanism of equality and social well-being. The public policy of economic growth and social welfare should facilitate access to and use of efficient, effective and equitable health services. The new economic and social context posited by the advent of the knowledge society places people at the centre of the competitive and social well-being scenario. In this new context, having a population with an optimal state of health is a fundamental requirement.

of efficient, effective and equitable health services. The new economic and social context posited by the advent of the knowledge society places people at the centre of the competitive and social well-being scenario. In this new context, having a population with an optimal state of health is a fundamental requirement. The need for resources to carry out public policies that favour UHC, as well as an innovative and efficient health system, leads one to believe that the level of economic development of the country is a determinant for this. Nevertheless, reality allows us to observe that, just as in certain economies where the level of development is low, the effort made—both by public institutions and by organisations and health professionals—is high.

m, leads one to believe that the level of economic development of the country is a determinant for this. Nevertheless, reality allows us to observe that, just as in certain economies where the level of development is low, the effort made—both by public institutions and by organisations and health professionals—is high. The results from the Third Global Survey on eHealth for the region of the Americas primarily show that the use of ICTs has intensified, and that the level of income, as well as countries’ health expenditure, are not observed as a determining variable in the implementation of eHealth programmes. It would be up to the countries to bet on the benefits of the use of eHealth, which would favour its implementation. This is the case of economies considered poor ones, which allocate high percentages of their GDP towards health expenditure. In that vein, these countries strive to offer UHC, to develop advanced practices in the field of eHealth, to train their health professionals in the field of ICTs and to have advanced legislation in the field of digital health. This could explain how Cuba, a country qualified as low-middle income and with low healthcare expenditure, has a public, universal and free healthcare system. To this system an increasing development of the services infrastructure and human capital is added, along with a significant number of eHealth programmes, making Cuba one of the nations with the best health indicators in the world.5

nd with low healthcare expenditure, has a public, universal and free healthcare system. To this system an increasing development of the services infrastructure and human capital is added, along with a significant number of eHealth programmes, making Cuba one of the nations with the best health indicators in the world.5 Contributors: DN-O and FS-R conceptualised the study and designed the paper. DN-O, MD’A, ETK, FB-P and EMD contributed to the methodology to obtain the data from countries and checked the accuracy of extracted data. FS-R, JT-S, AJ-Z and DN-O wrote the first draft of the review paper. EMD, MD’A, FB-P and ETK contributed to the final draft of this manuscript. All authors read and approved the final manuscript. Funding: None. Disclaimer: DN-O, EMD, MD’A and FB-P are staff members of the Pan American Health Organization/World Health Organization (PAHO/WHO) and are themselves alone responsible for the views expressed in the paper, which do not necessarily represent the views, decisions or policies of the PAHO/WHO. Competing interests: None declared. Patient consent: Not required. Provenance and peer review: Not commissioned; externally peer reviewed.

Introduction WHO data from 2015 state that the lifetime risk of maternal death in high-income countries is 1 per 6000 pregnancies compared with 1 in 36 in Sub-Saharan Africa.1 Efforts to rectify this stark inequity is a global health priority. Identifying mechanisms to achieve this is the challenge. Approximately 800 women die each day from complications of pregnancy and childbirth.2 Obstetric haemorrhage, hypertensive disorders, sepsis and abortion complications (often from haemorrhage or sepsis) contribute to 60% of all maternal deaths. Each of these conditions is associated with abnormalities in blood pressure and pulse.2 The majority of maternal deaths are preventable through simple intervention if administered promptly, including induction of labour, antihypertensives, misoprostol, uterine balloon tamponade and antibiotics.3 These interventions are often available even in the most low-resource settings; however, a common barrier to prompt intervention is a failure to recognise the severity of the condition.4 5 Detection of abnormal vital signs, including blood pressure and pulse, allows early recognition of women who require urgent treatment or transfer to higher intensity of care, and is critical in preventing mortality and morbidity. Despite the universally recognised importance of blood pressure and pulse measurement, access to accurate blood pressure devices, healthcare provider training on measurement and adequate escalation pathways are lacking for many women in low-income and middle-income countries (LMICs).6

reventing mortality and morbidity. Despite the universally recognised importance of blood pressure and pulse measurement, access to accurate blood pressure devices, healthcare provider training on measurement and adequate escalation pathways are lacking for many women in low-income and middle-income countries (LMICs).6 Even in high-income countries, it is acknowledged that failures by healthcare providers to immediately recognise and act on signs of life-threatening conditions, particularly haemorrhage, pre-eclampsia and sepsis, contribute to avoidable maternal deaths.7 8 To aid earlier recognition of compromise, modified obstetric early warning system charts are now encouraged internationally for use in the care of pregnant and postpartum women.7–11 However, these charts are based on limited evidence, are labour-intensive and require clinician interpretation. In this article, we report the progress of the CRADLE programme: the development and evaluation of the CRADLE Vital Signs Alert (VSA) device, a low-cost, accurate and easy-to-use vital signs measuring device developed over the past 15 years. It is a hand-held, upper-arm, semiautomated device developed in collaboration with Microlife, with the primary goal of ensuring accuracy and applicability in pregnancy.

ation of the CRADLE Vital Signs Alert (VSA) device, a low-cost, accurate and easy-to-use vital signs measuring device developed over the past 15 years. It is a hand-held, upper-arm, semiautomated device developed in collaboration with Microlife, with the primary goal of ensuring accuracy and applicability in pregnancy. Ensuring accuracy In 2006, we validated the device for use in a non-pregnant adult population, according to a formal validation protocol for assessing accuracy of automated devices.12 The cohort included participants with a range of ages, gender distribution, arm circumferences and blood pressure values, as stipulated by the protocol. For each participant, the validation process involved multiple measurements of blood pressure by trained observers under test conditions, alternating between the test device and two calibrated mercury sphygmomanometers. Assessment of accuracy was determined by comparing the mean difference (SD) between measurements from the test device and the mercury sphygmomanometers. Validation in non-pregnant adults does not imply accuracy in pregnant women. Following the British Hypertension Society (BHS) guidelines,13 in 2012 we successfully validated the device in pregnancy, including women with pre-eclampsia.14 Our device is one of only a few devices suitable for non-pregnant and pregnant adults. Uniquely, we have also validated the device in pregnant women with low blood pressure (in South Africa, a LMIC), demonstrating the device maintains accuracy at the lower end of the range.15

ancy, including women with pre-eclampsia.14 Our device is one of only a few devices suitable for non-pregnant and pregnant adults. Uniquely, we have also validated the device in pregnant women with low blood pressure (in South Africa, a LMIC), demonstrating the device maintains accuracy at the lower end of the range.15 Creation of the VSA device Through Bill & Melinda Gates Foundation funding, we modified the VSA device to ensure suitability for community healthcare providers in LMICs. It now meets the WHO’s requirements, being affordable (US$20), robust, easy-to-use, portable, requiring infrequent calibration and has low power requirements (charging through a micro-USB port) and dual auscultatory/oscillometric function.14 16 We incorporated a traffic light early warning system, aiming to alert untrained healthcare providers to vital sign abnormalities from hypertension, or shock secondary to obstetric haemorrhage and sepsis (figure 1). For maternal hypertension, established blood pressure thresholds trigger the green, yellow and red traffic lights of the early warning system. For shock (from haemorrhage or sepsis), we determined the early warning system thresholds through retrospective analysis of vital signs in women with obstetric haemorrhage (in both low-income and high-income settings).17 18 Shock Index (pulse/systolic blood pressure) is a novel marker known to be an early predictor of complications in trauma/combat settings19–23 and a more consistent predictor of adverse outcomes than individual conventional vital signs (blood pressure, pulse, mean arterial pressure and pulse pressure) in obstetric haemorrhage.17 18 We confirmed Shock Index normal range in a prospective observational study.24 Shock Index thresholds were selected for incorporation into the traffic light early warning system according to associated risk of adverse outcomes (figure 2).17 18

se, mean arterial pressure and pulse pressure) in obstetric haemorrhage.17 18 We confirmed Shock Index normal range in a prospective observational study.24 Shock Index thresholds were selected for incorporation into the traffic light early warning system according to associated risk of adverse outcomes (figure 2).17 18 Figure 1 The prototype Microlife 3AS1-2 and the final CRADLE Vital Signs Alert, with traffic light early warning system. Figure 2 Vital Signs Alert device traffic light early warning system display options. DBP, diastolic blood pressure; SBP, systolic blood pressure; SI, Shock Index.

se, mean arterial pressure and pulse pressure) in obstetric haemorrhage.17 18 We confirmed Shock Index normal range in a prospective observational study.24 Shock Index thresholds were selected for incorporation into the traffic light early warning system according to associated risk of adverse outcomes (figure 2).17 18 Figure 1 The prototype Microlife 3AS1-2 and the final CRADLE Vital Signs Alert, with traffic light early warning system. Figure 2 Vital Signs Alert device traffic light early warning system display options. DBP, diastolic blood pressure; SBP, systolic blood pressure; SI, Shock Index. Methods Preliminary fieldwork In 2013, we identified that over 90% of health clinics in a rural district of Tanzania either had no blood pressure devices or broken devices. We introduced 19 of our devices, containing tally counters to monitor use, into these clinics and showed that they were used frequently; focus groups confirmed they were well accepted by healthcare providers. This exploratory work led to a preintervention and postintervention study, conducted over 12 months in three rural hospitals in Tanzania, Zimbabwe and Zambia. During the 3-month preintervention phase, blood pressure was measured on consecutive pregnant women at 20 weeks’ gestation or more presenting to hospital for any reason. Blood pressure was measured twice (with a 5 min rest period between measurements) using the validated CRADLE prototype device (ie, without the traffic light early warning system) by trained volunteers. Urine dipstick results were recorded if available, together with demographic details and antenatal history, including whether blood pressure had ever been taken in the index pregnancy prior to presentation. These data were entered onto a central study database. During the 3-month intervention phase, 20 rural and semirural peripheral antenatal clinics that referred to each of the three hospitals received one to two of the CRADLE prototype devices (n=100) together with two face-to-face training sessions, translated instructions on its use and a guide on how to refer based on the blood pressure readings. One year later, in a matched time of year to the preintervention period, data collection was repeated as per the preintervention phase.

CRADLE prototype devices (n=100) together with two face-to-face training sessions, translated instructions on its use and a guide on how to refer based on the blood pressure readings. One year later, in a matched time of year to the preintervention period, data collection was repeated as per the preintervention phase. Iterative qualitative evaluation An iterative qualitative evaluation was undertaken to ensure that the VSA’s final traffic light early warning system and display symbols were understandable to users across a range of settings. Participants were shown a video animation of display options for hypertension and shock alerts and device utilities (low battery, error). Multiple choice questionnaires directly relating to the video animation and exploring the understanding of the options for the early warning system and the display symbols were distributed to a range of clinical and non-clinical participants from academic collaborative centres in India, the UK and Canada.

tery, error). Multiple choice questionnaires directly relating to the video animation and exploring the understanding of the options for the early warning system and the display symbols were distributed to a range of clinical and non-clinical participants from academic collaborative centres in India, the UK and Canada. Wide-range cuff validation Incorrect cuff size leads to blood pressure measurement inaccuracies.25–27 All VSA device validations have included individuals with a range of arm circumferences, reflecting the general population. The VSA device comes with a medium cuff as standard and large cuffs as required. Obesity is a global problem, with prevalence reaching more than 70% in the Americas and Caribbean and more than 40% in Africa.28 In pregnancy, obesity is associated with increased risk of pre-eclampsia, obstetric haemorrhage and sepsis.29–31 To ensure accuracy in all women, particularly those with large arm circumferences, we endeavoured to validate the VSA device in combination with a novel wide-range cuff. This would also mean that healthcare providers would not need to estimate arm circumference and select the appropriate cuff prior to measurement.

9–31 To ensure accuracy in all women, particularly those with large arm circumferences, we endeavoured to validate the VSA device in combination with a novel wide-range cuff. This would also mean that healthcare providers would not need to estimate arm circumference and select the appropriate cuff prior to measurement. The prospective validation study included 52 pregnant women recruited from a large tertiary hospital. Twenty-six women had an arm circumference of 22–32 cm and 26 had an arm circumference of 33–42 cm. Each woman had seven sequential same-arm measurements taken, alternating between mercury sphygmomanometer and the VSA prototype device and appropriate, inappropriate and wide-range cuffs. Data were analysed using adapted validation methods recommended by the BHS, the International Society of Hypertension (ISH) and the International Organization for Standardization (ISO) protocols. Auscultatory blood pressure validation The VSA device can be used as an auscultatory device, using a stethoscope, and the falling digital display in a similar fashion to a mercury column while still providing a digital reading derived from oscillometry. This is particularly important in those with heart arrhythmias, as most automated devices fail to give accurate readings in these individuals, or when wanting to verify an unexpected automated reading. We undertook a prospective evaluation of clinician auscultatory VSA device measurement compared with VSA oscillometric measurement.

cularly important in those with heart arrhythmias, as most automated devices fail to give accurate readings in these individuals, or when wanting to verify an unexpected automated reading. We undertook a prospective evaluation of clinician auscultatory VSA device measurement compared with VSA oscillometric measurement. Prospective clinical validation A 15-month prospective validation of the early warning system of the VSA device was undertaken at three tertiary hospitals in South Africa. At the three maternity departments, existing blood pressure devices were replaced with VSA devices. Data, including vital signs and adverse clinical outcomes, were collected on all pregnant/postpartum women admitted with either pre-eclampsia, obstetric haemorrhage or sepsis. The performance of the traffic light early warning system thresholds to predict adverse clinical outcomes was evaluated. Results Preliminary fieldwork Data from 1241 women were analysed (n=694 preintervention; n=547 postintervention). Introduction of the device to rural clinics was associated with improved antenatal surveillance of blood pressure (a decrease in the proportion of women who never had blood pressure measured in pregnancy from 25.1% to 16.9%, OR 0.58, p<0.001, 95% CI 0.42 to 0.79). This informed us that the device was acceptable to healthcare providers and may facilitate early treatment for women with pre-eclampsia.

l surveillance of blood pressure (a decrease in the proportion of women who never had blood pressure measured in pregnancy from 25.1% to 16.9%, OR 0.58, p<0.001, 95% CI 0.42 to 0.79). This informed us that the device was acceptable to healthcare providers and may facilitate early treatment for women with pre-eclampsia. Iterative qualitative evaluation Eighty-eight questionnaires were distributed and fully completed by doctors, nurses, midwives, healthcare assistants, medical students and non-clinical participants from India, the UK and Canada. Options not well understood were abandoned, including a blue traffic light to represent shock and an ‘E’ to represent an error. The options understood by the highest number of participants were selected for the final VSA device. With our team’s awareness that shock, particularly from haemorrhage, often requires more urgent intervention than hypertension, we ensured that the early warning system prioritised shock over hypertension and the shock alert demonstrated a higher level of urgency than hypertension (flashing down arrow as opposed to constant up arrow).

awareness that shock, particularly from haemorrhage, often requires more urgent intervention than hypertension, we ensured that the early warning system prioritised shock over hypertension and the shock alert demonstrated a higher level of urgency than hypertension (flashing down arrow as opposed to constant up arrow). Wide-range cuff validation The wide-range cuff in combination with the VSA device failed validation in pregnant women with arm circumferences 22–42 cm, significantly overestimating both systolic and diastolic blood pressure in those with large arm circumferences, according to adapted BHS, ISH and ISO protocols. Therefore, the wide-range cuff cannot be recommended for use as a cuff suitable for a range of arm circumferences in a pregnant population. We went on to successfully validate the VSA device in combination with an extra-large cuff, meaning that the device is accurate when used in combination with medium, large and extra-large cuff sizes. Auscultatory blood pressure validation Blood pressure readings (n=255) taken by 12 clinicians in the antenatal clinic, taking simultaneous and masked auscultatory and oscillometric readings, were analysed according to formal validation protocols. The mean difference (±SD) between clinician auscultation and oscillometry was 2.2 (±6.1) mm Hg and 2.3 (±5.2) mm Hg for systolic and diastolic blood pressure, respectively, well within the ISO protocol requirements (≤5±8 mm Hg). This confirmed that the VSA device maintains accuracy when used as an auscultatory device by trained healthcare providers.

en clinician auscultation and oscillometry was 2.2 (±6.1) mm Hg and 2.3 (±5.2) mm Hg for systolic and diastolic blood pressure, respectively, well within the ISO protocol requirements (≤5±8 mm Hg). This confirmed that the VSA device maintains accuracy when used as an auscultatory device by trained healthcare providers. Prospective clinical validation For obstetric haemorrhage and sepsis, the VSA traffic light thresholds strongly predicted risk of complications, including intensive care unit admission and emergency hysterectomy. For hypertension, the risk of renal impairment, preterm delivery, maternal magnesium sulfate use and intensive care unit admission increased with worsening VSA traffic light thresholds.32 The studies represent the first prospective cohort studies evaluating thresholds of blood pressure and pulse as predictors of adverse outcomes in pregnancy and the first to evaluate thresholds in LMICs.

rnal magnesium sulfate use and intensive care unit admission increased with worsening VSA traffic light thresholds.32 The studies represent the first prospective cohort studies evaluating thresholds of blood pressure and pulse as predictors of adverse outcomes in pregnancy and the first to evaluate thresholds in LMICs. Discussion Current impact To date (June 2018), over 6700 VSA devices have been used clinically in 12 countries across Africa, Asia and the Caribbean. An earlier version of the VSA device has been used as part of the intervention package in the community-level intervention for pre-eclampsia cluster randomised controlled trial in Mozambique, India and Pakistan, which enrolled a total of 75 532 pregnant women across the intervention and control clusters and completed recruitment in February 2017. In some cases, the VSA device was the first blood pressure device available to a clinic, and in most cases the first accurate blood pressure device for pregnancy in clinics and hospitals. In addition to the benefit to women in LMICs, the device is also suited for use in pregnancy in high-income countries, for non-pregnant adults and for home monitoring. Device accuracy when used as a self-monitor by pregnant women has been confirmed in a recent Danish prospective observational study.33 The device has also been adopted by groups requiring vital sign measurement away from clinical settings, including the fire brigade, army and mountain rescue in the UK.

home monitoring. Device accuracy when used as a self-monitor by pregnant women has been confirmed in a recent Danish prospective observational study.33 The device has also been adopted by groups requiring vital sign measurement away from clinical settings, including the fire brigade, army and mountain rescue in the UK. We anticipate that women who would have previously gone unnoticed would now be identified early as being at risk. This simple technology may prevent maternal deaths and could contribute to meeting the United Nations Sustainable Development Goal 3: to reduce the global maternal mortality ratio to less than 70 000 per 100 000 live births by 2030. The potential contribution of the VSA device was recognised by the PATH Innovation Countdown 2030 Report, being nominated as one of 30 life-saving innovations with great promise to accelerate progress over the next 15 years to reach the Sustainable Development Goals.34

less than 70 000 per 100 000 live births by 2030. The potential contribution of the VSA device was recognised by the PATH Innovation Countdown 2030 Report, being nominated as one of 30 life-saving innovations with great promise to accelerate progress over the next 15 years to reach the Sustainable Development Goals.34 Challenges and lessons learnt It is recommended that all blood pressure devices are validated as accurate in the population for which they are intended.13 In 2006, although the VSA device passed non-pregnant adult validation, it concurrently failed validation in pregnancy, including pre-eclampsia. Over the next 5 years, the device’s algorithm was tweaked and revalidated. Only in 2012 did the device pass validation in pregnancy, including pre-eclampsia, highlighting the challenge of achieving accuracy in this particular population. This process delayed further device development and clinical evaluation but was crucial in ensuring a device that identifies abnormality and prevents morbidity.

y in 2012 did the device pass validation in pregnancy, including pre-eclampsia, highlighting the challenge of achieving accuracy in this particular population. This process delayed further device development and clinical evaluation but was crucial in ensuring a device that identifies abnormality and prevents morbidity. An obstetric early warning system triggered by accurate blood pressure and pulse measurement is important for all settings. Maternity units in high-income countries frequently rely on non-validated blood pressure devices and early warning charts. Considering the high volume of women and frequency of blood pressure and pulse measurement in well-resourced tertiary units, the potential of a fully automated version of the device is being explored. The fully automated VSA device would be more suitable for well-resourced settings (and many LMIC facility-level settings), where power supply and cost are less of an issue. The current VSA device, with its low power requirements and low cost, would remain appropriate for low-resource settings.

device is being explored. The fully automated VSA device would be more suitable for well-resourced settings (and many LMIC facility-level settings), where power supply and cost are less of an issue. The current VSA device, with its low power requirements and low cost, would remain appropriate for low-resource settings. Where is the project heading next? A stepped-wedge, cluster randomised controlled trial (funded by the UK Medical Research Council/Department for International Development/Department of Biotechnology, India) to determine whether implementation of the CRADLE VSA device and a simple training package (figure 3) into every level of routine maternity care will reduce maternal mortality and morbidity in 10 LMIC sites (ISRCTN41244132) completed data collection in December 2017. The trial is guided by the Medical Research Council Complex Intervention Framework35 and commenced with a formative pilot phase where the training package was developed with multiple stakeholders input and then introduced, with the VSA device, into three LMIC sites. Feasibility and acceptability of the intervention and its implementation were assessed using questionnaires, observation, interviews and focus groups.36 This work informed the final training package, implementation strategies and data collection method for the trial. We will further determine the effectiveness of the CRADLE VSA intervention package and how the intervention may work in different contexts. The knowledge gained from this work will inform eventual scale-up.

his work informed the final training package, implementation strategies and data collection method for the trial. We will further determine the effectiveness of the CRADLE VSA intervention package and how the intervention may work in different contexts. The knowledge gained from this work will inform eventual scale-up. Figure 3 CRADLE trial training package poster for community healthcare providers. VSA, Vital Signs Alert. Thank you to the research assistants for collecting the data locally. Contributors: Design: HLN, NV, SG, MFG, NLH, AHS. Planning: HLN, NV, AdG, NLH, NS, NT, SG, MFG, LCC, AHS. Conduct: HLN, EL, NS, NT, AdG, JS. Data analysis: HLN, EL, NS, NT, JS, LCC, AHS. Manuscript writing: HLN, NV, LCC, AHS. Responsible for the overall content: AHS. Funding: Funded by Bill & Melinda Gates Foundation (Grant ID: OPP1086183). Competing interests: None declared. Patient consent: Not required. Ethics approval: Stellenbosch University Ethics Committee (N14/06068), University of Cape Town Ethics Committees (410/2014) and the University of the Free State Ethics Committee (230408-011). Provenance and peer review: Not commissioned; externally peer reviewed.

Introduction Administration of tablets and capsules in people with dysphagia has been shown to be related to a threefold increase in medication error rate and this is largely attributed to the frequent need to modify the solid dose formulation by crushing or dispersing in water prior to administration.1 Guidance frequently recommends solid dose formulation alteration prior to administration in order to minimise the risk of choking and make oral bolus formation easier.2 Crushing modified release solid dose formulations can result in immediate dose release and therefore increase likelihood of patient harm as they receive the whole dose more rapidly than anticipated.3 Similarly, disrupting enteric coatings can increase likelihood of harm to the patient’s stomach, harm to the chemical entity or the drug being released higher up the gastrointestinal (GI) tract than anticipated.3 Consequently, in such cases, alternative routes of administration or formulations are required.4

icipated.3 Similarly, disrupting enteric coatings can increase likelihood of harm to the patient’s stomach, harm to the chemical entity or the drug being released higher up the gastrointestinal (GI) tract than anticipated.3 Consequently, in such cases, alternative routes of administration or formulations are required.4 Liquid formulations are available for most medicines and their prescription immediately overcomes concerns regarding bolus formation in those with oral phase dysphagia. The texture of liquid medicines for patients with pharyngeal phase dysphagia can however be inappropriate as they may increase the likelihood of aspiration.3 Due to the small market share and increased costs of production and storage, liquid medicines are more expensive than solid dose formulations and consequently cost is an additional barrier to their prescription.5 It is not possible for the liquid formulation to provide enteric coatings or sustained release characteristics, which are routinely seen with solid dose oral formulation, and consequently liquid medicines can require more frequent dosing and cannot be formulated to protect the stomach, active ingredients, or to ensure that these are released past the stomach and duodenum.

de enteric coatings or sustained release characteristics, which are routinely seen with solid dose oral formulation, and consequently liquid medicines can require more frequent dosing and cannot be formulated to protect the stomach, active ingredients, or to ensure that these are released past the stomach and duodenum. With a desire to provide patients with tablets or capsules intact but lubricated to ease the swallow and give the mouthfeel of a liquid medicine, we developed a new formulation gel which enabled tablets or capsules to be encapsulated prior to swallow, thereby removing the need for altering tablet form prior to administration. Using pharmaceutically inert materials, the gel (SMART Swallowing Aid) was designed and patented (US Patent Application No 12/866715, Japanese Patent Application No 2010-545564, European Patent Application No 09708982.5, Chinese Patent Application No 200980104351.4) to not react with the medicines or their coatings and to additionally become more lubricated when in contact with saliva. The formulation did not contain any sugar making it suitable for those patients on a controlled diet. The SMART Swallowing Aid prototype gel was developed with the proposed advantages of: Increasing the comfort of the swallow. Providing lubrication to the tablet thereby negating the need for concomitant administration of water. Eliminating the need to crush medication. Enabling medicines designed to be absorbed further down the GI tract to be delivered intact. Masking the taste of the medication.

With a desire to provide patients with tablets or capsules intact but lubricated to ease the swallow and give the mouthfeel of a liquid medicine, we developed a new formulation gel which enabled tablets or capsules to be encapsulated prior to swallow, thereby removing the need for altering tablet form prior to administration. Using pharmaceutically inert materials, the gel (SMART Swallowing Aid) was designed and patented (US Patent Application No 12/866715, Japanese Patent Application No 2010-545564, European Patent Application No 09708982.5, Chinese Patent Application No 200980104351.4) to not react with the medicines or their coatings and to additionally become more lubricated when in contact with saliva. The formulation did not contain any sugar making it suitable for those patients on a controlled diet. The SMART Swallowing Aid prototype gel was developed with the proposed advantages of: Increasing the comfort of the swallow. Providing lubrication to the tablet thereby negating the need for concomitant administration of water. Eliminating the need to crush medication. Enabling medicines designed to be absorbed further down the GI tract to be delivered intact. Masking the taste of the medication. Protecting the GI mucosa, in particular the oesophagus, to reduce oesophagitis.6 Increasing the weight of the tablet and thereby aid passage down the oesophagus into the stomach. Minimising accusations of administering medicines covertly7 by ensuring that the gel was transparent.

Enabling medicines designed to be absorbed further down the GI tract to be delivered intact. Masking the taste of the medication. Protecting the GI mucosa, in particular the oesophagus, to reduce oesophagitis.6 Increasing the weight of the tablet and thereby aid passage down the oesophagus into the stomach. Minimising accusations of administering medicines covertly7 by ensuring that the gel was transparent. Other tablet swallowing aids are currently available (Pill glide,8 Medcoat9) which are designed either to coat the tongue or the tablet, thereby masking the flavour and making it easier to swallow the medication. All products consist of ingredients, which are considered to be safe for humans, and are believed not to affect absorption of the drug from the tablet. Using a standard in vitro model, Medcoat has been shown to not adversely affect tablet disintegration.9 While the products claim to not affect drug absorption, none have been tested in patients to determine bioequivalence. With concerns regarding patient safety and effectiveness surrounding the encapsulation of a tablet prior to swallowing and a desire to register SMART Swallowing Aid as a medicinal device we decided to undertake in vivo testing.

o not affect drug absorption, none have been tested in patients to determine bioequivalence. With concerns regarding patient safety and effectiveness surrounding the encapsulation of a tablet prior to swallowing and a desire to register SMART Swallowing Aid as a medicinal device we decided to undertake in vivo testing. The aim of this study was to investigate the safety, efficacy and acceptability of SMART Swallowing Aid prototype. The primary objectives of this study were to establish bioequivalence of aspirin with and without SMART Swallowing Aid (drug plasma concentration and drug effect on platelet activity) and test the safety of SMART Swallowing Aid in human subjects. As a secondary outcome, the study determined whether the gel was easy and comfortable to swallow in comparison with the tablet administered with water and whether it separated from the tablet on swallowing. Method The study consisted of a phase IV open-label randomised controlled cross-over trial, comparing the bioavailability of aspirin encapsulated within SMART Swallowing Aid with the bioavailability of aspirin alone administered with water.

The aim of this study was to investigate the safety, efficacy and acceptability of SMART Swallowing Aid prototype. The primary objectives of this study were to establish bioequivalence of aspirin with and without SMART Swallowing Aid (drug plasma concentration and drug effect on platelet activity) and test the safety of SMART Swallowing Aid in human subjects. As a secondary outcome, the study determined whether the gel was easy and comfortable to swallow in comparison with the tablet administered with water and whether it separated from the tablet on swallowing. Method The study consisted of a phase IV open-label randomised controlled cross-over trial, comparing the bioavailability of aspirin encapsulated within SMART Swallowing Aid with the bioavailability of aspirin alone administered with water. SMART Swallowing Aid Gelatin, from a bovine origin and certified to be bovine spongiform encephalopathy/transmissible spongiform encephalopathy free, formed the main ingredient as it is odourless and tasteless and forms a gel at room temperature and melts around body temperature (<35○C). To improve the gel’s swelling and lubrication properties hydroxypropylmethylcellulose was added as it becomes less tacky on exposure to water and therefore in the presence of saliva will lubricate the formulation. Citric acid was added to increase saliva production and as a preservative. Potassium sorbate was also included as the amount of citric acid was insufficient alone to prevent microbial contamination. All ingredients were selected due to their wide routine use within pharmaceutical and food industries. The aspirin and gel combination was compared with aspirin alone in vitro via a model gut10 to confirm bioequivalence.

so included as the amount of citric acid was insufficient alone to prevent microbial contamination. All ingredients were selected due to their wide routine use within pharmaceutical and food industries. The aspirin and gel combination was compared with aspirin alone in vitro via a model gut10 to confirm bioequivalence. Trial The overall trial process is summarised in figure 1. Twelve healthy male volunteers were recruited via poster advertising within the University of East Anglia (UEA) and the Norfolk and Norwich University Hospital. As we expected small variation within individuals and we have evidence of equivalence based on lab-based approaches we felt that a sample size of 12 individuals would be sufficient to have high power to demonstrate bioequivalence. Figure 1 Study progress flow chart. CRTU, Clinical Research and Trials Unit; NNUH, Norfolk and Norwich University Hospital. The study inclusion criteria were participants who were: Healthy. Male. Aged 18–35 years. We selected only male participants as Aspirin should be used with caution in pregnancy and then at low dose. Testing in one gender also limits pharmacokinetic variation. The study exclusion criteria were: Student at either the UEA School of Pharmacy or Medicine. Allergy or hypersensitivity to aspirin (or other non-steroidal anti-inflammatory drugs) or formulation ingredients. Lactose intolerance or coeliac disease. Dysphagia or difficulty swallowing. Active peptic ulceration. Haemophilia or other bleeding disorders. Taking aspirin regularly or who have taken it in the last 7 days.

Student at either the UEA School of Pharmacy or Medicine. Allergy or hypersensitivity to aspirin (or other non-steroidal anti-inflammatory drugs) or formulation ingredients. Lactose intolerance or coeliac disease. Dysphagia or difficulty swallowing. Active peptic ulceration. Haemophilia or other bleeding disorders. Taking aspirin regularly or who have taken it in the last 7 days. Taking other medications, particularly those containing salicylates. Parallel participation in another research study. Related to or living with any member of the study team. Language difficulties. Participants were withdrawn if they suffered an adverse event or reaction in response to aspirin or the swallowing aid or if they demonstrated difficulty in swallowing using SMART. The study was planned to be terminated if the gel was deemed a choking hazard. Participants were randomised using sealed envelopes, which were presented in numerical order to each participant at the start of the first visit day. Block randomisation was used with a block size of 4. While this study could not be blinded, only the study statistician was aware of the randomisation details prior to the visit days. Participants attended the Clinical Research and Trials Unit (CRTU) based at the Norfolk and Norwich University Hospital having not consumed any food or drink for at least 2 hours prior to each visit. Dependent on randomised allocation each participant was asked to swallow either one SMART Swallowing Aid containing one 300 mg aspirin uncoated tablet (figure 2), or to swallow one 300 mg aspirin uncoated tablet with a glass of potable water.

Participants attended the Clinical Research and Trials Unit (CRTU) based at the Norfolk and Norwich University Hospital having not consumed any food or drink for at least 2 hours prior to each visit. Dependent on randomised allocation each participant was asked to swallow either one SMART Swallowing Aid containing one 300 mg aspirin uncoated tablet (figure 2), or to swallow one 300 mg aspirin uncoated tablet with a glass of potable water. Figure 2 Tablet encapsulated within SMART Swallowing Aid. NNUH, Norfolk and Norwich University Hospital. Aspirin should be taken with food to avoid the risk of stomach ulceration, and consequently to standardise stomach contents prior to administration all participants were given two Weetabix on arrival with a measured volume of semiskimmed milk. Aspirin is known to exhibit a half-life of 15–20 min, with that of its metabolite salicylate being 3–6 hours. A washout period of 1 week allowed elimination of aspirin and its metabolite in the region of 504 and 28 half-lives, respectively, which was adequate for the purposes of this study. Participants were asked not to take aspirin or any medications containing aspirin or salicylates for 7 days prior to each data collection event and this was confirmed on arrival at the visit day.

and its metabolite in the region of 504 and 28 half-lives, respectively, which was adequate for the purposes of this study. Participants were asked not to take aspirin or any medications containing aspirin or salicylates for 7 days prior to each data collection event and this was confirmed on arrival at the visit day. Venous blood samples were taken at time 0 and 20, 40, 60, 90 and 120 min after the dose.11 Serum salicylate levels were measured with Advia 2400, Clinical Chemistry System by contract research company Viapath, Kings College, London.12 Platelet function was measured by impedance aggregometry using the Multiplate system (Verum Diagnostica, Munich, Germany).13 Platelet function included both measurements of effect on arachidonic acid and collagen. Participants completed a questionnaire after taking the tablet with SMART Swallowing Aid and after taking the tablet with water. Questionnaires consisted of three visual analogue scales (VAS) in response to ‘The tablet was easy to swallow’, ‘The tablet was comfortable to swallow’ and ‘I could taste the tablet during the swallow’. The lower the score the more positive the response. Additionally, a single yes/no question ‘The gel aid broke/the tablet separated from the gel aid’ was included when the tablet was administered with the gel. Participants then returned after at least 1 week to repeat the study with the alternative method of administration and were remunerated for their participation. The process within the CRTU is summarised in figure 3. Figure 3 Visit progress flow chart.

Participants completed a questionnaire after taking the tablet with SMART Swallowing Aid and after taking the tablet with water. Questionnaires consisted of three visual analogue scales (VAS) in response to ‘The tablet was easy to swallow’, ‘The tablet was comfortable to swallow’ and ‘I could taste the tablet during the swallow’. The lower the score the more positive the response. Additionally, a single yes/no question ‘The gel aid broke/the tablet separated from the gel aid’ was included when the tablet was administered with the gel. Participants then returned after at least 1 week to repeat the study with the alternative method of administration and were remunerated for their participation. The process within the CRTU is summarised in figure 3. Figure 3 Visit progress flow chart. Outcomes The outcome measures were: (1) the maximum concentration of salicylate; (2) area under the curve (AUC) of the concentration of salicylate (salicylate concentration measured at 0, 20, 40, 60, 90 and 120 min after swallowing); (3) the maximum collagen; (4) the AUC of the collagen; (5) the maximum ASP-arachidonic acid AUC; and (6) the AUC of ASP-arachidonic acid. Collagen AUC and ASP-arachidonic acid AUC are measured at 0, 20, 60 and 120 min after swallowing. Data analysis Analysis was undertaken using the software package SAS 9.4. Data generated from the participant taking SMART Swallowing Aid will be referred to as T (or test) and the data generated from the participant taking the non-gel as R (or reference).

Outcomes The outcome measures were: (1) the maximum concentration of salicylate; (2) area under the curve (AUC) of the concentration of salicylate (salicylate concentration measured at 0, 20, 40, 60, 90 and 120 min after swallowing); (3) the maximum collagen; (4) the AUC of the collagen; (5) the maximum ASP-arachidonic acid AUC; and (6) the AUC of ASP-arachidonic acid. Collagen AUC and ASP-arachidonic acid AUC are measured at 0, 20, 60 and 120 min after swallowing. Data analysis Analysis was undertaken using the software package SAS 9.4. Data generated from the participant taking SMART Swallowing Aid will be referred to as T (or test) and the data generated from the participant taking the non-gel as R (or reference). Derived variables The AUC for aspirin serum concentration was calculated using the trapezoidal rule using the equation: AUC=(C0+C20)2×20+(C20+C40)2×20+(C40+C60)2×20+(C60+C90)2×30+(C90+C120)2×30 where Ct is the aspirin serum concentration at time t. Any missing or incomplete data were replaced using last observation carried forward in order to calculate the AUC. AUC for the platelet function test was calculated in a similar fashion. Cmax for both aspirin serum concentration and platelet function test was calculated as the maximum value of the variables at the observed time points. As an equivalence study, only individuals who adhered to the protocol were included in the analysis.

where Ct is the aspirin serum concentration at time t. Any missing or incomplete data were replaced using last observation carried forward in order to calculate the AUC. AUC for the platelet function test was calculated in a similar fashion. Cmax for both aspirin serum concentration and platelet function test was calculated as the maximum value of the variables at the observed time points. As an equivalence study, only individuals who adhered to the protocol were included in the analysis. Average bioequivalence In order to demonstrate average bioequivalence (ABE), both the mean AUC and the mean Cmax for treatment should not significantly differ from the mean AUC and mean Cmax for control using a two one-sided test procedure. The trial team set the limits that ABE would be demonstrated if the 90% two-sided CI for the mean difference falls within the acceptance limits of –ln(1.25).14 To be precise, if µT denotes the true mean log(AUC) {or log(Cmax)} for those with the test and µR for those with the reference group, ABE is demonstrated if there is good evidence that: 0.8≤expμT-μR≤1.25 The model used to estimate ABE was based on a random effects cross-over model using the Kenward-Roger df.14 Population bioequivalence The level of both population bioequivalence (PBE) was measured using the standard guidelines issued from the Food and Drug Administration and estimated using a mixed model approach to estimate the within-subject variances required for the measured PBE. For PBE, the metric is given by equation whose 90% CI must be below zero.14

l of both population bioequivalence (PBE) was measured using the standard guidelines issued from the Food and Drug Administration and estimated using a mixed model approach to estimate the within-subject variances required for the measured PBE. For PBE, the metric is given by equation whose 90% CI must be below zero.14 For all ABE and PBE and random effects (participants as random), a cross-over model was used to estimate the required parameters. The aggregate measures are functions of the estimated variances and easily calculable from the computer output. It is not possible to estimate the power of this study since there are not enough data on which to base an estimate of the within-subject SD. Although given that SMART Swallowing Aid is unlikely to have a large impact on the AUC or Cmax, the within-subject SD was likely to be small and hence the power was likely to be high. It should be noted, however, that cross-over trials used to assess for PBE generally use in excess of 20–30 subjects so these results may be more exploratory than confirmatory. Questionnaire analysis The ease of swallowing and the comfort of swallowing and taste obtained from the questionnaires were analysed using a non-parametric Mann-Whitney test. Regulatory approvals The trial was registered with EudraCT 2011-005208-14 and ISRCTN 13972867.

It is not possible to estimate the power of this study since there are not enough data on which to base an estimate of the within-subject SD. Although given that SMART Swallowing Aid is unlikely to have a large impact on the AUC or Cmax, the within-subject SD was likely to be small and hence the power was likely to be high. It should be noted, however, that cross-over trials used to assess for PBE generally use in excess of 20–30 subjects so these results may be more exploratory than confirmatory. Questionnaire analysis The ease of swallowing and the comfort of swallowing and taste obtained from the questionnaires were analysed using a non-parametric Mann-Whitney test. Regulatory approvals The trial was registered with EudraCT 2011-005208-14 and ISRCTN 13972867. Results Twelve healthy male volunteers were randomised to receive either ‘Gel/no gel’ or ‘No gel/gel’. All participants who were randomised finished the trial and no participant was lost to follow-up. All participants took the intervention as intended and no adverse events were observed during the course of this trial. The analysis was performed in compliance with the statistical analysis plan, due to the nature of bioequivalence it is necessary to know the treatment (T) and the reference (R) so this analysis was not blinded. Bioequivalence results A summary of the outcome measure at the end of each period is given in table 1. This shows roughly equal means for collagen and ASP-arachidonic acid in each period and each sequence, but a slight difference in salicylate AUC and Cmax. Table 1 Summary statistics of outcome measures Sequence

The analysis was performed in compliance with the statistical analysis plan, due to the nature of bioequivalence it is necessary to know the treatment (T) and the reference (R) so this analysis was not blinded. Bioequivalence results A summary of the outcome measure at the end of each period is given in table 1. This shows roughly equal means for collagen and ASP-arachidonic acid in each period and each sequence, but a slight difference in salicylate AUC and Cmax. Table 1 Summary statistics of outcome measures Sequence No gel/with gel With gel/no gel Period 1 Period 2 Period 1 Period 2 Log AUC collagen 8.98 (0.19) 9.04 (0.19) 8.98 (0.21) 8.86 (0.18) Log Cmax collagen 4.43 (0.13) 4.42 (0.12) 4.37 (0.23) 4.24 (0.11) Log AUC ASP-arachidonic acid 8.80 (0.42) 8.98 (0.33) 8.76 (0.54) 8.63 (0.24) Log Cmax ASP-arachidonic acid 4.66 (0.22) 4.59 (0.17) 4.41 (0.52) 4.38 (0.18) Log AUC salicylate 6.02 (1.07) 6.25 (0.83) 5.61 (1.06) 6.36 (0.66) Log Cmax salicylate 2.24 (0.86) 2.40 (0.74) 1.83 (0.82) 2.55 (0.54) ASP, aspirin; AUC, area under the curve.

P-arachidonic acid 8.80 (0.42) 8.98 (0.33) 8.76 (0.54) 8.63 (0.24) Log Cmax ASP-arachidonic acid 4.66 (0.22) 4.59 (0.17) 4.41 (0.52) 4.38 (0.18) Log AUC salicylate 6.02 (1.07) 6.25 (0.83) 5.61 (1.06) 6.36 (0.66) Log Cmax salicylate 2.24 (0.86) 2.40 (0.74) 1.83 (0.82) 2.55 (0.54) ASP, aspirin; AUC, area under the curve. ABE results are presented in table 2. These have been estimated based on random effects cross-over model using the Kenward-Roger df.14 These show that since the 90% CI does not fall within the (0.80, 1.25) window, ABE cannot be concluded. The change in salicylate concentrations, collagen and ASP-arachidonic acid activity over time is shown in figures 4–6, respectively. The figures show an increase in salicylate levels and corresponding reduction in platelet activity as measured by ASP-arachidonic acid, but limited or no change in platelet activity as measured by collagen. Giving the aspirin with the gel results in lower salicylate concentrations and higher platelet activity. Figure 4 Salicylate concentration (mg/L) over time for uncoated aspirin administered with and without gel. Figure 5 Collagen platelet activity over time for uncoated aspirin administered with and without gel. Figure 6 ASP-Arachidonic acid platelet activity over time for uncoated aspirin administered with and without gel. Table 2 Estimate of ratio of means (with gel/no gel) and 90% CI

Figure 4 Salicylate concentration (mg/L) over time for uncoated aspirin administered with and without gel. Figure 5 Collagen platelet activity over time for uncoated aspirin administered with and without gel. Figure 6 ASP-Arachidonic acid platelet activity over time for uncoated aspirin administered with and without gel. Table 2 Estimate of ratio of means (with gel/no gel) and 90% CI Variable Estimate of mean ratio 90% CI Log AUC collagen 1.09 (0.96 to 1.24) Log Cmax collagen 1.07 (0.99 to 1.16) Log AUC ASP-arachidonic acid 1.16 (0.88 to 1.53)* Log Cmax ASP-arachidonic acid 0.98 (0.79 to 1.22)* Log AUC salicylate 0.77 (0.40 to 1.47)* Log Cmax salicylate 0.76 (0.44 to 1.31)* *Outside of range allowable to enable assumption of bioequivalence. ASP, aspirin; AUC, area under the curve. PBE was estimated using the metric provided in Jones and Kenward (2003, p 343).14 This metric is essentially a combination of the mean difference (squared) and the variances of the treatment response. The important factor is the upper 90% CI bound for the metric. If this is below zero then PBE can be declared. It should be noted that if ABE has not been demonstrated it is unlikely that PBE can be demonstrated. The values for each outcome are given in table 3. These show that T and R are PBE for collagen, but not for ASP-arachidonic acid or salicylate. Table 3 Lower 90% CI bound for PBE metric

PBE was estimated using the metric provided in Jones and Kenward (2003, p 343).14 This metric is essentially a combination of the mean difference (squared) and the variances of the treatment response. The important factor is the upper 90% CI bound for the metric. If this is below zero then PBE can be declared. It should be noted that if ABE has not been demonstrated it is unlikely that PBE can be demonstrated. The values for each outcome are given in table 3. These show that T and R are PBE for collagen, but not for ASP-arachidonic acid or salicylate. Table 3 Lower 90% CI bound for PBE metric Variable Lower bound Log AUC collagen −0.01 Log Cmax collagen −0.02 Log AUC ASP-arachidonic acid 0.19* Log Cmax ASP-arachidonic acid 0.16* Log AUC salicylate 0.54* Log Cmax salicylate 0.44 *Outside of range allowable to enable assumption of bioequivalence. ASP, aspirin; AUC, area under the curve; PBE, population bioequivalence. Questionnaire results The summary of the questionnaire VAS responses and formal comparisons is given in table 4. This shows that there is a significant treatment effect for taste with the gel group having a lower mean score than the no gel group. Out of the 12 individuals, 11 reported that the gel did not break and one reported that ‘first tablet spat out as it separated from gel, second tablet didn't separate.’ Table 4 Summary statistics of questionnaire data

Questionnaire results The summary of the questionnaire VAS responses and formal comparisons is given in table 4. This shows that there is a significant treatment effect for taste with the gel group having a lower mean score than the no gel group. Out of the 12 individuals, 11 reported that the gel did not break and one reported that ‘first tablet spat out as it separated from gel, second tablet didn't separate.’ Table 4 Summary statistics of questionnaire data Sequence P value Carry-over P value No gel/with gel With gel/no gel Period 1 Period 2 Period 1 Period 2 Ease of swallowing 9.55 (0.4) 8.73 (1.59) 7.48 (2.07) 9.05 (1.08) 0.1495 Comfort 9.13 (1.50) 7.75 (1.97) 8.93 (0.77) 9.22 (1.23) 0.1087 Taste 4.73 (4.57) 0.58 (0.62) 1.47 (1.54) 4.42 (4.35) 0.0370 0.5204 Discussion The study successfully recruited to target with no dropout. There were no safety issues identified during the trial. It cannot, however, be stated that bioequivalence was demonstrated when taking aspirin with and without the gel. The salicylate levels and amount absorbed seemed to be greater without the gel, which corresponded with lower platelet function test results. As anticipated, the patients who swallowed aspirin coated in the gel found the taste to be better. Furthermore, although water was not provided with the aspirin tablet when it was coated in the gel no significant effect on ease of swallow or swallow comfort was detected.

onded with lower platelet function test results. As anticipated, the patients who swallowed aspirin coated in the gel found the taste to be better. Furthermore, although water was not provided with the aspirin tablet when it was coated in the gel no significant effect on ease of swallow or swallow comfort was detected. As a randomised controlled cross-over study we were able to use a relatively small sample size to detect reasonable differences in bioequivalence. Although the trial was open label and this was unavoidable due to the nature of what was being tested we would not expect this to adversely affect clinical biomarkers. We chose to give Weetabix prior to administration of the tablet with and without gel as the trial occurred in the morning and this is a common breakfast within the UK. As a randomised control cross-over trial any effect that it may have itself on absorption will have been taken into account. While we followed a previously reported process for undertaking an evaluation to assess salicylate levels after dose, the results show that we did not capture the full pharmacokinetic profile.11 Tmax for normal coated aspirin is 2 hours15 and therefore considering the results seen, either the Cmax or Tmax is unlikely to be the same for aspirin when given with or without gel.

taking an evaluation to assess salicylate levels after dose, the results show that we did not capture the full pharmacokinetic profile.11 Tmax for normal coated aspirin is 2 hours15 and therefore considering the results seen, either the Cmax or Tmax is unlikely to be the same for aspirin when given with or without gel. The results suggest therefore that the gel may be reducing aspirin absorption, resulting in increased platelet activity. Our in vitro test for bioequivalence identified an average difference of 6 s in disintegration time and this we did not deem sufficient to significantly affect bioequivalence. The encapsulation may, however, be delaying the release of aspirin in vivo more than that identified in vitro and a longer follow-up time may have resulted in similar total aspirin release. An alternative hypothesis may be that the water content in the gel, which is in close contact with the uncoated tablet, may be hydrolysing the aspirin and therefore making less available for absorption and metabolism to salicylate in the first instance. Finally, an ingredient within the gel itself may be incompatible with aspirin and either physically or chemically interacting with the molecule to reduce absorption. The ingredients were selected due to their largely inert nature, however this may not be the case with this specific molecule.

late in the first instance. Finally, an ingredient within the gel itself may be incompatible with aspirin and either physically or chemically interacting with the molecule to reduce absorption. The ingredients were selected due to their largely inert nature, however this may not be the case with this specific molecule. While one option maybe to test the gel with different medicines which are known to be less susceptible to water or less likely to break down before the gel is absorbed, we would suggest that it would be more efficient to reformulate the gel to ensure that it is unlikely to affect absorption of all medicines. The finding that reductions in aspirin levels were related to increased platelet function as measured by arachidonic acid is to be expected as this is seen as an appropriate measure for this purpose16 whereas platelet binding to a collagen surface has been shown to be less sensitive to changes in aspirin levels.17 It could be argued that we selected a chemical entity and formulation which was more likely to be affected by the encapsulation and therefore this may not have provided an ideal model. However, aspirin was selected because it is frequently prescribed in individuals with dysphagia and due to it being very likely to react with ingredients and therefore a positive result would have provided greater reassurance regarding the inert nature of the gel.

nd therefore this may not have provided an ideal model. However, aspirin was selected because it is frequently prescribed in individuals with dysphagia and due to it being very likely to react with ingredients and therefore a positive result would have provided greater reassurance regarding the inert nature of the gel. The results bring into question the fact that other similar products have been allowed to be marketed to be used to ease medication administration without in vivo testing. Recent research has shown that mixing medicines’ food thickeners significantly reduces their effectiveness18 19 and although the extent of interaction and contact with swallowing aids and medicines is less, it seems that the likelihood of interaction needs to be explored more extensively prior to award of marketing authorisation. We therefore have to conclude that in using the gel to enable aspirin to be swallowed without either crushing or the addition of water we cannot assume bioequivalence and therefore further research is required to determine the exact reason for this result and to develop a product, which is less likely to affect bioequivalence.

fore have to conclude that in using the gel to enable aspirin to be swallowed without either crushing or the addition of water we cannot assume bioequivalence and therefore further research is required to determine the exact reason for this result and to develop a product, which is less likely to affect bioequivalence. Contributors: DJW and DQMC originated the idea for the gel, were coapplicants for the MRC funding and were responsible for managing the project. GM and SOW were responsible for gel development and testing in vitro. AB and VM were responsible for gel testing in vivo. AC provided statistical input at all stages of the study. AB, VM, SOW, DJW and JFP were responsible for development of the clinical trial protocol and JFP was the principal investigator responsible for management of the trial at the site. All authors contributed to the writing of the article and provided feedback on the different iterations. Funding: The study was funded by the Medical Research Council ‘Developmental Pathway Funding Scheme’ (grant reference: G902184). Competing interests: DJW undertakes consultancy work for and currently holds a research grant from Rosemont Pharmaceuticals who manufactures liquid medicines. Patient consent for publication: Not required Ethics approval: MHRA Device (CI/2012/0030 SMART Swallowing Aid) and NHS Research Ethics Committee approvals (Cambridge Central NRES Committee, East of England 14/EE/0156) were obtained prior to commencement of the study. Provenance and peer review: Not commissioned; externally peer reviewed.

Patient consent for publication: Not required Ethics approval: MHRA Device (CI/2012/0030 SMART Swallowing Aid) and NHS Research Ethics Committee approvals (Cambridge Central NRES Committee, East of England 14/EE/0156) were obtained prior to commencement of the study. Provenance and peer review: Not commissioned; externally peer reviewed. Data sharing statement: Data are available upon reasonable request.

In situ simulations are conducted in real clinical environments to allow healthcare teams to practice teamwork and communication, and clinical management skills, in a team's usual work setting with typically available resources and equipment. Such training improves reliability and safety of high-risk situations, allows for experiential learning, and improves the ability to identify and address latent threats and systems issues. Recent programmes and studies have demonstrated that in situ simulations improve unit safety culture, patient outcomes, knowledge, practical skills, communication and team performance, and reduce adverse event costs. Although multiple benefits of in situ simulations exist, support to individual hospitals and units for implementing in situ simulation has been limited. The purpose of this documentary-style video is to demonstrate how to plan and conduct in situ simulation training, with particular emphasis on how such training can be used to improve communication and teamwork. The video features an in situ simulation conducted at a labour and delivery unit. In addition to instructional content, most of the scenes show unscripted, yet authentic footage, of simulation facilitators and a healthcare team planning, running and debriefing a simulation, and demonstrate the use of visual media during the simulation debriefing—a process that provides a forum for shared learning experiences.

n to instructional content, most of the scenes show unscripted, yet authentic footage, of simulation facilitators and a healthcare team planning, running and debriefing a simulation, and demonstrate the use of visual media during the simulation debriefing—a process that provides a forum for shared learning experiences. This video will facilitate the implementation and routine use of in situ simulations, as it allows prospective users to visualise techniques difficult to convey in other media forms and training modalities. The video is part of a toolkit that will become available in late 2016. More information on the toolkit is available on the Agency for Healthcare Research and Quality website (http://www.ahrq.gov/professionals/quality-patient-safety/hais/tools/perinatal-care/index.html). 10.1136/bmjinnov-2015-000039.video01BMJ Journals Video Playerbmjinnov2015000039media1

The video is part of a toolkit that will become available in late 2016. More information on the toolkit is available on the Agency for Healthcare Research and Quality website (http://www.ahrq.gov/professionals/quality-patient-safety/hais/tools/perinatal-care/index.html). 10.1136/bmjinnov-2015-000039.video01BMJ Journals Video Playerbmjinnov2015000039media1 Contributors: AS was involved in the substantial contributions to the conception and design of the work; drafting the work and revising it critically for intellectual content; and final approval of the version to be published; accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. JP, JB, SD, KM, TI and LK were involved in the substantial contributions to the conception and design of the work; drafting the work and revising it critically for intellectual content; and final approval of the version to be published. SR was involved in the drafting the work and revising it critically for intellectual content and final approval of the version to be published. Competing interests: None declared. Funding: Funding for this work was provided by the Agency for Healthcare Research and Quality Contract No. HHSA290201000024I Task Order 3. Provenance and peer review: Not commissioned; internally peer reviewed.

The first smartphone was launched only 8 years ago and already in that time enormous advances have been made in mobile phone technology. It is now possible to transmit a myriad of data forms at the touch of a button, such as pictures, videos and Global Position System coordinates. It is estimated that, in the UK, approximately 72% of the population and 98.9% of doctors currently own a smartphone, with numbers continuing to rise.1 In hospitals, communication between doctors is ‘formally’ conducted via pagers. First introduced in the 1950s, the system is out-dated and impractical by today's standards. Contacting another doctor involves locating a landline, dialling their four-digit bleep number along with that phone extension, and then anchoring yourself to the phone and waiting for it to ring. In the meantime, you will be unable to continue working and will discourage others from using the line for fear of blocking the return call. On the receiving end, the message arrives with an ear-piercing alert and the four-digit extension. The communication is one-way and lacks useful information—namely, who is trying to contact you, for what purpose and with what level of urgency. Needless to say, this system has flaws, the most important of which is time wastage. Research suggests that a doctor who is frequently paged can spend up to a fifth of his or her time going through this rigmarole.2

useful information—namely, who is trying to contact you, for what purpose and with what level of urgency. Needless to say, this system has flaws, the most important of which is time wastage. Research suggests that a doctor who is frequently paged can spend up to a fifth of his or her time going through this rigmarole.2 Unsurprisingly then, the rising capability of smartphones is rendering this archaic system redundant. Latest research suggests that an astonishing number of doctors are now using their mobiles to communicate clinical information about patients at work: 64.7% are using SMS messaging and 33.1% are using equivalent web-based applications such as WhatsApp.1 This is a problem. National Health Service England has formal guidance on information governance, and on the use of WhatsApp in particular, stating that ‘whatever the other merits of WhatsApp, it should never be used for the sending of information in the professional healthcare environment’.3 The risk is, of course, around data security and therefore patient confidentiality. The public psyche is particularly tuned in to these issues, now more than ever, with incidents such as the recent TalkTalk data hacking scandal reducing trust in data security. This is clearly an issue that needs to be tackled with expediency.

k is, of course, around data security and therefore patient confidentiality. The public psyche is particularly tuned in to these issues, now more than ever, with incidents such as the recent TalkTalk data hacking scandal reducing trust in data security. This is clearly an issue that needs to be tackled with expediency. With these concerns in mind one might be forgiven for wondering what the advantages actually are in using such a system. We carried out a study in which we asked 40 junior doctors about their use of WhatsApp. The study produced two key findings. First, WhatsApp is used more by surgeons than by medics: as a percentage of their total day-to-day communication, it was 30% and 11%, respectively. We broke down what was being communicated and found that over a third of all WhatsApp messages consisted of advice or support. WhatsApp is performing an essential function, providing juniors whose teams are fragmented between theatres and wards with a hotline to senior opinion.

cation, it was 30% and 11%, respectively. We broke down what was being communicated and found that over a third of all WhatsApp messages consisted of advice or support. WhatsApp is performing an essential function, providing juniors whose teams are fragmented between theatres and wards with a hotline to senior opinion. The second interesting result was around doctors' perception of WhatsApp. It was felt to be a good thing in general by 72.5%, with 80% reporting that it improves the relationship between grades and breaks down the traditional hierarchies that can stunt effective communication within a team. The majority of doctors using WhatsApp form a ‘group’ (within the app), consisting of members of the clinical team, so when a message is sent it is visible to all. This group-chat dynamic not only improves team cohesion through informal conversation but it enables juniors to contact seniors more easily where they may not previously have felt able to phone them directly. It is clear that doctors will continue to use methods of communication that are most useful and efficient in practice. Clinical teams require either a bespoke system that can evolve to keep pace with technology while solving the issue of information governance, or a clearer set of regularly updated guidelines on appropriate use of web applications such as WhatsApp and its future iterations. Competing interests: None declared. Provenance and peer review: Not commissioned; externally peer reviewed.

Introduction Hospital readmissions numbered 2.3 million (19%) among Medicare enrolees within 30 days of a previous discharge, generating over $17 billion in healthcare costs in 2004.1 Up to $12 billion of these costs (70%) have been attributed to preventable rehospitalisations.1 With over 13 million hospitalisations driving $102 billion in healthcare costs in 2004, rates of rehospitalisation have become a measure of keen interest.2 In this era of healthcare reform, with increasing emphasis on higher quality healthcare at lower costs, the importance of care coordination during and after hospitalisation has been magnified.

lisations driving $102 billion in healthcare costs in 2004, rates of rehospitalisation have become a measure of keen interest.2 In this era of healthcare reform, with increasing emphasis on higher quality healthcare at lower costs, the importance of care coordination during and after hospitalisation has been magnified. Efforts to improve clinical management and reduce overall utilisation of healthcare have become a nearly universal initiative as the dictums of the Affordable Care Act become a reality. The goal of reducing hospital readmissions within 30 days of a previous discharge is a metric that has captured the attention of healthcare administrators, policymakers and providers. Factors driving readmission rates appear to be complex. Current data suggest that only 50% of the 2.3 million Medicare enrolees readmitted within 30 days in 2004 were seen by primary care providers in the interim and only 38% of heart failure patients are seen within 1 week of discharge.1 3 While a reasonable goal for an index margin of reduction remains undefined, a number of interventions focused on the risk predictive tools, discharge planning and postdischarge care have been introduced.4 Although many interventions demonstrate limited effect,3 there is evidence that educational tools and coordination of timely postdischarge care for patients can effectively reduce readmission rates.5–7

f interventions focused on the risk predictive tools, discharge planning and postdischarge care have been introduced.4 Although many interventions demonstrate limited effect,3 there is evidence that educational tools and coordination of timely postdischarge care for patients can effectively reduce readmission rates.5–7 The STate Action on Avoidable Rehospitalizations (STAAR) programme was launched by the Institute for Healthcare Improvement (IHI) with funding from the Commonwealth Fund in May 2009. The primary goal of this collaboration was to share best practices to improve transitions in care and reduce hospital readmissions. Twenty sites were identified in Massachusetts and this grew to 80 sites across multiple states by August 2013. At Massachusetts General Hospital (MGH), the STAAR initiative (October 2009 to September 2013) included three inpatient care units. This discussion will focus on the changes on one particular high-volume general medical unit averaging 180 monthly admissions. This quality improvement initiative looked to answer the questions of whether nursing and pharmacy interventions can improve care delivery in high-risk populations, and reduce hospital readmissions in a busy medical unit. Methods Owing to the non-invasive nature of implementing nursing and pharmacy interventions, no ethical issues were identified. Minimal concerns were raised regarding privacy or protection of participants. There were no identified potential author conflicts of interest.

The STate Action on Avoidable Rehospitalizations (STAAR) programme was launched by the Institute for Healthcare Improvement (IHI) with funding from the Commonwealth Fund in May 2009. The primary goal of this collaboration was to share best practices to improve transitions in care and reduce hospital readmissions. Twenty sites were identified in Massachusetts and this grew to 80 sites across multiple states by August 2013. At Massachusetts General Hospital (MGH), the STAAR initiative (October 2009 to September 2013) included three inpatient care units. This discussion will focus on the changes on one particular high-volume general medical unit averaging 180 monthly admissions. This quality improvement initiative looked to answer the questions of whether nursing and pharmacy interventions can improve care delivery in high-risk populations, and reduce hospital readmissions in a busy medical unit. Methods Owing to the non-invasive nature of implementing nursing and pharmacy interventions, no ethical issues were identified. Minimal concerns were raised regarding privacy or protection of participants. There were no identified potential author conflicts of interest. Given the volume of patient care occurring in the identified unit, this setting was chosen as a potentially beneficial place for a quality improvement effort. Prior to initiation, the expectation was that readmission rates would decrease and the overall quality of care would improve.

Methods Owing to the non-invasive nature of implementing nursing and pharmacy interventions, no ethical issues were identified. Minimal concerns were raised regarding privacy or protection of participants. There were no identified potential author conflicts of interest. Given the volume of patient care occurring in the identified unit, this setting was chosen as a potentially beneficial place for a quality improvement effort. Prior to initiation, the expectation was that readmission rates would decrease and the overall quality of care would improve. Study design This was a before–after study to determine the impact of D/C RN and Transitional Care Pharmacist (TC PharmD) interventions on readmission rates among patients at increased risk for readmission. The design rationale was based on the IHI recommended Plan-Do-Study-Act (PDSA) cycles with iterative episodes of change. Readmission rates were tracked in statistical process control charts during the study as were the participants in the study along with their primary reason for hospitalisation. Eligibility criteria Eligibility criteria were constructed using two domains (table 1): (1) criteria established in Project Red8 and (2) most common admission diagnoses associated with unit-specific readmissions. All patients admitted to the unit during the project interval (September 2009 to October 2013) who met criteria were categorised as high risk for readmission and identified for interventions. Minor changes to the eligibility criteria occurred during the project using the PDSA cycles. Patient characteristics are listed in table 2.

s admitted to the unit during the project interval (September 2009 to October 2013) who met criteria were categorised as high risk for readmission and identified for interventions. Minor changes to the eligibility criteria occurred during the project using the PDSA cycles. Patient characteristics are listed in table 2. Table 1 Inclusion/exclusion criteria Inclusion criteria Exclusion criteria Pneumonia <60 years old Atrial fibrillation Non-English speaking Altered mental status Left against medical advice Dehydration Transferred to the ICU Acute renal failure Urinary tract infection Cancer pain >10 Medications ICU, intensive care unit. Table 2 Patient characteristics Characteristic Intervention patients Non-intervention patients p Value Age (mean in years) 72.5 51 <0.01 Women (n, %) 1152 (0.44) 2983 (0.51) <0.01 Heart failure (n, %) 419 (0.16) 578 (0.1) <0.01 Pneumonia (n, %) 393 (0.15) 645 (0.13) <0.01 Atrial fibrillation (n, %) 340 (0.13) 446 (0.09) <0.01 Altered mental status (n, %) 314 (0.12 794 (0.16) <0.01 Dehydration (n, %) 314 (0.12) 844 (0.17) <0.01 Acute renal failure (n, %) 209 (0.08) 695 (0.14) <0.01 Urinary tract infection (n, %) 209 (0.08) 297 (0.06) <0.01 Cancer pain (n, %) 131 (0.05) 99 (0.02) <0.01 >10 Medications (n, %) 288 (0.11) NA NA NA, not applicable.

tered mental status (n, %) 314 (0.12 794 (0.16) <0.01 Dehydration (n, %) 314 (0.12) 844 (0.17) <0.01 Acute renal failure (n, %) 209 (0.08) 695 (0.14) <0.01 Urinary tract infection (n, %) 209 (0.08) 297 (0.06) <0.01 Cancer pain (n, %) 131 (0.05) 99 (0.02) <0.01 >10 Medications (n, %) 288 (0.11) NA NA NA, not applicable. Sample size Assuming an initial event rate of 21% for the primary outcome and an α error of 0.05, a pool of approximately 983 patients was estimated to provide 85% power to detect a 30% reduction within a 2.5% margin of error. The initiative continued throughout the entire 4-year time period to optimise results. There were 2620 patients admitted to the unit who met high risk for readmission criteria. Of these, all 2620 patients received nursing interventions and 539 received pharmacy interventions. Statistical methods Statistical process control charts were used to track readmission rates. The χ2 testing was used to differentiate rates of readmission for intervention and non-intervention groups. A p value less than 0.05 was considered to be statistically significant.

Sample size Assuming an initial event rate of 21% for the primary outcome and an α error of 0.05, a pool of approximately 983 patients was estimated to provide 85% power to detect a 30% reduction within a 2.5% margin of error. The initiative continued throughout the entire 4-year time period to optimise results. There were 2620 patients admitted to the unit who met high risk for readmission criteria. Of these, all 2620 patients received nursing interventions and 539 received pharmacy interventions. Statistical methods Statistical process control charts were used to track readmission rates. The χ2 testing was used to differentiate rates of readmission for intervention and non-intervention groups. A p value less than 0.05 was considered to be statistically significant. Intervention Nursing intervention Patients were classified on admission by the D/C RN as high risk for readmission according to pre-established inclusion criteria initially adapted from Project Red.8 These criteria were extended to include common admission diagnoses specific to readmissions to the medicine unit (table 1). Patients meeting criteria were identified by the D/C RN by adding an icon to the unit census board, which identified high-risk patients for the multidisciplinary care team. Daily, the D/C RN and the assigned staff nurse assisted patients in establishing a goal for the day, which was written on the patient's bedside board. The D/C RN provided education to each high-risk patient as well as a folder customised to include important information about their care team, follow-up appointments, treatment plan and individualised education materials. Within the folders, there was a place for the patient, or appropriate learner, to write questions for the care team, which were addressed by the D/C RN the following day.

der customised to include important information about their care team, follow-up appointments, treatment plan and individualised education materials. Within the folders, there was a place for the patient, or appropriate learner, to write questions for the care team, which were addressed by the D/C RN the following day. The D/C RN role included care coordination for all patients in the unit and not just those deemed as high risk. The D/C RN facilitated morning multidisciplinary rounds as well as coordinating daily afternoon huddles with senior house staff and the resource nurse to update the plan of care, and creating consensus regarding the anticipated day of patient discharge. Once a date was identified, this date was posted on the unit census board, and was updated twice daily to facilitate early discharges for the next day. The D/C RN assisted staff with arranging simple postdischarge services and collaborated with case management on more complex postdischarge planning.

patient discharge. Once a date was identified, this date was posted on the unit census board, and was updated twice daily to facilitate early discharges for the next day. The D/C RN assisted staff with arranging simple postdischarge services and collaborated with case management on more complex postdischarge planning. Pharmacy intervention I: postdischarge calls During the project, two pharmacy interventions were piloted. The first intervention was a stand-alone postdischarge call programme that was piloted during the first 6 months of the STAAR initiative. The purpose was to provide medication reconciliation and counselling after a patient was discharged home. After receiving notification of a discharge, the TC PharmD would review the patient medication profile and contact patients within 72 h to reconcile their home meds, provide counselling on new medications and to address adherence issues. The call was then documented and primary physicians were notified of any urgent issues. The pharmacy calls continued from January 2010 to June 2010, at which point findings were shared within the institution. As a result, primary care practices began calling patients postdischarge.

ions and to address adherence issues. The call was then documented and primary physicians were notified of any urgent issues. The pharmacy calls continued from January 2010 to June 2010, at which point findings were shared within the institution. As a result, primary care practices began calling patients postdischarge. Pharmacy intervention II: medication reconciliation and patient education The second pharmacy intervention occurred during hospitalisation and became the focus of the TC PharmD role. The TC PharmD anticipated medication-related changes for patients classified as high-risk for readmission. Partnering closely with the D/C RN, case managers and care team, the TC PharmD worked to resolve any medication-related issues over the course of the admission and confirm patient understanding of their new medication regimen. On the day of discharge, the TC PharmD completed medication reconciliation, provided counselling and addressed patient adherence barriers including facilitating the filling of discharge prescriptions or assisting with prior authorisations. Other interventions Multidisciplinary rounds were modified during this intervention. To assist with establishing the plan for the day, the timing of multidisciplinary rounds was moved and the focus was changed to include barriers to discharge as well as an estimated date of discharge. The electronic application for physicians to complete medication reconciliation was also updated to identify medication errors and provide additional safeguards during patient discharge.

Other interventions Multidisciplinary rounds were modified during this intervention. To assist with establishing the plan for the day, the timing of multidisciplinary rounds was moved and the focus was changed to include barriers to discharge as well as an estimated date of discharge. The electronic application for physicians to complete medication reconciliation was also updated to identify medication errors and provide additional safeguards during patient discharge. Outcomes The outcome measure was all-cause rates of readmission within 30 days. Secondary benefits were identified as a result of this project, but were not measured. All outcomes were tracked by the operations department staff who were blinded to intervention status.

The electronic application for physicians to complete medication reconciliation was also updated to identify medication errors and provide additional safeguards during patient discharge. Outcomes The outcome measure was all-cause rates of readmission within 30 days. Secondary benefits were identified as a result of this project, but were not measured. All outcomes were tracked by the operations department staff who were blinded to intervention status. From October 2009 to September 2013, overall readmission rates decreased by 30% from 21% (18% average) preintervention to 14.5% (14.9% average) postintervention (p<0.05; figure 1). From July 2010 to December 2011, rates of readmission among high-risk patients receiving the D/C RN intervention with or without the TC PharmD intervention decreased to 15.9% (p=0.59). From January 2010 to June 2010, rates of readmission among high-risk patients receiving the TC PharmD postdischarge calls decreased to 12.9% (p=0.55). Postdischarge TC PharmD calls demonstrated that 52% of patients deviated from medication instructions after leaving the hospital (table 3). Deviation from medication instructions included patient continuation of medications that had been discontinued during the hospitalisation, patient continuation of over the counter medications (non-steroidal anti-inflammatory drugs, etc) that were not mentioned during the previous hospitalisation and patient challenges following proper dosing instructions for medications that were initiated or changed on discharge. From July 2010 through December 2011, readmission rates for patients on the medical unit who did not receive the designated D/C RN or TC PharmD interventions decreased to 15.8% (p=0.61) and 16.2% (0.31), respectively. Patient characteristics in intervention and non-intervention groups are listed in table 2. Inpatient discharge experience satisfaction did not change during this study.

ts on the medical unit who did not receive the designated D/C RN or TC PharmD interventions decreased to 15.8% (p=0.61) and 16.2% (0.31), respectively. Patient characteristics in intervention and non-intervention groups are listed in table 2. Inpatient discharge experience satisfaction did not change during this study. Table 3 Transitional Care Pharmacists (TC PharmD) postdischarge calls to patients 11 January to 30 June 2010 Readmissions Discharges Readmission rate (%) Patients called by TC PharmD 13 101 12.9 Patients not called by TC PharmD 91 528 17.2 Total 104 629 NA NA, not applicable. Figure 1 All-cause 30-day readmissions. Discussion The STAAR initiative reduced hospital readmissions by 30% in this general medicine unit during the study period, including patients who received discharge interventions as well as those who did not. The D/C RN and TC PharmD interventions targeting high-risk populations alone did not appear to produce statistically significant changes in readmission rates. Additional impact was realised with the changes made to the care team structure, which affected all unit patients. Targeted interventions along with improved hospital coordination can augment processes of care for patients during and after admission. Implementation of similar interventions with improved care coordination may be an effective means of reducing readmissions in other hospitals.

ucture, which affected all unit patients. Targeted interventions along with improved hospital coordination can augment processes of care for patients during and after admission. Implementation of similar interventions with improved care coordination may be an effective means of reducing readmissions in other hospitals. Prior studies examining the impact of patient education-focused nursing interventions have demonstrated to effectively reduce readmissions in several different settings. Coleman et al9 reported reduced readmission rates with the addition of a nurse coordinator and care transition coach at the time of discharge. Jack et al demonstrated the effects of instituting nursing interventions to improve care transitions and reduce readmissions. Connections of readmission rates to patient health literacy have been described by Cloonan et al.10 Recently, McHugh and Ma's11 correlation of nurse workload with readmission rates among patients with heart failure, acute myocardial infarction and pneumonia, underlines a relationship between nursing inpatient business and readmission rates. Naylor et al's12 systematic review of care transitions describes 21 different adult inpatient interventions to reduce readmissions with emphasis on the impact of nursing-guided patient-centred interventions at the time of discharge. A systematic review by Chiu et al13 highlighted the value of nursing interventions focusing on communication tools, patient involvement, nurse-led coaching, education sessions and comprehensive discharge planning.

ions with emphasis on the impact of nursing-guided patient-centred interventions at the time of discharge. A systematic review by Chiu et al13 highlighted the value of nursing interventions focusing on communication tools, patient involvement, nurse-led coaching, education sessions and comprehensive discharge planning. Studies have established the impact of pharmacist-based interventions on readmission rates and care transitions. Schnipper et al14 demonstrated that postdischarge pharmacy calls reduce adverse drug events after hospitalisation. A pharmacist coaching intervention was noted to reduce readmissions in a study by Coleman et al.9 In elderly populations, pharmacist inpatient counselling paired with outpatient pharmacist home visits has been shown to effectively reduce readmissions.15 Overall readmission rates demonstrated improvement during the MGH STAAR initiative with a 30% rate reduction from ∼21% to 14.5%; however, the isolated nursing and pharmacist interventions did not demonstrate a statistically significant change in readmissions. This may be explained by the level of collaboration and engagement by unit staff, nursing, pharmacists and physicians. Furthermore, having a single point person who can be easily identified by patients and care team served to reduce inefficiencies and further streamline care. In addition, the D/C RN role included leading discussions at multidisciplinary rounds and focused on creating consensus around the estimated day of discharge for all unit patients. This method of engaging providers and broadcasting the care plan is an effective way to improve care coordination. Having support of key stakeholders including unit and institutional leadership allowed the initiative to build momentum that may have not otherwise occurred.

timated day of discharge for all unit patients. This method of engaging providers and broadcasting the care plan is an effective way to improve care coordination. Having support of key stakeholders including unit and institutional leadership allowed the initiative to build momentum that may have not otherwise occurred. Other key interventions that may have contributed to reducing readmission rates for all patients include changes to the electronic medication reconciliation application as well as the timing and structure of house staff and multidisciplinary rounds. These smaller changes had a significant impact on the flow and execution of care delivery in the unit.

have contributed to reducing readmission rates for all patients include changes to the electronic medication reconciliation application as well as the timing and structure of house staff and multidisciplinary rounds. These smaller changes had a significant impact on the flow and execution of care delivery in the unit. Although the goal of reducing unit readmission rates to 10% was not achieved, with the implementation of the D/C RN role, pharmacy medication reconciliation, postdischarge calls and improved care coordination, several unexpected outcomes occurred. During the process of interdisciplinary collaboration, awareness of readmission rates and the importance of reducing readmissions became a top priority. Although unable to quantify the cost savings of a 30% reduction in readmission rates, our belief is that we improved patient care outcomes as defined by The Institute of Medicine. The Institute of Medicine's (IOM) Crossing the Quality Chasm describes improving health as a function of upholding the dictums of safe, effective, patient-centred, timely, efficient and equitable healthcare.16 We believe that the work detailed here holds true to the IOM values and further demonstrates the call to action for healthcare systems in general.

ossing the Quality Chasm describes improving health as a function of upholding the dictums of safe, effective, patient-centred, timely, efficient and equitable healthcare.16 We believe that the work detailed here holds true to the IOM values and further demonstrates the call to action for healthcare systems in general. Limitations Inclusion criteria may have excluded patients who could have benefitted from the interventions; therefore, the D/C RN was afforded the opportunity to include additional individuals through consensus with the unit nurse director and project manager. In addition, the project required staff nurses to take on an additional patient as part of their assignment to create a cost-neutral dedicated D/C RN. A small percentage of a pharmacy resource was dedicated to this project and balanced with other responsibilities. For these reasons, this work may not be possible at some other institutions. All high-risk patients would have benefited from the TC PharmD intervention, however, due to timing of admissions, limited pharmacy resources and other unavoidable logistics, not all of the high-risk patients received both interventions possibly reducing favourable outcomes. Reorganisation of multidisciplinary rounds and other unit-based changes may have contributed to decreased readmissions in patients who were not identified as high risk for readmission in this study. During the fourth year of the project, the D/C RN role was changed to the Attending Nurse Role and expanded to other units in the hospital as a part of an institution-wide initiative. It is unclear how this influenced findings. In 2013, there was an institution-wide effort to increase the number of warm-handoffs to postacute care facilities and the impact on outcomes was unable to be quantified. Routine house staff and attending turnover may have negatively impacted readmission rates despite efforts to bridge care gaps.

s influenced findings. In 2013, there was an institution-wide effort to increase the number of warm-handoffs to postacute care facilities and the impact on outcomes was unable to be quantified. Routine house staff and attending turnover may have negatively impacted readmission rates despite efforts to bridge care gaps. Conclusion Single interventions, when evaluated in isolation, did not demonstrate statistically significant changes in readmission rates. A multidisciplinary approach to improving care coordination effectively reduced avoidable readmissions. This demonstrates the importance of multidisciplinary collaboration, which improved the care transitions and reduced rehospitalisations for our patients, the primary outcome measures of this quality improvement initiative. The authors would like to acknowledge the following individuals who contributed to this work: Theresa Gallivan, RN, Elizabeth Mort, MD, Andrew Karson, MD, Christine Annese, RN, Yanie Jackson, MBA, Gwendolyn Crevensten, MD, Christian Dankers, MD, Sarah Keegan, RN, Elizabeth Bishop, RN, Jenna Delgado, RN, Stefanie Green, RN, Catherine Mullin, RN, Joanne Kaufman, RN MPA, Laura Huber, CCM, Diane Hubbard, ACM, Lisa McDonough, RN, Patricia Durning, RN CM, Deborah Kiely, RN, Ashley Truong, PharmD, Andrea Bonnano, DPT, Regina Neives, RN and Mary Neagle.

n Dankers, MD, Sarah Keegan, RN, Elizabeth Bishop, RN, Jenna Delgado, RN, Stefanie Green, RN, Catherine Mullin, RN, Joanne Kaufman, RN MPA, Laura Huber, CCM, Diane Hubbard, ACM, Lisa McDonough, RN, Patricia Durning, RN CM, Deborah Kiely, RN, Ashley Truong, PharmD, Andrea Bonnano, DPT, Regina Neives, RN and Mary Neagle. Contributors: JAC designed the data collection, monitored data collection for the entire period, analysed data, and drafted and revised the paper. She is the guarantor. LSC, JC, JDP, KH, JS and LAT initiated the collaborative project, designed the data collection, monitored data collection for the entire period, analysed data, drafted and revised the paper, and revised and updated project criteria. JM initiated the collaborative project, designed the data collection, monitored data collection for the entire period, cleaned and analysed data, drafted and revised the paper, and revised and updated project criteria. Competing interests: None declared. Ethics approval: This study obtained Partners Healthcare Institutional Review Board (Chaired by Dr Elizabeth Hohmann, MD) ethical exemption due to the nature of the study and its quality improvement focus. Patients were informed at the time of their admission that efforts to provide optimal care by partnering with nursing and pharmacy resources would occur, and all patients were interested in these services. Provenance and peer review: Not commissioned; internally peer reviewed. Data sharing statement: No additional data are available.

Primary healthcare is key to achieving the Millennium Development Goals and the broader goal of ‘health for all’ by providing accessible, affordable and effective healthcare. Yet in many less developed countries, primary healthcare is constrained by a lack of adequately skilled and supervised health workers—the WHO estimates a global shortage of 4.3 million health workers.1 Through 14 years of innovation, the Knowledge Translation Unit (KTU) of the University Cape Town Lung Institute has found evidence-based ways to strengthen primary care in underserved areas. This article tells the story of that journey, and discusses how this innovation is being spread and might be spread further. The challenges The caseload in primary care has increased because of the growing burden of non-communicable disease (NCD) together with programmes to shift the care of patients with HIV and tuberculosis (TB) out-of-hospital.2 In addition, primary care in underserved areas has been dominated by vertical programmes that only treat patients with TB or HIV. This is no longer an adequate response. Primary care in less developed countries is ill prepared for the pandemic of NCD and mental illness now hitting it. And many patients have not one condition, but many—for example, TB and diabetes or HIV and hypertension, which means that treatments are complex and intricate.

HIV. This is no longer an adequate response. Primary care in less developed countries is ill prepared for the pandemic of NCD and mental illness now hitting it. And many patients have not one condition, but many—for example, TB and diabetes or HIV and hypertension, which means that treatments are complex and intricate. The 2008 Lancet series celebrating the 30th anniversary of the Alma Ata Declaration, which called for primary care to be available to all, made clear the need for strengthening primary care, building on the success of vertical programmes for priority conditions.3 It also noted that good primary care depends on the successful implementation of high-quality evidence-informed guidelines.4 In reality, primary care health workers in less developed settings are provided with guidelines that are long, dense, contain conflicting recommendations, are out of date, and do not help the patients who have multiple conditions. Plus, we have long known that simply sending guidelines to healthcare workers does not on its own improve performance; more is needed.5

s developed settings are provided with guidelines that are long, dense, contain conflicting recommendations, are out of date, and do not help the patients who have multiple conditions. Plus, we have long known that simply sending guidelines to healthcare workers does not on its own improve performance; more is needed.5 Two decades ago, the WHO revolutionised the care of young children with its Integrated Management of Childhood Illness (IMCI) strategy, which provided standardised treatment algorithms for serious conditions based on symptoms.6 7 The strategy has been implemented in over 100 less-developed countries, and evaluations have shown that it improved the quality of care and reduced costs.8–11 In order to reduce deaths, the programme needed to be implemented at scale, but scale up was constrained by dependence on intensive 2 week offsite training that often led to only one health worker in each facility trained to manage young children; the health worker being away also puts strain on the service. Nevertheless, IMCI remains the standard approach to training health workers managing children in resource constrained settings.

e on intensive 2 week offsite training that often led to only one health worker in each facility trained to manage young children; the health worker being away also puts strain on the service. Nevertheless, IMCI remains the standard approach to training health workers managing children in resource constrained settings. The WHO embarked on the development of an equivalent strategy for adults, first as the Practical Approach to Lung Health (PAL)12 and later as the Integrated Management of Adult and Adolescent Illness (IMAI).13 However, IMAI arrived just as major campaigns were being launched to scale up treatment of patients with HIV and so took on an HIV focus. Its implementation has been limited to a few districts in countries with high rates of HIV and has not been evaluated in the overarching way that IMCI has. The recently developed WHO Package of Essential NCD (PEN) interventions for primary healthcare in less-developed settings includes guidelines on diabetes, chronic respiratory disease, cancer, heart disease and stroke, and is currently being piloted in several countries, as yet without a published evaluation.14 Table 1 summarises the key features of currently available clinical management aids designed to standardise the approach to priority conditions.6 12 13–21 Table 1 Key features of currently available clinical management aids designed to standardise an approach to priority conditions

The WHO embarked on the development of an equivalent strategy for adults, first as the Practical Approach to Lung Health (PAL)12 and later as the Integrated Management of Adult and Adolescent Illness (IMAI).13 However, IMAI arrived just as major campaigns were being launched to scale up treatment of patients with HIV and so took on an HIV focus. Its implementation has been limited to a few districts in countries with high rates of HIV and has not been evaluated in the overarching way that IMCI has. The recently developed WHO Package of Essential NCD (PEN) interventions for primary healthcare in less-developed settings includes guidelines on diabetes, chronic respiratory disease, cancer, heart disease and stroke, and is currently being piloted in several countries, as yet without a published evaluation.14 Table 1 summarises the key features of currently available clinical management aids designed to standardise the approach to priority conditions.6 12 13–21 Table 1 Key features of currently available clinical management aids designed to standardise an approach to priority conditions Management aid Integrated acute and chronic conditions (NCDs, IDs, MH) Symptom based Evidence based Less developed setting Primary healthcare based User friendly Comment IMCI6 ✓ ✓ ✓ ✓ Limited to childhood illnesses IMAI13 ✓ ✓ ✓ ✓ ✓ ▸ HIV focused ▸ Limited coverage of chronic conditions ▸ Modular PEN Interventions for Primary Health Care in Low-Resource Settings14 ✓ ✓ ✓ ▸ Text format ▸ Limited to NCDs ▸ Focus on HIV IMPAC15 ✓ ✓ ✓ ✓ ✓ ▸ Limited to obstetric and newborn care ▸ Modular mhGAP Intervention Guide16 ✓ ✓ ✓ ✓ ✓ Limited to mental, neurological and substance use disorders TEAMcare intervention manual17 ✓ ✓ ✓ ▸ Integrated to depression, diabetes and coronary heart disease ▸ Developed settings FNIHB Clinical Practice Guidelines for Nurses in Primary Care18 ✓ ✓ ✓ ▸ Text format ▸ Modules Queensland Government PCCM19 ✓ ✓ ✓ ✓ Text format New Zealand Primary Care Handbook 201220 ✓ ✓ ✓ ▸ Developed setting ▸ Text format with some algorithms FNIHB, first Nations and inuit health branch; IDs, infectious diseases; IMAI, Integrated Management of Adult and Adolescent Illness; IMCI, Integrated Management of Childhood Illness; MH, mental health; mhGAp, Mental Health Gap Action Programme; NCDs, Non-communicable diseases; PCCM, primary clinical care manual; PEN, package of essential Non-communicable disease.

IDs, infectious diseases; IMAI, Integrated Management of Adult and Adolescent Illness; IMCI, Integrated Management of Childhood Illness; MH, mental health; mhGAp, Mental Health Gap Action Programme; NCDs, Non-communicable diseases; PCCM, primary clinical care manual; PEN, package of essential Non-communicable disease. So it remains the case that primary care health workers in less developed countries face a burgeoning caseload without the tools they need to manage adult patients. Developing a package to improve primary care in South Africa Over the past 14 years KTU has developed, tested, expanded, refined and implemented an integrated approach to common priority conditions in primary care throughout South Africa. PACK (Practical Approach to Care Kit) Adult, the latest iteration of this work, covers all significant and common presentations of adults in primary care in South Africa, and has four components: policy compliant and evidence-informed guidelines; onsite, team and case-based training; non-physician prescribing and a cascade system for jurisdiction-wide scaling up.

the latest iteration of this work, covers all significant and common presentations of adults in primary care in South Africa, and has four components: policy compliant and evidence-informed guidelines; onsite, team and case-based training; non-physician prescribing and a cascade system for jurisdiction-wide scaling up. The first iteration of the package (PALSA; PAL in South Africa) was developed to address adults presenting to first-line facilities with cough or difficulty breathing, among whom the index of suspicion for a diagnosis of TB or another acute or chronic respiratory disease should be high. It was developed to address the already high incidence rate of adult TB in South Africa and the delay in diagnosis. A cluster randomised trial of 40 clinics with over 200 nurses showed modest improvements in quality of care, and substantial improvements in detection of cases of TB, confirming the WHO's reasoning that improved management of the symptomatic patient was necessary to augment passive case detection of TB.21

in diagnosis. A cluster randomised trial of 40 clinics with over 200 nurses showed modest improvements in quality of care, and substantial improvements in detection of cases of TB, confirming the WHO's reasoning that improved management of the symptomatic patient was necessary to augment passive case detection of TB.21 The next iteration of the programme, known as PALSA PLUS, was developed in response to South Africa's decision to implement a public sector antiretroviral treatment (ART) programme. It extended the training of clinic nurses to include HIV/AIDS screening, and referral to physicians for diagnosis and initial prescribing of ART, with patients returning to nurses for monitoring. A second cluster randomised trial of this wider programme again confirmed modest improvements in the quality of care, and a substantial additive positive impact on case detection of TB.22 There was also a surprising improvement in successful treatment outcomes for retreatment patients with TB, suggesting that the training might have had a positive impact on nurse–patient relationships.22

med modest improvements in the quality of care, and a substantial additive positive impact on case detection of TB.22 There was also a surprising improvement in successful treatment outcomes for retreatment patients with TB, suggesting that the training might have had a positive impact on nurse–patient relationships.22 As the South African HIV/AIDS treatment programme scaled up, it became clear that the physician initiation of ART (mostly in hospital or during once a week facility visits) was a severe bottleneck preventing access to early treatment.23 So the KTU expanded the scope of the PALSA PLUS package to support Nurse Initiated and Managed ART (NIMART). A third cluster randomised trial found that the intervention supporting nurse-led care of HIV/AIDS resulted in patients being managed as effectively as in physician-led programmes, and again showed modest improvements in quality of care and outcomes for patients,24 a rare finding in real world evaluations of health system innovations.5 Qualitative evaluations alongside these trials repeatedly showed that front line clinic staff felt empowered25 26 by their training, setting to rest a fear that responsibility for clinical diagnosis and treatment would be overpowering and result in burnout. PALSA PLUS has now been scaled up across South Africa as part of a 2010 push to provide ART in each of the country's 3500 primary care clinics. For the first time, mortality has begun to fall, and life expectancy rose from 54 years in 2005 to 60 years in 2012.27 Today, South Africa has the world's largest ART programme with an estimated 2 million people on treatment, many initiated by nurses.28

to provide ART in each of the country's 3500 primary care clinics. For the first time, mortality has begun to fall, and life expectancy rose from 54 years in 2005 to 60 years in 2012.27 Today, South Africa has the world's largest ART programme with an estimated 2 million people on treatment, many initiated by nurses.28 PACK Adult The PACK Adult guideline covers 40 common symptoms and 20 chronic conditions seen among adults attending primary care clinics in South Africa29 (see online supplementary appendix 1), and is an expansion of PALSA PLUS to include NCD (hypertension, diabetes, cardiovascular disease), mental health, end-of-life care and women's health (antenatal care, contraception). A concise tool, it uses symptom-based algorithms as its entry point and a standardised checklist format to assist health workers to ‘Assess, Advise and Treat’ patients' chronic conditions. Usability has steadily improved by incorporating feedback from users and the qualitative evaluations. The guideline prompts users at appropriate opportunities to consider diagnosis of a priority condition, speeding initiation of routine care for that condition with criteria and paths for referral where appropriate.

ns. Usability has steadily improved by incorporating feedback from users and the qualitative evaluations. The guideline prompts users at appropriate opportunities to consider diagnosis of a priority condition, speeding initiation of routine care for that condition with criteria and paths for referral where appropriate. PACK Adult is revised every year to remain current with new evidence and to stay aligned with fast changing national policies and programmes. Compliance with local medication lists and regulations on prescribing is critical, and experience gained from initial implementation has been used to lobby for greater access to essential medications. Updating of guidelines is a global problem, and KTU has worked hard to develop sustainable methods of updating. It has now formed a partnership with BMJ that provides access to the evidence updates that the BMJ produces every month. The guideline's relevance and being up-to-date now makes PACK Adult the predominant knowledge resource for local health workers, their managers and educators. PALSA PLUS was used as the Department of Health training vehicle to pilot NIMART in South Africa,30 and, more recently, successful PACK pilot implementation has led to nurses increasingly being allowed to initiate and titrate medication for hypertension, diabetes and other NCDs. The PACK Adult training programme supporting the guideline is the culmination of repeated refinements, four large pragmatic trials of implementation, the scale up of PALSA PLUS throughout South Africa and a pilot in three of South Africa's health districts.

PACK Adult is revised every year to remain current with new evidence and to stay aligned with fast changing national policies and programmes. Compliance with local medication lists and regulations on prescribing is critical, and experience gained from initial implementation has been used to lobby for greater access to essential medications. Updating of guidelines is a global problem, and KTU has worked hard to develop sustainable methods of updating. It has now formed a partnership with BMJ that provides access to the evidence updates that the BMJ produces every month. The guideline's relevance and being up-to-date now makes PACK Adult the predominant knowledge resource for local health workers, their managers and educators. PALSA PLUS was used as the Department of Health training vehicle to pilot NIMART in South Africa,30 and, more recently, successful PACK pilot implementation has led to nurses increasingly being allowed to initiate and titrate medication for hypertension, diabetes and other NCDs. The PACK Adult training programme supporting the guideline is the culmination of repeated refinements, four large pragmatic trials of implementation, the scale up of PALSA PLUS throughout South Africa and a pilot in three of South Africa's health districts. The training programme is based on educational outreach, or repeated brief visits by clinicians trained as educators to the clinics where they facilitate information, encouragement and participatory education for health professional teams to embed the use of PACK Adult in daily practice.31

The PACK Adult training programme supporting the guideline is the culmination of repeated refinements, four large pragmatic trials of implementation, the scale up of PALSA PLUS throughout South Africa and a pilot in three of South Africa's health districts. The training programme is based on educational outreach, or repeated brief visits by clinicians trained as educators to the clinics where they facilitate information, encouragement and participatory education for health professional teams to embed the use of PACK Adult in daily practice.31 PACK Adult, as in the preceding programmes, is being scaled up in South Africa using a cascade system: master trainers based in provincial government training centres are selected by health department leaders and the KTU to be equipped as master trainers during a 4- day residential workshop. They are then supported to plan implementation at district level, identify suitable facility outreach trainers, conduct workshops for these outreach trainers, and then monitor and support them as they go out to provide training in clinics. Since 2007 the KTU has trained 74 master trainers who have equipped 1430 facility outreach trainers to train almost 20 000 health workers at more than 2000 South African primary care facilities. This has been the largest and most rapid scale up of health worker training in the country's history. Over 260 000 copies of the PALSA PLUS and PACK guidelines have been distributed, signalling the enduring demand for the programme.

n almost 20 000 health workers at more than 2000 South African primary care facilities. This has been the largest and most rapid scale up of health worker training in the country's history. Over 260 000 copies of the PALSA PLUS and PACK guidelines have been distributed, signalling the enduring demand for the programme. The curriculum for the educational outreach sessions comprises cases designed to create familiarity with the guideline and to highlight content where practice needs to change. The facility outreach trainers initially facilitate 8 to 12 60–90 min training sessions to the clinic team, thus avoiding information overload, and allowing health workers to test out the guideline and discuss their experience of it at the next session. Soon the nurses are bringing their own real-life cases to their training sessions. The trainers do not instruct the nurses but, rather, help them to learn from the guidelines and each other. The length of this initial training is flexible, to ensure that required competence levels are reached by all nurses, with completion marked by a certificate. Follow-up training varies from monthly sessions to multiple visits and will provide updates, address local requests, align with national initiatives such as TB and HIV awareness days, or provide training for new staff. A hybrid training model to target new staff members combining subdistrict-based sessions with clinic-based outreach has also been developed.

sessions to multiple visits and will provide updates, address local requests, align with national initiatives such as TB and HIV awareness days, or provide training for new staff. A hybrid training model to target new staff members combining subdistrict-based sessions with clinic-based outreach has also been developed. This training strategy is a significant departure from usual training programmes, which take individual health workers out of their facilities, away from their own setting and patients, and rarely provide follow-up support. The instructing of trainers also differs, paying equal importance to skills for their role as educator and to clinical content; this ensures that trainers facilitate the nurses' own learning and use of the guidelines rather than simply offering them didactic content. The strategy has influenced training across the public health sector in South Africa.

differs, paying equal importance to skills for their role as educator and to clinical content; this ensures that trainers facilitate the nurses' own learning and use of the guidelines rather than simply offering them didactic content. The strategy has influenced training across the public health sector in South Africa. Once trained, nurses who previously were permitted only to prescribe a limited number of medications are granted prescribing permission for the larger number of medications necessary to control priority chronic diseases such as HIV, hypertension and diabetes. In the case of HIV, this is followed by a mentoring programme where an experienced prescriber, either a physician or nurse clinician, is partnered with a nurse who is starting to prescribe. This relationship means that the new prescriber has direct access to a named person for troubleshooting and queries, and allows for an assessment of competency necessary to meet local regulations. The KTU is working with provincial health departments in South Africa to expand this model to NCD. Ensuring all nurses are competent prescribers is important for making optimal use of the large state funded system of primary care clinics in South Africa, where nurses see 9 of 10 patients without physician input.30

ons. The KTU is working with provincial health departments in South Africa to expand this model to NCD. Ensuring all nurses are competent prescribers is important for making optimal use of the large state funded system of primary care clinics in South Africa, where nurses see 9 of 10 patients without physician input.30 PACK Adult is being implemented in the 11 pilot districts identified for health system strengthening as part of plans to introduce National Health Insurance in South Africa over the next 15 years.32 This follows PACK Adult's successful pilot in four districts during 2012, and the political profile it received when presented as one of 11 examples of innovation for NCD at the August 2012 Global Health Policy Summit in London.33 The South African Deputy Minister of Health, who attended that meeting, has since championed the introduction of the programme in these pilot districts, even travelling out to clinics to see it in action. PACK Adult is now being used to train medical and nursing students through the Medical Education Partnership Initiative (MEPI). The programme at Stellenbosch University prepares medical students for primary care using PACK Adult during their rural healthcare programme. Many physicians have little training in primary healthcare, and they need to be equipped for the special challenges of large patient numbers, wide scope of practice, limited options for investigation and treatment, working alongside nurse clinicians and difficulties referring patients to the next level of care.

gramme. Many physicians have little training in primary healthcare, and they need to be equipped for the special challenges of large patient numbers, wide scope of practice, limited options for investigation and treatment, working alongside nurse clinicians and difficulties referring patients to the next level of care. Evidence for the effectiveness of the PACK Adult approach Four large randomised controlled trials, completed between 2003 and 2013, support the effectiveness of PACK Adult to improve health outcomes and strengthen health systems.21–23 The first trial tested the respiratory components of the package, the second, its expansion to cover HIV/AIDS, the third, nurse prescribing of ART, and the most recent, the expansion to NCD and mental health (still to be published). All of these trials have been pragmatic trials completed in real world circumstances, including frequent turnover of staff, medication stock outs and variable quality of implementation. Other studies using different designs, including qualitative research and economic evaluations, have complemented the trials.25 26 34 35 The studies, between them, have shown: Modest but consistent and reproducible improvements across a range of outcomes and behaviours (prescribing, referral, screening). Simultaneous improvements in the care of patients with both communicable and NCD. This is important because strengthening one part of the health system is often accompanied by weakening another part. Improvements extend beyond quality of care indicators to patient health outcomes, which is unusual in trials of health systems.

Simultaneous improvements in the care of patients with both communicable and NCD. This is important because strengthening one part of the health system is often accompanied by weakening another part. Improvements extend beyond quality of care indicators to patient health outcomes, which is unusual in trials of health systems. Shifts in healthcare utilisation with reductions in the length and duration of hospital admissions. Low-intensity delivery of educational outreach is effective, but there is a dose–response effect, suggesting that more training is more effective. The combination of algorithmic guidelines based on symptoms, educational outreach training, and non-physician prescribing is particularly empowering and effective. The approach is highly acceptable, and popular among nurses and trainers. No evidence of harm has been observed in any of the trials. This evidence makes PACK Adult one of few programmes for training health workers to be based on robust evidence. A fifth trial of adding counselling for depression is underway, and there have also been studies of implementing an adaptation in Malawi.36 37 Experience beyond South Africa PALSA PLUS has been adapted and implemented in pilot sites in Malawi, the Gambia, Brazil and Mexico, and PACK Adult in Botswana.

This evidence makes PACK Adult one of few programmes for training health workers to be based on robust evidence. A fifth trial of adding counselling for depression is underway, and there have also been studies of implementing an adaptation in Malawi.36 37 Experience beyond South Africa PALSA PLUS has been adapted and implemented in pilot sites in Malawi, the Gambia, Brazil and Mexico, and PACK Adult in Botswana. The KTU led the guideline adaptation in Malawi in partnership with the non-governmental organization (NGO) Dignitas International, and in Botswana with the University of Botswana. In both cases, this required lengthy consultation with many local clinicians and managers taking around 6 months of the time of a guideline developer over a year. In Botswana, the Ministry of Health convened a steering group with representatives of relevant programmes ranging from NCD to maternal health. Working groups from these programmes gave input into the guideline pages, resulting in a high level of investment and sense of ownership by the ministry. As in South Africa, existing national and provincial guidelines contained inconsistent recommendations, complicating their integration into a single coherent tool. For some conditions, local guidelines were either out of date or absent. Local clinicians assigned to the guideline adaptation struggled to commit the time and attention to detail to resolve discrepancies between policy and available resources.

ecommendations, complicating their integration into a single coherent tool. For some conditions, local guidelines were either out of date or absent. Local clinicians assigned to the guideline adaptation struggled to commit the time and attention to detail to resolve discrepancies between policy and available resources. KTU led the training workshops in Malawi and Botswana. All trainers went on to facilitate onsite training. This approach proved readily transferable, and initial resistance to an unconventional format without didactic presentation gave way to active participation and ownership. Clinic staff received the training with enthusiasm, but expressed frustration at the gap exposed by the guideline between policy and reality in the clinics. Further implementation was initially well supported in Malawi, especially by the NGO appointment of a programme coordinator, but later faltered because of political unrest, the global economic crisis and failure to position the programme within the Ministry. Both the Gambia and Rwanda approached KTU to implement the programme, but with unrealistic timelines. In both cases, the health ministries underestimated the cost of guideline adaptation and viewed the intervention as one off training rather than a programme requiring the creation of in-country capacity to sustain it. KTU was not involved in adapting or implementing PALSA PLUS in either Brazil or Mexico, rather the programmes were led by local clinicians. The instructing of trainer model was not used. The Brazilian adaptation is being evaluated ahead of scale up.

Both the Gambia and Rwanda approached KTU to implement the programme, but with unrealistic timelines. In both cases, the health ministries underestimated the cost of guideline adaptation and viewed the intervention as one off training rather than a programme requiring the creation of in-country capacity to sustain it. KTU was not involved in adapting or implementing PALSA PLUS in either Brazil or Mexico, rather the programmes were led by local clinicians. The instructing of trainer model was not used. The Brazilian adaptation is being evaluated ahead of scale up. Meeting global demand Over the past 2 years KTU has received numerous expressions of interest from countries as diverse as Zanzibar, Brazil, Mali, Nigeria, Uganda, Kenya, Tanzania, Rwanda, Cambodia, Vietnam, Bangladesh, Nepal, Oman, Yemen, China, India, Germany and Canada. It is clear, however, that a way must be found to enable cost-effective and sustainable implementation of PACK Adult in other countries, and KTU, together with the BMJ, Open University and other partners, has worked to develop a model to do so. Responding to the challenges of the global scaling up of PACK The challenges to scaling up something like PACK Adult are globally different from those of developing the components of PACK Adult and implementing it in South Africa, and meeting the challenges is giving KTU a rich opportunity to learn.

Meeting global demand Over the past 2 years KTU has received numerous expressions of interest from countries as diverse as Zanzibar, Brazil, Mali, Nigeria, Uganda, Kenya, Tanzania, Rwanda, Cambodia, Vietnam, Bangladesh, Nepal, Oman, Yemen, China, India, Germany and Canada. It is clear, however, that a way must be found to enable cost-effective and sustainable implementation of PACK Adult in other countries, and KTU, together with the BMJ, Open University and other partners, has worked to develop a model to do so. Responding to the challenges of the global scaling up of PACK The challenges to scaling up something like PACK Adult are globally different from those of developing the components of PACK Adult and implementing it in South Africa, and meeting the challenges is giving KTU a rich opportunity to learn. First, the guidelines must be adapted in response to local epidemiology and policies, and availability of health workers, medications and services. The aim is to establish local capacity to localise, revise and implement the guidelines to ensure satisfactory tailoring to the local setting, and ownership by those implementing it. This adaptation is a complex process that requires clinical, evidential and diplomatic skills. It can be difficult to find local people with the necessary skills. KTU is helping by preparing a generic version of PACK that provides the evidence for each of its 3500 recommendations along with guidance from BMJ Evidence, the WHO and other global bodies, and decision support tips with a limited resource focus. In addition, KTU can offer mentoring for those localising the guidelines and is producing a series of mentorship packages to guide local country teams to localise, pilot, evaluate, scale and revise PACK Adult.

e from BMJ Evidence, the WHO and other global bodies, and decision support tips with a limited resource focus. In addition, KTU can offer mentoring for those localising the guidelines and is producing a series of mentorship packages to guide local country teams to localise, pilot, evaluate, scale and revise PACK Adult. Sometimes, the guidelines must be translated into languages other than English, which requires translation skills and, ideally, a means to quickly modify the translated version when the generic English version changes. Second, ways must be found to make training more cost-effective. Training in South Africa has been mostly face-to-face, but it should be possible to produce onscreen support, and KTU is working on this with the Open University. Onscreen is more useful than online because many rural areas do not have easy access to the internet. Third, the South African programme has mostly used paper, although an interactive e-book tablet version has been developed and piloted in four rural clinics, with success. A sophisticated e-reader should be developed, but this requires skills and investment in information technology, plus very careful design. A great deal of thought and learning has gone into the production of the paper version, and this must now be repeated for electronic versions.

d in four rural clinics, with success. A sophisticated e-reader should be developed, but this requires skills and investment in information technology, plus very careful design. A great deal of thought and learning has gone into the production of the paper version, and this must now be repeated for electronic versions. Fourth, if PACK is to be used by non-physicians (which is not always the case but usually is in rural areas and slums in low and middle income countries) then they must be able to prescribe. This usually proves to be a long drawn out political process with physician organisations often resistant. Considerable political skills are necessary to make this happen. Fifth, there should ideally be a quality assurance system for the guidelines, the training and the daily use of the guidelines. More work is needed on this in South Africa as well as globally, although opportunities exist to form partnerships with other organisations that lead this work, but lack the appropriate clinical decision support. Finally, the biggest challenge may be obtaining funding and achieving a sustainable model. It seems to be much easier to find funding for research than for scaling up. Investment is needed to achieve the developments outlined above, and it is hard to find this investment. In terms of finding a sustainable model, there is probably only one long-term option—becoming a business of some sort, perhaps a social enterprise.

to be much easier to find funding for research than for scaling up. Investment is needed to achieve the developments outlined above, and it is hard to find this investment. In terms of finding a sustainable model, there is probably only one long-term option—becoming a business of some sort, perhaps a social enterprise. The business model has simplicity: you generate products and services, customers pay you for the value you provide, the surplus (profit) in what they pay you allows you to invest to produce still better products and services, the customers pay you for the added value—and a virtuous circle is established. If you cease to provide value you go out of business. In contrast, raising research grants or seeking donor funding requires spending a lot of time raising the funds, and once the research is over or a pilot is completed, funding usually ends and the search for funding must begin again. In addition to the challenge of finding investment, the ‘business model’ requires a set of skills that is usually not available in academic units. The units either have to develop the skills themselves or partner with organisations that have them. KTU is doing both, but the challenges of scaling up something like PACK Adult to other countries are very different from those needed to develop products and services in the first place—probably explaining why so many innovations do not spread. Global health would probably benefit from less time and money spent developing new innovations, and more on finding ways to spread those we already have.

to other countries are very different from those needed to develop products and services in the first place—probably explaining why so many innovations do not spread. Global health would probably benefit from less time and money spent developing new innovations, and more on finding ways to spread those we already have. Extending the vision PACK Adult is growing in response to requests from the health workers, trainers and managers who use it. These requests include a package for community care workers (CCWs) and one for the management of child health. Many low-income countries lack not only physicians but also nurses. The material developed for CCWs so far covers referral recommendations for seven priority chronic conditions. A 22-page pocket guide, PACK CCW, complemented by patient information leaflets and a training programme, has been developed, and the package is being piloted with 380 CCWs before further expansion. Following a request by the South African Provincial Western Cape Health Department and others, funding is now available to develop PACK Child, which would extend beyond the scope of IMCI to include common non-life-threatening conditions, as well as conditions in children between 5 and 15 years. The depression package currently being piloted will inform the development of PACK Counsellor, a generic tool for lay health workers providing counselling services across a range of conditions, behaviours and therapies.

Many low-income countries lack not only physicians but also nurses. The material developed for CCWs so far covers referral recommendations for seven priority chronic conditions. A 22-page pocket guide, PACK CCW, complemented by patient information leaflets and a training programme, has been developed, and the package is being piloted with 380 CCWs before further expansion. Following a request by the South African Provincial Western Cape Health Department and others, funding is now available to develop PACK Child, which would extend beyond the scope of IMCI to include common non-life-threatening conditions, as well as conditions in children between 5 and 15 years. The depression package currently being piloted will inform the development of PACK Counsellor, a generic tool for lay health workers providing counselling services across a range of conditions, behaviours and therapies. The KTU vision is to integrate the management of common priority conditions both within primary care and across levels of care, targeting all health workers from community health workers to professional clinicians based in first-level facilities, as well as those receiving patient referrals to district hospitals. PACK Adult is the first component of this package, and working out how best to adapt it quickly and cost-effectively for other countries in a way that fosters local ownership and sustainability will be key learning for this broader vision.

facilities, as well as those receiving patient referrals to district hospitals. PACK Adult is the first component of this package, and working out how best to adapt it quickly and cost-effectively for other countries in a way that fosters local ownership and sustainability will be key learning for this broader vision. Supplementary Material Web supplement The authors would like to acknowledge the more than 200 doctors and nurses, managers and academics who contribute regularly to the PACK Guidelines. They also acknowledge Sandy Picken and Camilla Wattrus for compiling table 1. Twitter: Follow Richard Smith at @Richard56 Contributors: LF and RS produced the first draft. All the authors commented critically and LF and RS brought together the comments. LF is the guarantor.

Supplementary Material Web supplement The authors would like to acknowledge the more than 200 doctors and nurses, managers and academics who contribute regularly to the PACK Guidelines. They also acknowledge Sandy Picken and Camilla Wattrus for compiling table 1. Twitter: Follow Richard Smith at @Richard56 Contributors: LF and RS produced the first draft. All the authors commented critically and LF and RS brought together the comments. LF is the guarantor. Funding: The authors would like to acknowledge past funders of the KTU, including the International Development Research Centre, the Belgian Technical Cooperation, Development Cooperation Ireland, Discovery Health South Africa, the Canadian International Development Agency, Britain's Medical Research Council, South Africa's Medical Research Council, the National Heart, Lung and Blood Institute (NIH), and USAID, through PEPFAR. Current funders of the KTU include the Sowerby Foundation and National Institute of Mental Health (NIH), and the Health Resources and Services Administration, US Department of Health and Human Services, through the Medical Education Partnership Programme. They also acknowledge our partners in the Western Cape and Free State governments, the National Department of Health (South Africa), Dignitas International, the University of Bostwana and the BMJ Publishing Group.

stration, US Department of Health and Human Services, through the Medical Education Partnership Programme. They also acknowledge our partners in the Western Cape and Free State governments, the National Department of Health (South Africa), Dignitas International, the University of Bostwana and the BMJ Publishing Group. Competing interests: LF, RC, NF and VT are employees of KTU. GF is an ex-employee and now a consultant to KTU. EB is director of the University of Cape Town Lung Institute, which includes KTU. MZ and MB have both worked on the trials of PACK Adult and its predecessors. RS directed the United Health Chronic Disease Initiative that provided some funding for the development and testing of PACK Adult. Provenance and peer review: Not commissioned; internally peer reviewed.

Background According to the United Nations Department of Economic and Social Affairs’ World Population Prospects, Brazil, Cambodia, Egypt, India, Turkey and Zimbabwe, are expected to see a rapid growth in the 65 and above population in the coming years (see online supplementary appendix A, table 1). Among low-income and middle-income countries (LMIC), Brazil, Cambodia, Egypt, India, and Turkey are expected to have the highest growth in 65 and above age group by the year 2050 and 2100 within each WHO region (see online supplementary appendix A, table 2).1 Zimbabwe, on the other hand, is expected to have the second highest growth among the LMIC in the WHO Region for Africa (AFRO) by 2050 and third highest by 2100 (see online supplementary appendix A, table 2).1 In addition to rapidly ageing populations, functional limitations (activity and participation restrictions) due to impairment and disability among older populations in LMIC are expected to be much higher than in high-income countries (HIC). Impairment is defined by the International Classification of Functioning, Disability and Health (ICF) as any loss or abnormality of psychological, physiological or anatomical structure or function.2 Disability is defined by the ICF as any restriction or lack (resulting from an impairment) of ability to perform an activity in the manner or within the range considered normal for a human being. With increasing age, older adults are more likely to experience impairment and disability that increase activity limitations where they may face difficulties in executing activities, and participation restrictions where they cannot become involved in social activities.2

in the range considered normal for a human being. With increasing age, older adults are more likely to experience impairment and disability that increase activity limitations where they may face difficulties in executing activities, and participation restrictions where they cannot become involved in social activities.2 For cognitive-related functional limitations, according to the World Alzheimer Report 2010, the growth in number of people with dementia in low-income countries is expected to increase from ∼35 million in 2010 to ∼115 million by 2050, compared to HIC where the increase is expected to be from ∼15 million in 2010 to close to ∼35 million by 2050.3 In 2012, years lived with disability (YLDs) due to falls in adults aged 50–69 in developing countries (66%) was nearly double the rate in developed countries (34%).4 Falls have been documented as a key health concern leading to YLDs that older adults are facing today, and occur at home, in institutions and in the community.5

2, years lived with disability (YLDs) due to falls in adults aged 50–69 in developing countries (66%) was nearly double the rate in developed countries (34%).4 Falls have been documented as a key health concern leading to YLDs that older adults are facing today, and occur at home, in institutions and in the community.5 AT are defined as “the application of organized knowledge and skills, procedures and systems related to provision of assistive products, whose primary purpose is to maintain or improve an individual's functioning and independence, facilitate participation, and enhance overall well-being and quality of life.”6 AT provide significant support and benefits for older adults in various ways: safety and prevention (ie, prevention of falls), mobility and independence, social connectivity and ease of living, preservation of cognitive abilities, delay in depression, decline in functional loss and improved well-being and quality of life.7–13 Safety has consistently been identified as one of the most important reasons for the use of AT among older adults.8 10 According to Heywood et al10 home adaptations have reported a 70% increase in feeling safe at home among older adults, while adaptations such as grab rails and handrails have been found to produce lasting positive consequences. A systematic review by Chase et al14 found that AT, home modifications, along with other interventions prevented falls among community-dwelling older adults. Evidence from the current literature suggest that more than half (62%) of older adults felt safer from a risk of an accident with the availability of even minor adaptations.7 According to a study published by the AARP Foundation in the United States of America (USA), home safety devices made the majority of older adults feel safer (85%) and gave them peace of mind (78%).7

than half (62%) of older adults felt safer from a risk of an accident with the availability of even minor adaptations.7 According to a study published by the AARP Foundation in the United States of America (USA), home safety devices made the majority of older adults feel safer (85%) and gave them peace of mind (78%).7 The definition of AT used in this review includes all the external technologies or products that are designed for and used by older adults. These technologies may assist in carrying out actions (ie, walking, bathing, remembering to take medications) and/or prevent severe health outcomes such as falls and injuries that may lead to more complicated health problems. There is a vast range of AT; according to the WHO, most of these products can be divided into several classes (table 1).6 Table 1 Classes of AT

The definition of AT used in this review includes all the external technologies or products that are designed for and used by older adults. These technologies may assist in carrying out actions (ie, walking, bathing, remembering to take medications) and/or prevent severe health outcomes such as falls and injuries that may lead to more complicated health problems. There is a vast range of AT; according to the WHO, most of these products can be divided into several classes (table 1).6 Table 1 Classes of AT Class Product examples Products for cognitive functions Memory devices and pill box reminders Products for sensory functions Spectacles, software for screen magnification and reading, hearing aids Products related to orthotics and prosthetics Artificial limbs (prosthesis), spinal orthotics and cervical collar Products for personal mobility Wheelchairs, canes and crutches Products for activities of daily living (ADLs—see online supplementary appendix B) including personal care and protection Toilet chairs, diapers and robots Products for communication and skills training Voice and speech training devices, Braille apparatus and screen readers Products for recreation and sports Modified sports equipment Products for housing, work and environmental improvement Home modification, handrails or grab bar and controlled lighting AT, assistive technologies.

aily activities, means tested criteria for the assessment of poverty among vulnerable groups47 Customs and Excise Regulations 2004 Government provides for a waiver of duty for all specially adapted for use by persons with disabilities imported less than 5 years from the date of manufacture47 AT, assistive technologies. Table 5 Findings on AT availability by country (cited at least once in publications)

aily activities, means tested criteria for the assessment of poverty among vulnerable groups47 Customs and Excise Regulations 2004 Government provides for a waiver of duty for all specially adapted for use by persons with disabilities imported less than 5 years from the date of manufacture47 AT, assistive technologies. Table 5 Findings on AT availability by country (cited at least once in publications) Country Assistive technology Brazil Canes, wheelchairs, walkers, bath chairs, toilet chairs, hearing aids, ‘Geriatric’ software, activity board, calendar, routine organiser, cane tips to detect puddles, Android application to access social media websites (ie, Twitter), capacitive touch switches, home automation systems with web interface, augmentative and alternative communication devices (AAC) and digitally creating video audio descriptions, Braille display, button mouse, audio transmitter, capacitive driver, cane tip for water detection17–20 Cambodia AT for locomotor disabilities, weaving kits, bamboo walking canes, wheelchairs, crutches, tricycles, prostheses, artificial limbs, knee joints, double gripping hook/bowl attachments for prosthetic arm, carved wooden carrying yokes, chairs with an attached trays, wooden bathing benches with rails at each end, metal/rubber bathing benches, wheelchair jump seats, wooden toilet seats with handrails on each side, wheelchair toilet seats and soft plastic washing devices to clean self after toilet use and different forms of seating21 Egypt Egypt-Japan Assistive Device (EJAD), a low-cost AT designed to help elders in walking and sit-to-stand activities.22 EJAD imitates the caregiver's motion during a support task, which helps maintain the stability of the users at any situation, Tablet PCs powered by Android systems containing Quran in large font, cooking recipes, and old songs and Cellular phones with large buttons48 India Security alarms, raised seat, grab bars, video intercom, medicine dispenser/cabinet, Jar opener/closer, talking alarm clock/watch, magnifying toenail clipper, lap desk with book holder, back scrubber with hand loops on each end, kitchen finger protector, utensils with finger bump grips or hand strap, adjustable height bed, bed rail, Velcro fastenings shoes, Velcro clothing, pressure modification stockings, cordless speaker phone with pre-set memory-dial, large buttons and numbers, mobile phone with one touch call to family members, friends, helplines, doctors, security, eye glasses/magnifiers, ‘E-Netra’ reads text and converts into voice, hearing aids, visual alerting systems, memory games, walking sti

for communication and skills training Voice and speech training devices, Braille apparatus and screen readers Products for recreation and sports Modified sports equipment Products for housing, work and environmental improvement Home modification, handrails or grab bar and controlled lighting AT, assistive technologies. AT are a fundamental part of broader, integrated health and social system solutions to supporting older adults. Yet, despite the documented benefits of AT, some parts of the world have minimal or no access to AT that are designed for older adults. According to the WHO, in many LMIC, only 5–15% of people who require AT have access to them.15 LMIC are expected to experience the most rapid and dramatic demographic change, in shortened timeframes as compared to higher income countries. For example, it took more than 100 years for the share of France's population aged 65 or older to double from 7% to 14%,16 whereby it has been estimated that countries like Brazil will take less than 25 years to reach the same threshold.16 With rapidly ageing populations, accompanied by increasing functional limitations, in LMIC, AT are likely to be in higher demand.

share of France's population aged 65 or older to double from 7% to 14%,16 whereby it has been estimated that countries like Brazil will take less than 25 years to reach the same threshold.16 With rapidly ageing populations, accompanied by increasing functional limitations, in LMIC, AT are likely to be in higher demand. Given the trends in LMICs, this review investigated the currently available AT, existing enabling legislation for provision of AT, and initiatives that promote AT for older adults in Brazil, Cambodia, Egypt, India, Turkey and Zimbabwe. The objective is to explore in the published literature, if currently available AT and existing legislation related to AT-provision are sufficient to support ‘ageing in place’ rather than ageing in institutionalised care homes; facilitate better health for older adults, especially in terms of functioning and independence (ie, carrying out Activities of Daily Living or ADLs and Instrumental Activities of Daily Living (IADLs)); encourage inclusion and full participation in community activities; and enhance older adults’ overall well-being and quality of life. Methods Data sources and search strategy A systematic approach was taken to conduct the systematic review. The search strategy aimed to retrieve literature in English language that focused on AT to assist older populations (60 and above) in Brazil, Cambodia, Egypt, India, Turkey and Zimbabwe. Databases searched were MEDLINE (1950 to week 1 July 2014), EMBASE (1980 to week 1 July 2014), Scopus (1966 to week 1 July 2014) and Cochrane Library (1996 to week 1 July 2014).

ture in English language that focused on AT to assist older populations (60 and above) in Brazil, Cambodia, Egypt, India, Turkey and Zimbabwe. Databases searched were MEDLINE (1950 to week 1 July 2014), EMBASE (1980 to week 1 July 2014), Scopus (1966 to week 1 July 2014) and Cochrane Library (1996 to week 1 July 2014). WHO regional indexes including African Index Medicus (AIM); the Eastern Mediterranean Region Library Network (EMLIBNET); Pan American Health Organization (PAHO) Library Institutional Memory Database; Europe (EURO) publications; Index Medicus for South-East Asia Region (IMSEAR); WHO/WPRO Library; Western Pacific Region Index Medicus (WPRIM); and the WHO Library Database (WHOLIS) using the WHO Global Health Library platform were searched for relevant publications. Three groups of keywords were combined: those relating to AT, ageing population and selected LMICs. Search terms used in each database are shown in figure 1. All retrieved articles were initially reviewed by title and abstract to find potentially relevant papers. Those papers were reviewed to obtain articles that met the inclusion criteria. Reference lists of articles that met the inclusion criteria were reviewed to identify any further relevant papers. Figure 1 Search terms used in each database. Study selection-Inclusion criteria Selected papers were assessed against the following inclusion criteria: (1) studies published in academic and peer-reviewed journals, research reports, government reports (2) studies answered ‘yes’ to all three screening questions (3) written in English.

Figure 1 Search terms used in each database. Study selection-Inclusion criteria Selected papers were assessed against the following inclusion criteria: (1) studies published in academic and peer-reviewed journals, research reports, government reports (2) studies answered ‘yes’ to all three screening questions (3) written in English. Screening Questions: Does the publication address AT as a primary research topic? Does the publication focus on one of the six countries? Does the sample size or intended population include older adults (≥60 years old)? Study selection-exclusion criteria If the abstract indicated that the study did not relate to AT for ageing populations in LMICs, the study was excluded (ie, material relating to disability studies involving only younger people). Studies that focused on the technology use in hospitals were excluded, since the rationale of the systematic review is to focus on AT that have a direct impact on older adults to enhance their ADLs and participation in the community and at home. Duplicates were excluded (figure 2). Figure 2 Literature search.

Study selection-exclusion criteria If the abstract indicated that the study did not relate to AT for ageing populations in LMICs, the study was excluded (ie, material relating to disability studies involving only younger people). Studies that focused on the technology use in hospitals were excluded, since the rationale of the systematic review is to focus on AT that have a direct impact on older adults to enhance their ADLs and participation in the community and at home. Duplicates were excluded (figure 2). Figure 2 Literature search. Results The systematic review retrieved 17 relevant studies; four studies related to AT in Brazil, seven related to India, three related to Turkey and one study per each country for Cambodia, Egypt and Zimbabwe (figure 3). Seven publications focused on ‘mobility’ AT while six on ‘more than one function’; and AT for cognitive functions, ADLs, vision and others had one citation each (figure 4). Relevant studies are classified into four types as provided in table 2 below. Tables 3–5,17–33 provide an overview of relevant studies, existing enabling legislation and findings on AT for each country, respectively. Table 2 Study classification Study type Description Research Includes all types of research studies Product development and intervention Project focusing on the development/production of AT and intervention to assess product impacts Product development Merely production or development of AT Product evaluation Evaluation of products AT, assistive technologies. Table 3 Table of Evidence

Study type Description Research Includes all types of research studies Product development and intervention Project focusing on the development/production of AT and intervention to assess product impacts Product development Merely production or development of AT Product evaluation Evaluation of products AT, assistive technologies. Table 3 Table of Evidence Country, Author, Year Population Description Study Type/Data Collection Method Relevant Outcomes Regarding Available AT and Other Findings Brazil Assis et al, 20039 N/A Product development and intervention, and Product evaluation Development of low-cost “Geriatric” software and equipment (activity board, calendar, routine organizer). Software and equipment were found useful and applicable to programmes for individuals with cognitive impairment. Post intervention showed improvement in cognitive functioning and daily activities at home. De Cruz et al, 20133 N=91, Mean Age=62.2 Research Cross-Sectional Most reported assistive resources were those to assist mobility (46%) (Predominately canes), and ADLs (29%). No direct association between AT and independence (only mild dependency). Mello MAF, 199917 N=84, Age range=73-102 Research (Doctoral Thesis) N/A 74% of participants used some type of AT. Canes were the most used type of AT (45%). 4% of the participants had all the assistive devices and home modifications they needed. 35.5% of users were satisfied with the use of these technologies. 2/3 of the AT were purchased by the user, a relative, or friend. Schüler, et al, 20135 N/A Product development Produced low-cost AT such as cane tips to detect puddles, Android application for social media websites (i.e. Twitter), capacitive touch switches, home automation systems with web interface, AAC, and digitally creating video audio descriptions for a lower cost. Cambodia Jones et al, 20037 N=100 (Disabled, non-disabled adults, and disabled children). Results only focused on older adults. Research Qualitative study (In-depth field work Semi-structured interviews and observation) Most had access to AT for bathing, drinking, toilet use, and transporting water from the source. Majority of solutions were low or no cost, made with locally available material and designs. AT benefited disabled individuals in terms of increased self-reliance, ability to contribute to the family or community, improved social status, reduced work-load, well-being, economic situation, and high self-esteem.

e source. Majority of solutions were low or no cost, made with locally available material and designs. AT benefited disabled individuals in terms of increased self-reliance, ability to contribute to the family or community, improved social status, reduced work-load, well-being, economic situation, and high self-esteem. Egypt Salah et al., 201110 N/A Product development Development of the EJAD, low cost AT to help older adults walk and complete sit-to stand activities. India Jefferds et al., 20106 PART survey age range=19-67,WST/QUEST age range=21-60 Research Quantitative Study (Survey Questionnaire) Custom-fitted wheelchairs provided by the Department of AT improved wheelchair skill scores and technology satisfaction among participants. Kumar et al, 200912 N=100, Age=60-85 years, Mean Age=N/A Research Qualitative Study (Survey Questionnaire) Inadequate availability of AT in India has been partially due to the lack of awareness among the users as well as among the professionals. Although respondents lacked awareness of the devices, they were willing to use them, provided these are made available at affordable costs. 27% had difficulty in the kitchen gadgets and appliances, 26% in recreation/entertainment, 21% in home elements, 19% with mobility and 18% with ADLs. Highest awareness was in communication devices, home elements, ADL mobility, and least was in high-tech AT. Langdon et al., 20138 N=19, Age range=25-84 Research Qualitative Study (Semi-structured Interview) Participants with impairments were enthusiastic to use new technical devices, but emphasized on need of training. Many users did not have access to assistive devices and interaction techniques. Manogna S et al., 201013 N/A Product Development Development of head movement based assist system to serve people with quadriplegia. The assist system has proven to be of simple implementation and of low cost. The system was easily handled by the patients. Shore, 200814 N=100 (from India), Mean Age=50±25 Age range=3-93 Research Quantitative Study (Survey) Use of wheelchair resulted in significant shift towards independent function. The impact on health and quality of life was generally viewed as positive (56.2%). Shore S et al, 201215 N=206, Mean Age=54, Age range=4-102 Research Quantitative Study Survey Receiving a simple and durable wheelchair improved the reported health, quality of life, and function of recipients following 12 months of use. 78% had not owned a wheelchair before, largely due to lack of money.

itive (56.2%). Shore S et al, 201215 N=206, Mean Age=54, Age range=4-102 Research Quantitative Study Survey Receiving a simple and durable wheelchair improved the reported health, quality of life, and function of recipients following 12 months of use. 78% had not owned a wheelchair before, largely due to lack of money. Increase in independence associated with the use of wheelchair (independence score changed from 3.9 to 5.0 over 12 months). Zipfel, et al., 20074 N/A Product Development Developed a manual folding cross-brace design with several points of adjustability, and higher grade bolts. The new chair design has so far proven to be more durable and comfortable than existing wheelchairs. Turkey Bengisu, 20102 N=80, Mean Age=35±9 Age range =18-61 Research Qualitative Study (Phone interview) Product evaluation Three most used AT were computer screen readers (46%), talking watches (26%), and screen readers for cellular phones (21%). Most desired AT were cellular phones with screen reading capability. Most frequently requested AT were devices for warning about barriers, reading printed documents and signs. Screen readers for computers were rated the highest (4.7/5) and 76% of the participants rated them very useful and 0 participants rated the product not useful. Screen readers for cellular phones were rated 4.6/5 and 72% of the participants rated them very useful and only 1% rated it not useful. Simsek et al., 201216 N=163, Mean age=73.18 ±6.62, Age range =58-105 Research Quantitative and qualitative Study (mixed methods) The most frequently used device to maintain mobility was a walking stick. A positive relationship was found between AT usage and mobility state, and AT usage and mobility activities required for developing basic life functions (P<0.05). AT usage was greater in women (39%) than in men (19%). Ucsular et al., 201111 N=24, Age range=13-65 Research Randomized Control Trial The Wheelchair Skill performance scores increased significantly in the training group (P=0.034) compared to control group (P=0.01). A significant improvement in safety scores in the training group compared with the control group. Zimbabwe Edie et al, 20031 N=3901 households, Mean Age (Disability)=43.2 Age range=0-98 Matabeleland, Manicaland and Midlands (44% of the total population). Research Qualitative Study (Survey Of those who used AT, 41.7% were among ≥61.

improvement in safety scores in the training group compared with the control group. Zimbabwe Edie et al, 20031 N=3901 households, Mean Age (Disability)=43.2 Age range=0-98 Matabeleland, Manicaland and Midlands (44% of the total population). Research Qualitative Study (Survey Of those who used AT, 41.7% were among ≥61. Of those who used AT, 80% used personal mobility devices, 19% used AT for communication and information, 2.2% of the users had AT for personal care & protection. 31.2% of the devices were provided by private sources, 27.4% from government health services, 8.3% from NGOs, 3.2% from other government services and 28.8% from other sources. 63.1 % received guidance on using the devices, 36.9% had not received any instructions. Significantly fewer disabled in rural areas had received guidance; 57% versus 73% in urban areas. 36.3% were responsible themselves for maintenance and repair, 14% had government assistance, 21 % relied on families and 14 % had devices not maintained or repaired. Zimbabwe had the smallest gap in AT/services compared to 3 other African countries (Malawi, Namibia, and Zambia). 35.1% urban Zimbabweans had access to AT compared to 22.2% rural populations. Table 4 Overview of legislation/national plans/policies and acts

ed on families and 14 % had devices not maintained or repaired. Zimbabwe had the smallest gap in AT/services compared to 3 other African countries (Malawi, Namibia, and Zambia). 35.1% urban Zimbabweans had access to AT compared to 22.2% rural populations. Table 4 Overview of legislation/national plans/policies and acts Country Legislation/national plan Significance Ratified by all 6 countries UN Convention on the Rights of Persons with Disabilities (CRPD) Online supplementary appendix C, UN CRPD Ratified by Brazil and Zimbabwe CRPD Optional Protocol (OP-CRPD) Allows for individual complaints to be submitted to the CRPD Committee alleging that rights have been violated under CRPD. If the CRPD Committee makes a finding that the State has failed in its obligations under the CRPD, it will issue a decision requiring that the violation be remedied and for the State party to provide follow-up information Brazil Decree No. 7612 Rights of Persons with Disabilities (PWDi) –Living Without Limits Plan. Promote accessibility for people with disability and frail adults through: ▸ Credit lines for the acquisition of AT ▸ Institutionalisation of the National Programme for AT. ▸ Tax rebate on AT products.34 Law 10 048/2000 Decree No. 5296 Chapter VII, Article 64, 65 and 66 Reduction or exemption from import taxes for products that are not available within the country. Reduction or exemption from excise tax on technical aids. Deduction of income tax for all AT for people with disabilities or reduced mobility.35 Recognize the area of AT as an area of knowledge. Encourage the formation and training of orthotics and prosthetics. Encourage studies to support the development of standards on technical aids. Detection of regional reference centres for technical aids, aiming at the formation of an integrated national network35 Decree No. 4360 Provision of long-term assistance for people with disabilities and older people36 Cambodia National Policy on the Health Care for Elderly and Disabled People 1999 ‘Further plan of action’ states that providing the elders and disabled to obtain the AT such as hearing aids, walking aids and kitchen utensils to assist living independently as one of the priorities for social protection37 Protection and Promotion of the Rights of Persons with Disabilities (PPRPD), Article 4, 20 and 26 Commits to: Giving priority to technologies at an affordable cost.

ed to obtain the AT such as hearing aids, walking aids and kitchen utensils to assist living independently as one of the priorities for social protection37 Protection and Promotion of the Rights of Persons with Disabilities (PPRPD), Article 4, 20 and 26 Commits to: Giving priority to technologies at an affordable cost. Providing accessible information to persons with disabilities about available AT, including new technologies, as well as other forms of assistance, support services and facilities. Facilitating access to quality mobility aids, devices, AT and forms of live assistance and intermediaries. Encouraging entities that produce AT to take into account all aspects of mobility for persons with disabilities Promoting the availability, knowledge and use of AT designed for PWDi38 Disability Action Council (DAC) Royal Sub-Decree No.

quality mobility aids, devices, AT and forms of live assistance and intermediaries. Encouraging entities that produce AT to take into account all aspects of mobility for persons with disabilities Promoting the availability, knowledge and use of AT designed for PWDi38 Disability Action Council (DAC) Royal Sub-Decree No. 59, 2009, PPRPD Involved in standardisation through subcommittee on prosthetics, orthotics and assistive devices and give advice and orientation at the national level Provides recommendations for the rights of persons with disabilities Endorsed in the law on the PPRPD, and receives core funding from USAID through its humanitarian assistance budget39 India The National Policy for Persons with Disabilities 2006 Physical Rehabilitations Strategies under Rehabilitation measures includes provision of AT as one of the commitments Commits to creating an environment that provides Persons with Disabilities equal opportunities, protection of their rights and full participation in society40 Persons with Disabilities (Equal Opportunities, Protection of Rights and Full Participation) Act 1995 The ‘education’ component suggests research for designing and developing new AT Governments shall initiate/encourage research by official and non-governmental agencies to design and develop new AT ‘Research and manpower development’ promotes sponsoring research in the development of AT41 Scheme of Assistance to Disabled Persons for Purchase/Fitting of Aids/Appliances (ADIP Scheme) 2005 (see online supplementary appendix D) Assist low income individuals with disabilities in obtaining assistive aids and appliances Those aids/appliances which do not cost more than US$100.46 are covered Those with income US$250 to US$333/month receive devices at a 50% subsidy Those with incomes of US$250/month, receive devices at full cost42 Provide 500 Cochlear implants free per year India's National Policy on Senior Citizens 2011 Increasing advancement in technology is one of the listed priority areas Highlights that technology is not only needed for the ≥80, but also for rural poor, disadvantaged seniors who will have longer years in the coming decades43 Turkey Turkish Disability Act Law no.

year India's National Policy on Senior Citizens 2011 Increasing advancement in technology is one of the listed priority areas Highlights that technology is not only needed for the ≥80, but also for rural poor, disadvantaged seniors who will have longer years in the coming decades43 Turkey Turkish Disability Act Law no. 5378/2005, Article 3 and 7 Commits to having buildings, open space, transportation and information services and information and communication technology, accessible and usable by people with disabilities to be safe and independent Commits to disclosure of information and communication technologies and services to be accessible for people with disabilities Value Added Tax (VAT) Law No. 3065, Article 167–4/4 and 17/4 Provides importation exemptions within the scope of customs law article 167-4/4 and general goods for the disabled Provides partial exemption for delivery of computer programmes and every kind of instruments (excluding vehicles), and equipment especially produced for daily life of the disabled.44 The services provided in elderly and disabled care houses are considered exceptions and the services provided in such places are also held exempt from the value added tax45 Social Assistance and Solidarity Law No.

d of instruments (excluding vehicles), and equipment especially produced for daily life of the disabled.44 The services provided in elderly and disabled care houses are considered exceptions and the services provided in such places are also held exempt from the value added tax45 Social Assistance and Solidarity Law No. 3294, Article 5 Covers the costs of treatment and function-acquisition related orthopaedic and other aid tools/equipment for PWDi who are covered by social security institutions and get salary or payment from institutions46 Zimbabwe National Disabled Persons’ Act 1992, Section 4 Established the National Disability Board that is responsible for the formulation and development of measures and policies to ensure that disabled people live independently and access all social services, and estimation of the cost of the measures for the welfare and rehabilitation of disabled persons47 Disabled Persons Fund (DPF) Provides public assistance, medical fees, purchases assistive devices and a revolving loan facility Allows exempt from the payment of import tax for all specially adapted cars purchased Ensures that disabled people are assisted with the purchase of AT like crutches, artificial limbs, spectacles and wheelchairs to enable carrying out day to day functions47 Social Welfare Assistance Act of 1988 Provides for the health provision, maintenance allowances, purchase of AT, import of special equipment for the disabled to assist in their daily activities, means tested criteria for the assessment of poverty among vulnerable groups47 Customs and Excise Regulations 2004 Government provides for a waiver of duty for all specially adapted for use by persons with disabilities imported less than 5 years from the date of manufacture47 AT, assistive technologies.

ication stockings, cordless speaker phone with pre-set memory-dial, large buttons and numbers, mobile phone with one touch call to family members, friends, helplines, doctors, security, eye glasses/magnifiers, ‘E-Netra’ reads text and converts into voice, hearing aids, visual alerting systems, memory games, walking sti ck, walkers, wheelchair, spinal braces, stair glide, handrails, prosthesis-artificial limbs, height adjustable/tilting chairs/stool/tables/desks, CDs on traditions, historic movements, pilgrimage and tourist places, folk songs, wheelchair and head movement-based assist system23–29 Turkey Reading pens, screen readers for cell phones, and computers, locators, laser canes, modified appliances, talking watches, walking sticks, crutches, walkers and wheelchairs30–32 Zimbabwe Wheelchairs, walking sticks, crutches, standing frame, white cane, eye glasses, magnifying glasses, hearing aids, Braille, portable writer and enlarged print, bath and shower seats, commode chairs, toilet seat raisers, safety rails and eating aids, aids for opening containers, flashing light on doorbell, vibrating alarm clock and amplified telephones33 AT, assistive technologies. Figure 3 Number of publications retrieved by country. Figure 4 Number of publications by assistive technologies category.

ck, walkers, wheelchair, spinal braces, stair glide, handrails, prosthesis-artificial limbs, height adjustable/tilting chairs/stool/tables/desks, CDs on traditions, historic movements, pilgrimage and tourist places, folk songs, wheelchair and head movement-based assist system23–29 Turkey Reading pens, screen readers for cell phones, and computers, locators, laser canes, modified appliances, talking watches, walking sticks, crutches, walkers and wheelchairs30–32 Zimbabwe Wheelchairs, walking sticks, crutches, standing frame, white cane, eye glasses, magnifying glasses, hearing aids, Braille, portable writer and enlarged print, bath and shower seats, commode chairs, toilet seat raisers, safety rails and eating aids, aids for opening containers, flashing light on doorbell, vibrating alarm clock and amplified telephones33 AT, assistive technologies. Figure 3 Number of publications retrieved by country. Figure 4 Number of publications by assistive technologies category. AT- related National Initiatives Brazil The Assistive Technology Innovation Programme (PITA) finances technological development and innovation of products, processes and services targeted at people with disabilities, older adults and as well as others with difficulties of mobility.49 A budget of US$70 million is anticipated for this programme until 2014 directed to ‘innovation activities’ carried out by companies, universities and research institutions. Innovation activities include development and/or enhancement of products, processes and services related to AT and/or contribution toward prevention, alleviation or elimination of deficiencies.49

s programme until 2014 directed to ‘innovation activities’ carried out by companies, universities and research institutions. Innovation activities include development and/or enhancement of products, processes and services related to AT and/or contribution toward prevention, alleviation or elimination of deficiencies.49 Future plans include the establishment of the National Centre for Excellence in Adaptive Technology, composed of 20 research centres implemented in public universities to develop strategic technologies, with an emphasis on prevention, rehabilitation and accessibility. The programme launched a virtual catalogue of more than 1200 AT products available in Brazil (http://assistiva.mct.gov.br).34 The website is to be updated at least once a year, so consumers and professionals are aware of availability of AT products in the country.34 Cambodia Phnom Penh's National Component Factory is the only source of Prosthetic and Orthotic (P&O) devices in the country. It monitors and evaluates the production of prosthetics and AT materials, in order to supply for workshops and centres that produce P&O.50 They supply close to 15 000 those in need a year. Since January 2009, Phnom Penh's National Component Factory has produced and supplied more than 3000 artificial limbs, knee joints and alignment systems to physical rehabilitation centres in the country, free of charge.51

ly for workshops and centres that produce P&O.50 They supply close to 15 000 those in need a year. Since January 2009, Phnom Penh's National Component Factory has produced and supplied more than 3000 artificial limbs, knee joints and alignment systems to physical rehabilitation centres in the country, free of charge.51 A number of workshops are available to provide prostheses, wheelchairs and crutches, repair and maintenance services in Cambodia, where each workshop is operated in partnership between the Government and NGOs. Currently 21 workshops provide AT for people with locomotor disabilities. Some of the workshops are run by numerous associations, committees and organisations. Online supplementary appendix E provides an overview of the associations and services provided. Egypt In Egypt, non-governmental organisations (NGOs) such as the Egyptian Society for Elderly Care and Misr al-Kheir provide assistance in acquisition of ATs such as canes and wheelchairs.48 The Arabic Doctors’ Union (an NGO) provides hearing aids and cochlear implants to those of age 65 and above while the Egyptian Society for Hearing Impaired Persons is responsible for provision and training the use and maintenance of hearing aids.48 Furthermore, Wafaa and Amal48 compound established and run by the Egyptian Veterans Association includes an assistive device factory for manufacture, maintenance and some export of devices for disabled individuals and veterans, where majority are older adults.

ion and training the use and maintenance of hearing aids.48 Furthermore, Wafaa and Amal48 compound established and run by the Egyptian Veterans Association includes an assistive device factory for manufacture, maintenance and some export of devices for disabled individuals and veterans, where majority are older adults. The government contributes to the provision of AT through the development, production, distribution, maintenance and repair. The government also provides prostheses/orthoses, wheelchairs, crutches, hearing devices and visual devices.52 In addition, the Ministry of Health and the Ministry of Social Solidarity provide healthcare and rehabilitation services, including technical aids and devices through their institutions. Ministry of Social Affairs provides services such as physiotherapy, prosthesis and assistive devices and rehabilitation. India The Artificial Limb Manufacturing Company (ALIMCO), fully supported by the Government of India's Ministry of Welfare, manufactures a large number of tools, machines and components for producing AT and finished devices. So far, ALIMCO has provided AT to 4.2 million users in India and serves approximately 200 000 users annually. ALIMCO also provides AT to Afghanistan, Nepal and Bhutan among others.53

ment of India's Ministry of Welfare, manufactures a large number of tools, machines and components for producing AT and finished devices. So far, ALIMCO has provided AT to 4.2 million users in India and serves approximately 200 000 users annually. ALIMCO also provides AT to Afghanistan, Nepal and Bhutan among others.53 The Indian Drugs and Pharmaceuticals Ltd. is the only other Indian company that manufactures components for AT in large numbers.54 In addition, there are at least 15 agencies producing wheelchairs and 20 producing tricycles.54 The most commonly produced are hearing aids, by as many as 32 manufacturing agencies mainly in the private sector.54 Most of the time, AT are distributed through NGOs, district rehabilitation centres, medical colleges and hospitals, special schools and camps and national institutions in India. NGOs and government supported institutions have spread over 19 states and union territories, which are financially supported under the ADIP Scheme run by the Government's Ministry of Welfare.54 The Department of Rural Development is involved in distribution efforts, and some NGOs not supported by the government also give out assistive devices. As many as 25 out of 31 states and union territories have their own schemes of providing AT through government hospitals and voluntary agencies.54 Many district rehabilitation centres also provide AT in the rural areas of specified regions.

and some NGOs not supported by the government also give out assistive devices. As many as 25 out of 31 states and union territories have their own schemes of providing AT through government hospitals and voluntary agencies.54 Many district rehabilitation centres also provide AT in the rural areas of specified regions. The Government of India launched the scheme of Science and Technology—Project in Mission Mode 1988, to provide suitable and cost-effective aids and appliances and improve mobility.54 The project concentrated on developing cost-effective technologies and appliances and guiding these to large-scale use. Furthermore, there are initiatives to encourage proposals, research and development of technological solutions for older adults. One such example is the Initiative on Science and Technology Interventions 2010 for the elderly population, introduced by the Department of Science and Technology, under Science for Equity, Empowerment and Development Division (SEED). The initiative included a number of valuable indicatives (see online supplementary Appendix F).55 Turkey No further initiatives were found. Zimbabwe No further initiatives were found. However, assistive devices and equipment for persons with disabilities are partially financed by Government Ministries, persons with disabilities and NGOs. The Government of Zimbabwe is involved in the provision of assistive devices, and provides prostheses/orthoses, wheelchairs, crutches and hearing devices.56

found. However, assistive devices and equipment for persons with disabilities are partially financed by Government Ministries, persons with disabilities and NGOs. The Government of Zimbabwe is involved in the provision of assistive devices, and provides prostheses/orthoses, wheelchairs, crutches and hearing devices.56 Discussion The systematic review found that AT designed for older adults, have received some attention in LMIC. The publications (n=17) retrieved, heavily focused on products that assisted older adults with existing impairment and disability and only a few projects17 20 22 24 30 focused on AT designed for preventing impairment and disability among older adults (ie, fall preventive devices or smart home technologies). While existing disability related AT (eg, prosthetics, orthotics) are somewhat available in LMIC, other products designed to prevent impairment, disability and other health outcomes among older adults have received minimal attention, putting aside the needs of older adults who do not have existing disabilities but are at risk for developing impairment and disability due to increasing age, who would otherwise benefit from AT. Therefore, AT for the entire ageing populations including older adults with disabilities and older adults without existing disabilities but are at high risk for impairment and disability, such as those who are living alone at home, are necessary.

nd disability due to increasing age, who would otherwise benefit from AT. Therefore, AT for the entire ageing populations including older adults with disabilities and older adults without existing disabilities but are at high risk for impairment and disability, such as those who are living alone at home, are necessary. Based on the country-level findings and an overall analysis of all materials gathered for this systematic review, a set of high-level findings are presented and discussed below that could be considered as broad action points in moving forward with the relevant agenda on AT for ageing populations in LMIC.

nd disability due to increasing age, who would otherwise benefit from AT. Therefore, AT for the entire ageing populations including older adults with disabilities and older adults without existing disabilities but are at high risk for impairment and disability, such as those who are living alone at home, are necessary. Based on the country-level findings and an overall analysis of all materials gathered for this systematic review, a set of high-level findings are presented and discussed below that could be considered as broad action points in moving forward with the relevant agenda on AT for ageing populations in LMIC. Promoting initiatives for low-cost AT In many LMIC, production of AT is low; where access might be possible, costs are excessive.15 20 However, this systematic review found that of those LMIC that have AT initiatives in place, typically for disabled populations, many have started the production of low-cost AT (ie, Brazil, Cambodia, Egypt and India). In Cambodia, the majority of the technological solutions provided included low-cost or even, no-cost AT, which are produced by using locally available materials and designs.50 The construction of low-cost equipment in Brazil is a model example, where production cost of devices was 30% lower compared with the cost of imported equipment.20 In the past 5 years, Assistive Technology Centers in Brazil have been developing AT similar to the commercial products, but with use of low-cost material and price reduction of up to 70% (see online supplementary appendix G, table 6).34 Strategies such as producing low-cost products by using locally available materials increase potential for continuous growth and demand for AT and allow for mass production to all those needing them. It is one of the most suitable strategies for LMIC, and highly recommended for resource-poor contexts.57 Mass production can lower costs if the technologies follow universal design principles and are marketed widely. ALIMCO in India is an example where large amount of AT are produced and marketed internationally.53 In the countries that have large production, markets should be expanded beyond regional boundaries to allow for cost advantages, operational efficiency and lower variable costs. In particular, there is a priority need for producers of AT to tailor design, distribution and cost to the realities and demand of older adults.

the countries that have large production, markets should be expanded beyond regional boundaries to allow for cost advantages, operational efficiency and lower variable costs. In particular, there is a priority need for producers of AT to tailor design, distribution and cost to the realities and demand of older adults. In terms of production, many LMIC produced AT locally. Cambodia and India produce a considerable number of AT and components for AT with materials indigenous in these countries. This may be partly explained by the fact that the production workshops or companies themselves are partially or fully supported by or working in partnership with the government. Providing funds for local workshops such as the Egyptian Fund for Supporting Small Industries,48 further encourages local production. Manufacturing or assembling products locally, using local materials, can ensure that devices are suitable for the context.58 It may either be a locally produced, finished product or locally assembled products with imported materials. Where local production is not sustainable, reducing import and duty taxes on AT devices or components can help reduce costs. For example, Brazil does not impose import or duty taxes on imported materials,59 Turkey has exemptions for imported goods and partial exemptions for selected equipment, and Zimbabwe also allows for exemption from paying import tax for some equipment produced for older adults and disabled citizens.45 Therefore, local production and reduction of taxes, where import is required, are appropriate strategies that governments can acknowledge and support.

al exemptions for selected equipment, and Zimbabwe also allows for exemption from paying import tax for some equipment produced for older adults and disabled citizens.45 Therefore, local production and reduction of taxes, where import is required, are appropriate strategies that governments can acknowledge and support. Awareness raising and capacity building on AT Producing and providing AT in and of themselves is not adequate. Significantly fewer AT users in rural areas receive guidance and instructions on the use of AT, and only a few older adults are aware of the existing AT and their benefits in LMIC. For example, a major constraint to the availability of AT in India has been partly the lack of awareness among the users and professionals.24 Even if schemes are available to provide free or subsidised AT, unless users and professionals are aware of their existence and benefits, AT will not be able to prove their worth and purposes. Therefore, information sharing, awareness raising and training are all crucial for successful implementation and usability of AT. In addition, difficulty of use is one of the major reasons for disuse of AT among older adults.60 Our review found that provision of community-based skills programmes has increased total performance and sense of security with the devices to a greater extent among older adults (ie, wheelchairs) in Turkey.32 Hence, the provision of training along with guidance and instruction on the use of AT will reduce the hesitancy to use AT among older adults.

of community-based skills programmes has increased total performance and sense of security with the devices to a greater extent among older adults (ie, wheelchairs) in Turkey.32 Hence, the provision of training along with guidance and instruction on the use of AT will reduce the hesitancy to use AT among older adults. Adequately trained professionals are also an important part of successful implementations and use of AT. Many developing countries including Brazil, India and Zimbabwe58 are lacking occupational therapist (OT) and physical therapists (PT) mainly due to lower pay scales and lack of college or university training programmes. AT must be ensured to be appropriate, adaptable and high quality before distributing to older adults. They must suit the environment and the user and include adequate follow-up in order to ensure safety and efficient use. Especially when countries receive AT as donations from HIC, they need to be assured for quality and safety. For example In Egypt, a highlighted concern is that AT are largely inappropriate, and that there are no quality standards or criteria for monitoring of these services.61 Therefore, implementing policies and programmes for training professionals to assess, provide and promote AT are relevant approaches to successfully providing AT to older adults in LMIC. Furthermore, while training OT and PT, it is important to include other experts in the field such as geriatricians to foster a multidisciplinary team approach in order to provide appropriate technologies and services. Standards and guidelines on provision, development, production and distribution of AT in LMIC must be addressed with the input of experts in the field in collaboration with other countries.

field such as geriatricians to foster a multidisciplinary team approach in order to provide appropriate technologies and services. Standards and guidelines on provision, development, production and distribution of AT in LMIC must be addressed with the input of experts in the field in collaboration with other countries. Bridging the gap between AT policy and practice for older adults The review findings point to the existing gap between legislation and practice. Examples (table 4) were found with some legislation and policies, but whether they have resulted in significant improvement in provision of AT for all older adults in practice is unclear. This is partly due to lack of documentation or lack of evidence available in the literature. Countries that have absolutely no legislation could possibly start with ratifying the UN-CRPD. Countries that have ratified the UN-CRPD and have legislation in place could make amendments and adopt new legislation that is more focused on improving function and well-being of older adults.

ce available in the literature. Countries that have absolutely no legislation could possibly start with ratifying the UN-CRPD. Countries that have ratified the UN-CRPD and have legislation in place could make amendments and adopt new legislation that is more focused on improving function and well-being of older adults. Most often, planning for and responding to health and social needs of older adults falls under either the Ministry of Social Solidarity and/or Ministry of Health, under umbrella terms such as ‘disabled care’ or ‘family care’ or ‘family medicine’ (ie, in Egypt).48 This partly explains the fewer existing legislation that focus on providing AT for older adults. Legislation tends to be specific to a population group (such as the disabled), or to specific health technologies, but rarely linking the two. There is thus, a need to help countries further develop legislation that supports both production and financing of AT along with their integration with health and social services, and in particular for older adult's functional and cognitive needs.

sabled), or to specific health technologies, but rarely linking the two. There is thus, a need to help countries further develop legislation that supports both production and financing of AT along with their integration with health and social services, and in particular for older adult's functional and cognitive needs. Fostering targeted research on AT An interesting finding of this review that offers opportunities for further research is that AT for mobility, especially canes, were the most commonly used with the least difficulty in usability among older adults. It may be explained, in part, that canes are an inexpensive, versatile and light weight mobility option that are well suited for LMIC contexts. Production costs are low and products can be made by material readily available such as wood; hence may be among the most appropriate AT in LMIC. While mobility aids provide assistance, support, strength and balance and also prevent falls, some studies5 62 have reported that mobility aids such as canes could also cause falls. Therefore, additional research is required to understand the mechanisms leading to negative consequences, as well as to better design AT and identify use and environmental measures to prevent such negative consequences along with maximising appropriateness of different AT.

ity aids such as canes could also cause falls. Therefore, additional research is required to understand the mechanisms leading to negative consequences, as well as to better design AT and identify use and environmental measures to prevent such negative consequences along with maximising appropriateness of different AT. The review also found that benefits of AT are rated high and disadvantages low among older adults.21 As seen in the Cambodian study,21 availability of AT to fulfil ADL have increased self-reliance and contribution to the community by physically vulnerable older adults. Improvement in social status, self-esteem, not being a burden to others, improved well-being and economic situation are among the achieved benefits of AT. Whether older adults consider AT as beneficial or not beneficial in LMIC is an important precondition for successful implementations of AT. Therefore, attitudes, aspirations and demands towards technology contribute to their usefulness and usability and perceived advantages among the users, and must be explored further prior to supplying AT to older adults.

eficial or not beneficial in LMIC is an important precondition for successful implementations of AT. Therefore, attitudes, aspirations and demands towards technology contribute to their usefulness and usability and perceived advantages among the users, and must be explored further prior to supplying AT to older adults. Another observation from the review is that of the countries that had available AT, such AT were found to be suitable for the local population. For example, in Cambodia, the most common AT were different forms of seating, which enabled individuals to sit comfortably and bathe from a water source. In Cambodia, it is traditional to bathe by scooping water from a source, using a household cup or bowl and pouring it over the body. Similarly, in Turkey, walking sticks carry traditional importance in some regions, and these were one of the most used mobility aids in Turkey. When introducing AT to LMIC settings, it is important to take into account traditional and cultural values and provide AT that are likely to be accepted by the targeted population. Social acceptability of AT contributes to the usability of devices and avoids development of devices that may be stigmatised in the society. When there is more social acceptability, users are more likely to use their devices.

and cultural values and provide AT that are likely to be accepted by the targeted population. Social acceptability of AT contributes to the usability of devices and avoids development of devices that may be stigmatised in the society. When there is more social acceptability, users are more likely to use their devices. In addition, in future research, education and action on innovation, AT require to be seen differently, which is as highly useful products and devices that support older adults with existing impairment and disability, and as a way to prevent severe health outcomes among older adults who do not yet have disabilities.

and cultural values and provide AT that are likely to be accepted by the targeted population. Social acceptability of AT contributes to the usability of devices and avoids development of devices that may be stigmatised in the society. When there is more social acceptability, users are more likely to use their devices. In addition, in future research, education and action on innovation, AT require to be seen differently, which is as highly useful products and devices that support older adults with existing impairment and disability, and as a way to prevent severe health outcomes among older adults who do not yet have disabilities. Limitations The systematic review retrieved a limited number of relevant publications published in peer-reviewed journals, for each country, in English. This can be partly explained by the limited research in the field of AT, especially in LMIC. In many LMIC, numerous other health issues (eg, treatment of communicable and non-communicable diseases; health insecurity) are considered as pressing priorities. Only recently did AT for older adults are becoming a health priority consideration for LMIC. At the same time, there are a number of barriers that keep LMIC from having ready access to affordable technologies. These include cost, limited funding, lack of awareness on existing AT and their benefits, cultural perceptions (or misperceptions), inappropriate environmental conditions, heterogeneity of older adults terms of functional limitations and comorbidities and a limited number of trained practitioners to train individuals on the use of AT.19 48 Lack of research conducted on the subject in resource-poor contexts strongly emphasises the need for more research on AT for older adults in LMIC.

ns, heterogeneity of older adults terms of functional limitations and comorbidities and a limited number of trained practitioners to train individuals on the use of AT.19 48 Lack of research conducted on the subject in resource-poor contexts strongly emphasises the need for more research on AT for older adults in LMIC. Another limitation is that the search was limited to English language. It is possible that many countries have some research published in different languages. For example, in the case of Egypt, most documents could have been published in Arabic and not translated to English, which is possibly a reason for the lack of evidence available on Egypt. Similarly with Brazil, the authors came across material that was not published in English. Additional limitations may include those related to lack of consistency in the use of the AT terminology; and lack of awareness or knowledge base to date of the needs of older populations in the six study countries. Despite the limitations, the authors believe that the results are important and represent the most comprehensive information currently available on AT (in English) for older adults in these LMIC.

T terminology; and lack of awareness or knowledge base to date of the needs of older populations in the six study countries. Despite the limitations, the authors believe that the results are important and represent the most comprehensive information currently available on AT (in English) for older adults in these LMIC. Conclusion and recommendation AT for ageing populations have received some attention in LMIC. AT that assisted older adults with existing impairment and disability have received more attention than AT that are designed for preventing impairment and disability among older adults who are inevitably at a high risk for impairment and disability as a result of ageing. More systematic planning and efforts are required to design, produce and ensure the availability of affordable, acceptable, accessible and adaptable AT to support older adults in meeting specific functional and cognitive limitation needs of a rapidly growing ageing population in the midst of advancements in assistive technological innovations globally.

ts are required to design, produce and ensure the availability of affordable, acceptable, accessible and adaptable AT to support older adults in meeting specific functional and cognitive limitation needs of a rapidly growing ageing population in the midst of advancements in assistive technological innovations globally. With rapidly increasing ageing populations in LMIC, as exemplified in the six countries selected for this review, AT for the entire ageing population including older adults with existing disabilities and without existing disabilities, but are at high risk for impairment and disability, are necessary with the support of national enabling legislation. Analysis of review findings indicate the need for a comprehensive, integrated health and social system approach to increase the current availability of AT for ageing populations in LMIC. These would entail, yet not be limited to, work on: (1) promoting initiatives for low-cost AT; (2) awareness raising and capacity building on AT; (3) bridging the gap between AT policy and practice; and (4) fostering targeted research on AT.

ach to increase the current availability of AT for ageing populations in LMIC. These would entail, yet not be limited to, work on: (1) promoting initiatives for low-cost AT; (2) awareness raising and capacity building on AT; (3) bridging the gap between AT policy and practice; and (4) fostering targeted research on AT. Supplementary Material Web appendices Web table This systematic review was carried out in 2014 by the WHO Centre for Health Development, also known as the WHO Kobe Centre (WKC), scientific work stream to build the evidence base for Innovation for Healthy Ageing (IHA). The authors would like to thank the technical inputs, review and kind assistance of WHO/WKC colleagues, namely: Mr Loic Garçon, Technical Officer, WKC; Mr Amit Prasad, Technical Officer, WKC; Mr Paul Rosenberg, Technical Officer, WKC; Mr Chapal Khasnabis, Technical Officer, WHO HQ NMH/VIP/DAR; Dr Said Arnaout, Coordinator, Healthy Life Style Team, WHO Regional Office for the Eastern Mediterranean (WHO EMRO); Egyptian team from Ain Shams University—Dr Sarah Hamza, Dr Mostafa Manar, Dr Mousa Shereen, Dr Shawky Ahmed, Dr Rashidy Doha and Dr Hala Sakrha. They also would like to acknowledge and appreciate the support of other members of the IHA team, WKC. Competing interests: None declared. Provenance and peer review: Not commissioned; externally peer reviewed.

Introduction The WHO, in its report on the global status on alcohol and health, has clearly highlighted that alcohol use disorders are a major worldwide problem that has resulted in millions of deaths each year.1 Of particular concern, it should also be noted that the highest alcohol consumption rates are usually found in the developed world, including most parts of Western and Eastern Europe. In addition, higher income countries generally also have the highest alcohol rates.1 Typically hazardous levels of drinking for males would be drinking more than 40 g/day alcohol on average (which is around 5 units) and for females, drinking more than 24 g/day (which is around 3 units).2 Hazardous drinking habits have been linked with increased incidences of liver disease, hypertension and violent deaths for males; and for females, increased incidences of developing liver diseases and breast cancer.2 In Canada, a recent study looking at data obtained from the Centre for Addiction and Mental Health in Ontario between the period of January 2005 and December 2010, and involving a cumulative sample size of 13 557 individuals, has noted that the prevalence of alcohol usage and risky drinking was much lower in foreign-born as compared to Canadian-born individuals.3 Ethnicity was noted to be a mediating factor affecting whether individuals were at-risk heavy drinkers.3 In its report, the WHO has clearly stated that governmental organisations now have increased recognition of the gravity of alcohol-use disorders and its complications and have instituted various measures to help cope with the increasing prevalence and mortality arising from risky usage of alcohol.1

isk heavy drinkers.3 In its report, the WHO has clearly stated that governmental organisations now have increased recognition of the gravity of alcohol-use disorders and its complications and have instituted various measures to help cope with the increasing prevalence and mortality arising from risky usage of alcohol.1 In the past decade, there have been major advances in Internet and Smartphone-based technologies that have had an impact on medicine and patient care. There are existing published papers on the utilisation of smartphone technologies for chronic disease management,4 patient monitoring,5 rehabilitation6 and medical diagnosis,7 as well as for outreach efforts in developing countries.8 Such advances in Internet and Smartphone-based technologies have also been utilised to help individuals with alcohol use disorders. Examples of such interventions include the usage of text messages as a mechanism of behavioural support for youths who are at risk of hazardous drinking.9 Previous evaluations have identified that such a text-message based intervention has produced small reductions in self-reported binge drinking and the number of drinks consumed per drinking day.9 Other interventions, such as the Location-Based Monitoring and Intervention for Alcohol use Disorder (LBMI-A)10 include features such as interventions for on-going drinking, cravings, social connections, management of life problems, high-risk location alerting and activity scheduling. These features have demonstrated clinical efficacy as there have been significant reductions in the amount of hazardous alcohol use after an intervention,10 with the intervention largely catered for individuals diagnosed with alcohol-use disorder. The drink monitor toolkit featured in LBMI-A allows the users to press a button within the application when they are drinking, hence enabling measurement of drinking in vivo. Other research has also demonstrated the clinical efficacy of a smartphone application to help support recovery from alcoholism.11

-use disorder. The drink monitor toolkit featured in LBMI-A allows the users to press a button within the application when they are drinking, hence enabling measurement of drinking in vivo. Other research has also demonstrated the clinical efficacy of a smartphone application to help support recovery from alcoholism.11 A search through the existing published applications on both the Apple IOS and the Android Play store, using the keyword ‘Alcohol’, yielded a total of 1378 applications on the Apple IOS store alone. Prior research12 shows a content analysis of 500 smartphone applications made available on the store and have highlighted that the vast majority of applications (50%) are not clinical-based, and largely for entertainment purposes. An estimated 39% of the applications are blood alcohol concentration (BAC) applications, and only 11% of the applications were noted to be health promotion or stop drinking applications.12 One of the issues highlighted by the previous review and analysis was that the vast majority of applications appeared to help promote drinking and even though there were applications estimating BAC, they were highly unreliable. Of particular concern was that applications also gave a specific time to which users could return to their normal activities, such as drinking.12 One of the recommendations based on the review includes the recommendation that health professionals need to get more involved and help to regulate and review the evidence base of existing published applications. A recent review looking into the behaviour change techniques in popular alcohol reduction apps has highlighted that the vast majority of apps still implicitly or explicitly promoted the usage of alcohol.13 Again, issues with regard to the evidence base of the app were raised. Taking this into consideration, the following study aims to illustrate how the authors have made use of innovative methodologies to overcome the issues relating to the accuracy of tracking the amount of alcohol one has consumed by means of an innovative smartphone application; and the following study also aims to determine user perceptions about the innovative tracker and various other features of an alcohol self-management application among a group of individuals from the general population of a developed country (Canada).

l one has consumed by means of an innovative smartphone application; and the following study also aims to determine user perceptions about the innovative tracker and various other features of an alcohol self-management application among a group of individuals from the general population of a developed country (Canada). The difference between the current innovation and the innovations previously reviewed in the existing literature is that the current innovation serves to target at-risk drinkers, and less so, those who are diagnosed with alcohol-use disorders, or those who have a history of alcohol-use disorders and are in remission. Methodology The alcohol self-management application The first author, MWBZ, in consultation with the last author, RCMH, formulated the initial scope of the application and delineated the contents to be included prior to the actual development of the smartphone application. The development of the application has been previously described by Zhang et al14 in their video commentary to BMJ Innovations. The following contents were included as they were deemed to be helpful for those who are at-risk drinkers and for those who are abusing or are dependent on alcohol. Unique alcohol tracker that enables individuals to log down the absolute number of beers, wine and shots they have taken on a daily basis; this in turn would be translated into units of alcohol and stored in the smartphone application

e who are at-risk drinkers and for those who are abusing or are dependent on alcohol. Unique alcohol tracker that enables individuals to log down the absolute number of beers, wine and shots they have taken on a daily basis; this in turn would be translated into units of alcohol and stored in the smartphone application Unique imediate notification if users have exceeded the limit for the day/or their limit for the week (The notifications are pre-programmed based on the recommended number of units for males and females in accordance with the National Institute for Health and Care Excellence (NICE) (UK) and the Canadian Guidelines.) Calendar view of the total number of units of alcohol consumed on a weekly basis Immediate links to a hotline for those who wish to seek help for their alcohol issues Information about signs and symptoms of alcohol abuse and dependence Information about the medical complications that might arise from both short-term and long-term alcohol usage Psychological therapies, which included that of a functional analysis chart as well as a behavioural goals chart Integration of an Alcohol Use Disorders Identification Test (AUDIT) questionnaire for self-monitoring and diagnosis.

Information about the medical complications that might arise from both short-term and long-term alcohol usage Psychological therapies, which included that of a functional analysis chart as well as a behavioural goals chart Integration of an Alcohol Use Disorders Identification Test (AUDIT) questionnaire for self-monitoring and diagnosis. Following the formulation of the scope of the application as well as defining the contents of the application, the development of the android version of the application was undertaken by the first author, MWBZ, in consultation with a mobile application software company. Android Java developmental kit (ADK) was utilised to programme the core features of the application. MySQL programming language was utilised to programme the core features of the server to gather the survey results of the AUDIT questionnaire. The entire developmental phase took approximately 6 weeks to complete, with 2 weeks of testing prior to deployment onto the android play store. The application was made available on the android play store since 4 February 2015.

mme the core features of the server to gather the survey results of the AUDIT questionnaire. The entire developmental phase took approximately 6 weeks to complete, with 2 weeks of testing prior to deployment onto the android play store. The application was made available on the android play store since 4 February 2015. In order to determine user perspectives towards this new innovative application, the authors took advantage and made use of crowdsourcing to acquire user perspectives. Crowdsourcing is a method in which online users could help to perform specific tasks in exchange for a nominal fee. A total of 100 Canadians aged between 18 and 75 years were recruited online by means of the crowdsourcing website. All participants were paid US$0.10 for their participation in the evaluation and the subsequent 10 min survey. All the participants had to consent to the terms and conditions on the crowdsourcing website prior to their participation in the study. In order to ensure the reliability of the participants, the approval rating for the participants was set at 90%, which meant that they had had past experiences and were validated to help out in another survey. A minimum time of 200 s was set for each questionnaire, and participants needed to spend a minimum of 200 s to fill up the survey. This was to ensure that participants did not fill up the questionnaire at random. A demographic survey was crafted, along with a user perspective survey, which asked questions largely pertaining to the user's perception of the individual features of the application. Participants were to rate their perception on a five-point Likert rating scale. The survey was disseminated via the crowdsourcing website and participants were allowed 2 weeks for the completion of the survey.

ich asked questions largely pertaining to the user's perception of the individual features of the application. Participants were to rate their perception on a five-point Likert rating scale. The survey was disseminated via the crowdsourcing website and participants were allowed 2 weeks for the completion of the survey. As this was an application developed and launched internationally, the authors MWBZ and RCMH worked with collaborators from Shandong University and the University of British Columbia, Vancouver BC. The Institutional Review Board of the Department of Medicine, Shandong University, People's Republic of China approved of this international collaborative research. All statistical analyses were performed using the SPSS statistical program V.18.0 for Windows. Categorical variables were expressed by number (N) and percentage (%). Continuous variables were expressed as mean and SD. Further analysis of the data-set were conducted using T-test. Statistical significance was set at p<0.05 for all analyses.

yses were performed using the SPSS statistical program V.18.0 for Windows. Categorical variables were expressed by number (N) and percentage (%). Continuous variables were expressed as mean and SD. Further analysis of the data-set were conducted using T-test. Statistical significance was set at p<0.05 for all analyses. Results Demographic information A cumulative total of 100 Canadians took part in the user perception survey. 40% were male, and 58% were female, with 2% preferring not to disclose their gender. The vast majority of individuals was from Ontario (58%) and within the age range of 25–54. 72% were Caucasian in ethnicity, and the vast majority were educated with a Bachelor's degree. Most of our participants were married (58%) and 83% indicated that they owned a smartphone, with 45% using the android platform and 29% using the apple platform. Further analysis also showed that nearly 55% of males were android phone users, whereas only 38% of females were android users. Most of the females were apple users instead. Table 1 shows the basic demographic information of our participants. Table 1 Basic demographic characteristics of participants Province % Response (% of total population in 2013) Ontario 58 (39) Quebec 14 (23) British Columbia 10 (13) Nova Scotia 5 (3) New Brunswick 4 (2) Alberta 3 (11) Manitoba 2 (4) Newfoundland and Labrador 1 (2) Unknown 3 (–) Age Range Proportion (%) 18–24 16 25–34 28 35–44 21 45–54 20 55–64 12 65–74 3 Ethnicity Proportion (%) White 72 Asian 20 Hispanic 2 American Indian 2 Other 2 Prefer not to answer 2 Education Proportion (%)

Province % Response (% of total population in 2013) Ontario 58 (39) Quebec 14 (23) British Columbia 10 (13) Nova Scotia 5 (3) New Brunswick 4 (2) Alberta 3 (11) Manitoba 2 (4) Newfoundland and Labrador 1 (2) Unknown 3 (–) Age Range Proportion (%) 18–24 16 25–34 28 35–44 21 45–54 20 55–64 12 65–74 3 Ethnicity Proportion (%) White 72 Asian 20 Hispanic 2 American Indian 2 Other 2 Prefer not to answer 2 Education Proportion (%) Some high school, no diploma 4 High school/HED 19 Some college, no degree 18 Technical/trade/vocational training 8 Associate degree 9 Bachelor's degree 27 Master's degree 13 Professional degree 1 Doctoral degree 1 Marital Status Proportion (%) Single, never married 34 Married or domestic partnership 58 Widowed 3 Divorced 4 Separated 1 HED, Higher Education Diploma. Smartphone ownership Further analysis also showed that smartphone ownership was highest among the age group of 35–44 years (91%) and lowest among those aged between 55 and 64 (58%). In addition, the analysis also showed that education attainment was correlated with smartphone ownership. Table 2 shows the percentage of smartphone ownership according to age and educational status. Table 2 Percentage of smartphone ownership according to age and educational status

Smartphone ownership Further analysis also showed that smartphone ownership was highest among the age group of 35–44 years (91%) and lowest among those aged between 55 and 64 (58%). In addition, the analysis also showed that education attainment was correlated with smartphone ownership. Table 2 shows the percentage of smartphone ownership according to age and educational status. Table 2 Percentage of smartphone ownership according to age and educational status Dimension Owns smartphone (%) 18–24 87.5 25–34 89.3 35–44 90.5 45–54 80.0 55–64 58.3 65–74 66.7 Some high school, no diploma 75.0 High school/HED 89.5 Some college, no degree 66.7 Technical/trade/vocational training 87.5 Associate degree 77.8 Bachelor's degree 88.9 Master's degree 92.3 Professional degree 0 Doctorate degree 100 Alcohol consumption and smartphone ownership With regard to alcohol consumption, individuals have been classified into three categories: Heavy frequency consumption (which includes daily drinking, 2–3 times per week or once per week), infrequent consumption (which includes drinking every 2 weeks, every month and social drinking) and never drinking. Our analysis showed that 66% of the individuals have self-reported heavy consumption, while 25% are infrequent drinkers and 9% have never drunk. Of importance, by considering the average frequency of consumption using a six-point scale to represent responses (daily drinking=6, drinking every 2 weeks=3, never=0), males (mean of 3.85) drink more frequently than females (mean of 2.93). The effect size between males and females is moderate (d=0.57). Our analysis also showed that 25–34-year-olds and 35–44-year-olds drink more frequently than the other groups, with an effect size of d=0.84 for those aged 25–34-years and d=0.81 for those between 35 and 44 years. Further analysis showed that 29% of males aged 18–24 were heavy drinkers with an iPhone or (as another example) 75% of all males aged 35–44 were heavy drinkers with an Android phone.

frequently than the other groups, with an effect size of d=0.84 for those aged 25–34-years and d=0.81 for those between 35 and 44 years. Further analysis showed that 29% of males aged 18–24 were heavy drinkers with an iPhone or (as another example) 75% of all males aged 35–44 were heavy drinkers with an Android phone. Perceptions of application The user perception survey asked participants to rate multiple dimensions of the application using a five-point Likert scale. The mean survey scores for each dimension are presented in table 3. The six dimensions have good internal consistency—with a Cronbach's α of 0.95. Table 3 App perceptions Users Mean score (SD) How useful is the alcohol tracker? 1.67 (1.14) How useful are the provided psychological therapies? 1.57 (1.12) How useful is the information on alcohol dependence? 1.83 (1.21) How useful is the notification service? 1.74 (1.24) How useful is the audit questionnaire? 1.61 (1.22) Overall, how useful is the app? 1.78 (1.21) Usefulness rating scale: 4=extremely; 3=very; 2=moderately; 1=slightly; 0=not at all.

therapies? 1.57 (1.12) How useful is the information on alcohol dependence? 1.83 (1.21) How useful is the notification service? 1.74 (1.24) How useful is the audit questionnaire? 1.61 (1.22) Overall, how useful is the app? 1.78 (1.21) Usefulness rating scale: 4=extremely; 3=very; 2=moderately; 1=slightly; 0=not at all. Results suggest that notification and information were the two most useful functions, with psychotherapy expected to be the least useful. Perception appears to depend on gender, with females slightly more positive about the app (and its individual functions) than males, as illustrated in table 4. Females reported a higher mean score than males on every domain, but the difference was not significant. The only domain where there appeared to be a large difference in perception was the alcohol tracker (d=0.4, p=0.06), which suggests that females may be more likely than males to use such a functionality. Females indicated that notification service was the most useful function, while males preferred the information component. In addition, the results in table 5 indicated that the app was rated most useful overall by those in the youngest age group (18–24), and least useful overall by those aged 45–54. The notification service was the highest rated function for those aged 18–44, while for older groups information was rated as the most useful function. Table 4 App perceptions in accordance with gender

Results suggest that notification and information were the two most useful functions, with psychotherapy expected to be the least useful. Perception appears to depend on gender, with females slightly more positive about the app (and its individual functions) than males, as illustrated in table 4. Females reported a higher mean score than males on every domain, but the difference was not significant. The only domain where there appeared to be a large difference in perception was the alcohol tracker (d=0.4, p=0.06), which suggests that females may be more likely than males to use such a functionality. Females indicated that notification service was the most useful function, while males preferred the information component. In addition, the results in table 5 indicated that the app was rated most useful overall by those in the youngest age group (18–24), and least useful overall by those aged 45–54. The notification service was the highest rated function for those aged 18–44, while for older groups information was rated as the most useful function. Table 4 App perceptions in accordance with gender Females Males Mean score (SD) Mean score (SD) How useful is the alcohol tracker? 1.86 (1.12) 1.40 (1.22) How useful are the provided psychological therapies? 1.69 (1.16) 1.43 (1.22) How useful is the information on alcohol dependence? 1.83 (1.17) 1.65 (1.23) How useful is the notification service? 1.97 (1.31) 1.63 (1.15) How useful is the audit questionnaire? 1.62 (1.14) 1.55 (1.30) Overall, how useful is the app? 1.84 (1.15) 1.63 (1.29) Usefulness rating scale: 4=extremely; 3=very; 2=moderately; 1=slightly; 0=not at all.

tion on alcohol dependence? 1.83 (1.17) 1.65 (1.23) How useful is the notification service? 1.97 (1.31) 1.63 (1.15) How useful is the audit questionnaire? 1.62 (1.14) 1.55 (1.30) Overall, how useful is the app? 1.84 (1.15) 1.63 (1.29) Usefulness rating scale: 4=extremely; 3=very; 2=moderately; 1=slightly; 0=not at all. Table 5 App perceptions in accordance with age group

tion on alcohol dependence? 1.83 (1.17) 1.65 (1.23) How useful is the notification service? 1.97 (1.31) 1.63 (1.15) How useful is the audit questionnaire? 1.62 (1.14) 1.55 (1.30) Overall, how useful is the app? 1.84 (1.15) 1.63 (1.29) Usefulness rating scale: 4=extremely; 3=very; 2=moderately; 1=slightly; 0=not at all. Table 5 App perceptions in accordance with age group 18–24 25–34 35–44 45–54 55–64 65–74 Tracker 1.63 1.68 1.57 1.65 1.75 2.33 Psychotherapy 1.50 1.32 1.81 1.75 1.50 2.00 Information 1.50 1.68 1.90 1.80 1.92 2.33 Notification 1.75 1.93 1.95 1.60 1.75 2.00 Audit 1.25 1.68 1.81 1.50 1.33 2.67 Overall 1.94 1.68 1.90 1.55 1.67 2.00 Discussion On the basis of our current knowledge, this is perhaps one of the first few studies to describe the methodology of developing a new alcohol intervention application and one of the first few studies to evaluate a new methodology of tracking alcohol consumption. The current study was methodologically feasible and the initial users’ perspectives demonstrated the usefulness of certain components of the current smartphone application. The vast majority of our sampled users was female and had at least a Bachelor's degree. Of significance, our results showed that the smartphone ownership rates were highest among the 35–44 years age group, and correspondingly lower in the 55–64 years age group. This finding is not surprising, given that other studies have clearly identified that smartphone utilisation is highest among the younger population. In our sampled population, 66% of the users reported heavy consumption of alcohol (daily drinking), and the incidence of drinking was also highest among the younger as compared to the older age group. Interestingly, our results showed that 75% of males aged between 35 and 44 years were heavy drinkers and owned only an Android phone. This in itself would have downstream implications for application design and deployment. Pertaining to users’ perception of the features within the application, the results we have obtained suggested that notification services and information were perceived to be the most useful functionality, while psychotherapy was correspondingly being perceived to be least useful. There appears to be a gender difference in users’ preference as well, with females tending to prefer the notification services, whereas males tend to prefer the information service. Age is another mediating factor affecting the perception of the usefulness of notification services versus informational content.

useful. There appears to be a gender difference in users’ preference as well, with females tending to prefer the notification services, whereas males tend to prefer the information service. Age is another mediating factor affecting the perception of the usefulness of notification services versus informational content. Previous research has highlighted the limitations with regard to the BAC methodology of tracking and notifying users with regard to the drinking habits.15 Among 28 574 students sampled, it was noted that such computation did not have an effect on the absolute amount of alcohol consumption and that it might even lead to more alcohol consumption in certain genders, such as the male gender.15 Prior evaluation of existing applications, which included that of ‘Promillekoll’ and ‘Party Planner’, has highlighted that in order for applications to be more effective, applications need to enable users to be able to manipulate how they track their own alcohol consumption, as well as integrate features that would appeal to both genders.15 This might be one of the reasons why our currently sampled population perceived the notification service to be useful, especially so for at-risk female drinkers. In our current application, we have enabled users to log down the amount of beer, wine and spirits that they have drunk on a daily basis. The smartphone application is able to automatically compute and convert the amount they have drunk into relevant alcoholic units. Notifications would be displayed if they have exceeded their daily limits, namely 14 units for females and 21 units for males. In addition, the notification service and the innovative tracker might be perceived to be useful as it enables real-time tracking of relative consumption. Previous studies looked at in-vivo as well as retrospective reports and highlighted that, unlike in-vivo recording, retrospective recording tends to have limitations, as in, it is very likely to result in underestimation of the actual amount of alcohol which has been consumed.16

l-time tracking of relative consumption. Previous studies looked at in-vivo as well as retrospective reports and highlighted that, unlike in-vivo recording, retrospective recording tends to have limitations, as in, it is very likely to result in underestimation of the actual amount of alcohol which has been consumed.16 Our current findings are also in line and congruent with the findings of other previous interventions for alcohol use disorders.17 Previous interventions have determined that individuals tend to find the following features most useful in a phone, which are that of provision of awareness as well as tracking and prompts.17 The strength of this study is that we managed to make use of existing technological advances to develop a self-help alcohol smartphone application that could circumvent the previous issues pertaining to the concerns of using BAC levels as a tracker. We have managed to evaluate the prototype that we have developed in a group of at-risk and heavy drinkers and have identified their user preferences. Of interest, we managed to also identify the differences between the genders with regard to their perceptions of the usefulness of individual components of the applications. This would be informative and would affect how future applications might be developed.

heavy drinkers and have identified their user preferences. Of interest, we managed to also identify the differences between the genders with regard to their perceptions of the usefulness of individual components of the applications. This would be informative and would affect how future applications might be developed. The limitations in this study include the fact that our sample size is limited to a total size of 100 individuals. Preferences might differ if the sample size was larger. In addition, the educational level of our participants tends to be higher as compared to that of a normal population. Also, the authors have developed only an English version of the application and this limits the further evaluation of the application. In addition, the authors have not included within the application more comprehensive psychological interventions and this might have affected user perceptions. Conclusions This study has demonstrated how the authors have made use of innovative technologies to overcome the existing concern pertaining to the utilisation of the BAC levels as a tracker. In addition, the authors have managed to evaluate the applications in a sample group and have highlighted user preferences that should be considered when future applications for alcohol use disorders are being formulated and developed. Twitter: Follow Melvyn Zhang at @melvynzhang Competing interests: None declared. Ethics approval: Shandong University, China (on behalf of the international consortium): LL-201501062. Provenance and peer review: Not commissioned; internally peer reviewed.

Introduction Five hundred million individuals use mental health apps worldwide, with these numbers expected to reach 1.7 billion by 2018.1 The potential for mobile devices to revolutionise healthcare and clinical research has not been lost on either industry2 3 or academia.4–6 Notable examples of initiatives to collect behavioural data using mobile technology are the Patient-Centered Outcomes Research Institute (PCORI)-funded Patient-Centered Clinical Outcomes Research Networks, National Institutes of Health (NIH's) Precision Medicine Initiative, and Apple's Research Kit. Indeed, the use of smart technology appears to be a clear avenue to increase research participation.7 8 Mobile technologies may be particularly useful in improving participant access to and reducing expenses of randomised clinical trials (RCTs). Typical RCTs cost millions of dollars and recruit 200–300 participants in 3–5 years.9 Sample demographics are determined by the location of the research institution, limiting the representativeness of many RCT samples. Expense and access problems are exacerbated when trying to study populations who are challenging to recruit, such as those with mental illnesses, people living in rural areas or racial/ethnic minority populations.

termined by the location of the research institution, limiting the representativeness of many RCT samples. Expense and access problems are exacerbated when trying to study populations who are challenging to recruit, such as those with mental illnesses, people living in rural areas or racial/ethnic minority populations. One solution to overcome access and cost issues has been the use of the internet to conduct randomised control studies. These trials are beneficial from the cost perspective, with estimated cost reductions of more than 50% compared with conventional trials,10 11 and from the access perspective, these studies recruit very large samples in short periods of time.12 13 However, retention issues are particularly problematic for internet studies, with one recent internet-based trial reporting a drop-out rate of over 90% in a sample of 3000 individuals.14 Drop-out is likely due in part to the need to access a WiFi connection and dependence on an immobile device (eg, a desktop computer). A potential advantage to research using mobile devices (smart phones and tablets) is that data can be collected anywhere at any time. These devices also facilitate passive data collection such as Global Positioning System (GPS) information from the phone's accelerometer and media usage to gauge social and physical activity that can supplement self-reports. However, while mobile technology may be able to further expand the reach of clinical research, this approach has yet to be tested.

ve data collection such as Global Positioning System (GPS) information from the phone's accelerometer and media usage to gauge social and physical activity that can supplement self-reports. However, while mobile technology may be able to further expand the reach of clinical research, this approach has yet to be tested. The purpose of this study is to determine the feasibility of conducting a fully remote RCT using smart devices in depressed adults 18 years old and older. We elected to study depression as our clinical focus given its ubiquitous presence in mental illnesses and disability.15 It is the leading cause of disability worldwide,16 17 and the enrolment of depressed individuals into clinical trials is difficult.18 19 In this paper, we report data on population access (sample representativeness), engagement assessment and cost to complete the study. Methods Ethical approval for the trial was granted by the UCSF Committee for Human Research.

The purpose of this study is to determine the feasibility of conducting a fully remote RCT using smart devices in depressed adults 18 years old and older. We elected to study depression as our clinical focus given its ubiquitous presence in mental illnesses and disability.15 It is the leading cause of disability worldwide,16 17 and the enrolment of depressed individuals into clinical trials is difficult.18 19 In this paper, we report data on population access (sample representativeness), engagement assessment and cost to complete the study. Methods Ethical approval for the trial was granted by the UCSF Committee for Human Research. Recruitment To test our hypotheses about access, we used three different types of recruitment approaches: traditional, social networking and search engine-based methods. Traditional methods were written ads placed in city buses, newspapers and Craigslist postings throughout the USA. Social networking methods included regular postings on sites such as Facebook and Twitter, and contextual-targeting methods to identify and directly push recruitment ads to potential participants, based on their Twitter and other social media comments. Our search engine-based method included using Google Adwords, a historically successful recruitment tool.20 Each approach (described further in see online supplementary materials) provided potentially interested participants a link to our custom study website (http://www.brightenstudy.com).

other social media comments. Our search engine-based method included using Google Adwords, a historically successful recruitment tool.20 Each approach (described further in see online supplementary materials) provided potentially interested participants a link to our custom study website (http://www.brightenstudy.com). Participant eligibility Participants had to speak English, be 18 years old or older, own a smartphone (iPhone or Android) with WiFi or 3G/4G capabilities, and own an iPad2 or newer device. iPad ownership was required as our cognitive assessment tool was only available on this device at the time of the study. To characterise recruiting logistics without this restriction, individuals without an iPad but with a smartphone were given the opportunity to participate in phone-only study arms that were not part of the randomised sample. A Patient Health Questionnaire (PHQ-9),21 score of 5 or greater, or a score of 2 or greater on PHQ item 10 (indicating that they felt disabled in their life because of their mood), was also required for enrolment. Procedure Screening Potential participants were directed to a website (http://www.brightenstudy.com) explaining the study purpose and procedures. Interested participants completed an online brief screening consisting of questions about mobile device ownership.

Participant eligibility Participants had to speak English, be 18 years old or older, own a smartphone (iPhone or Android) with WiFi or 3G/4G capabilities, and own an iPad2 or newer device. iPad ownership was required as our cognitive assessment tool was only available on this device at the time of the study. To characterise recruiting logistics without this restriction, individuals without an iPad but with a smartphone were given the opportunity to participate in phone-only study arms that were not part of the randomised sample. A Patient Health Questionnaire (PHQ-9),21 score of 5 or greater, or a score of 2 or greater on PHQ item 10 (indicating that they felt disabled in their life because of their mood), was also required for enrolment. Procedure Screening Potential participants were directed to a website (http://www.brightenstudy.com) explaining the study purpose and procedures. Interested participants completed an online brief screening consisting of questions about mobile device ownership. Consent We used a combination of a written consent and custom videos posted on YouTube to explain the study. Participants had to pass a quiz that tested their understanding that the study was voluntary, was not a substitute for treatment and that they were to be randomised. Each question had to be answered correctly before moving on to baseline assessment and randomisation. Eligibility was established after consent was obtained.

ticipants had to pass a quiz that tested their understanding that the study was voluntary, was not a substitute for treatment and that they were to be randomised. Each question had to be answered correctly before moving on to baseline assessment and randomisation. Eligibility was established after consent was obtained. Randomisation Participants were randomised to one of three treatment arms where they viewed a brief video explaining how to download and use the assessment and assigned treatment app. Participants were also given a link to view a custom dashboard of their study progress. Treatment Participants were asked to use their assigned app for 1 month. The first app was a cognitive intervention video game (Project: EVO™, or EVO) designed to modulate cognitive control abilities, a common neurological deficit underlying depression.22 The second intervention was an app based on an evidence-based treatment for depression (problem-solving therapy, or PST).23 The final intervention app, an information control, provided daily health tips (HT) for overcoming depressed mood such as self-care (eg, taking a shower) or physical activity (eg, taking a walk; see online supplementary materials for further descriptions of each).

ent for depression (problem-solving therapy, or PST).23 The final intervention app, an information control, provided daily health tips (HT) for overcoming depressed mood such as self-care (eg, taking a shower) or physical activity (eg, taking a walk; see online supplementary materials for further descriptions of each). Assessment We used two apps to collect baseline and 4, 8 and 12 weeks of outcome data. The first app, developed by Ginger.io™ was used to collect self-reported mood, function and passive analytics such as communication data (text logs including call/text time, call duration, text length and screen usage), and mobility data (activity type and distance travelled using the phone's accelerometer and GPS). The second app was a mobile cognitive assessment app (Adaptive Cognitive Evaluation (ACE)), to measure cognitive control processes (see online supplementary etable 1). Participants were automatically notified every 8 h for 24 h if they had not completed a survey within 8 h of its original delivery. An assessment was considered missing if it was not completed within this 24 h time frame.

ive Evaluation (ACE)), to measure cognitive control processes (see online supplementary etable 1). Participants were automatically notified every 8 h for 24 h if they had not completed a survey within 8 h of its original delivery. An assessment was considered missing if it was not completed within this 24 h time frame. The baseline assessment included the collection of demographics including age, race/ethnicity, martial and employment status, income, education, smart device ownership, use of other health apps, and use of mental health services, including use of medications and psychotherapy. We collected information on mental health status using the PHQ-924 for depression, the generalized anxiety disorder (GAD)-7,22 for generalised anxiety, a four-item mania and psychosis-screening instrument25 and the four-item National Institute on Alcohol Abuse and Alcoholism (NIAAA) Alcohol Screening Test.26 To assess for self-reported disability, we used the Sheehan Disability Scale.27 28 We also asked participants to rate their health on a scale of excellent to poor. Daily assessments were a combination of self-report and passive data collection. Participants completed the PHQ-2 (mood and enjoyment) every morning. The Ginger.io app collected passive analytics daily. Private information such as actual content of voice calls or text messages or emails was not collected.

The baseline assessment included the collection of demographics including age, race/ethnicity, martial and employment status, income, education, smart device ownership, use of other health apps, and use of mental health services, including use of medications and psychotherapy. We collected information on mental health status using the PHQ-924 for depression, the generalized anxiety disorder (GAD)-7,22 for generalised anxiety, a four-item mania and psychosis-screening instrument25 and the four-item National Institute on Alcohol Abuse and Alcoholism (NIAAA) Alcohol Screening Test.26 To assess for self-reported disability, we used the Sheehan Disability Scale.27 28 We also asked participants to rate their health on a scale of excellent to poor. Daily assessments were a combination of self-report and passive data collection. Participants completed the PHQ-2 (mood and enjoyment) every morning. The Ginger.io app collected passive analytics daily. Private information such as actual content of voice calls or text messages or emails was not collected. The 4-week, 8-week and 12-week assessments included the PHQ-9 to measure changes in mood, ACE for changes in cognitive control, and the Sheehan for changes in disability. Participants were also asked this question: ‘since using this app, I feel that I am: (1) much worse, (2) worse, (3) no different, (4) improved, (5) much improved’.

week and 12-week assessments included the PHQ-9 to measure changes in mood, ACE for changes in cognitive control, and the Sheehan for changes in disability. Participants were also asked this question: ‘since using this app, I feel that I am: (1) much worse, (2) worse, (3) no different, (4) improved, (5) much improved’. Payment Randomised participants were paid a total of $75 for completing all assessments over the 12 weeks via Amazon gift vouchers, while participants in the phone-only arms were paid $45 as they did not complete the cognitive assessment. To test if increased payment led to increased adherence and retention, a subset of participants (n=144) were given $75 in bonus pay if they completed all assessments. Procedures to reduce gaming ‘Gaming’ is a situation where a user fraudulently enrols in a study solely to acquire research payment. We utilised the following safeguards to prevent this: (1) locking the eligibility survey if a participant tried to change a submitted answer, (2) using study links that are valid for one user/device, and (3) tracking IP addresses to minimise duplicate enrolment.

tly enrols in a study solely to acquire research payment. We utilised the following safeguards to prevent this: (1) locking the eligibility survey if a participant tried to change a submitted answer, (2) using study links that are valid for one user/device, and (3) tracking IP addresses to minimise duplicate enrolment. Statistical analyses To assess participant access, we describe the sample demographics, clinical characteristics and sample comorbidities using the appropriate descriptive statistics. To assess participant engagement, we examined the proportion of study drop-outs and the proportion of enrolled individuals who responded to the primary mood outcome measures at each time point using a mixed-model analysis of variance (with Greenhouse-Geisser corrections when needed). To calculate time to drop-out, we tested a survival analysis model with the distribution of the ‘survival’ times for those assigned each app estimated and non-parametric estimates of the survivor function computed by the Kaplan-Meier method, with curves tested using the log-rank test using Stata V.14.0. We also examined whether there was a significant difference in drop-out rates among the three interventions using Pearson's χ2 test. Pairwise log-rank tests were conducted to determine where there were significant differences between the distributions, and a Bonferroni correction was set at p<0.017 to correct for multiple comparisons. We also compared these outcomes for the entire sample and by sample type (randomised and non-randomised). To assess issues surrounding cost, we describe a total study cost approach factoring in β testing, staff time and efforts beyond those payments made for recruitment and participant remuneration.

ltiple comparisons. We also compared these outcomes for the entire sample and by sample type (randomised and non-randomised). To assess issues surrounding cost, we describe a total study cost approach factoring in β testing, staff time and efforts beyond those payments made for recruitment and participant remuneration. Results Access Recruitment rate National recruitment began in August of 2014, and was conducted in five, 2-week advertising waves (total of 5 months of recruitment). We recruited a total of 2923 participants. Of these recruited individuals, 1098 were enrolled to the randomised (N=626) and non-randomised (N=472) arms of the study (see figure 1). Eighty-nine per cent of the sample came from traditional recruitment approaches, <1% came from social networking, <1% came from search engine-based methods, and 10.3% came from unanticipated means (own search, referrals). We were able to successfully recruit individuals from 8 of the 15 most rural states in the USA29 without any targeted recruitment efforts (see figure 2A). Figure 1 CONSORT diagram (HT, health tips; PHQ, Patient Health Questionnaire). Figure 2 Demographic characteristics. (A) Percentage of recruited participants across the USA. (B) Percentage of participants within different age ranges from the recruited sample. (C) Ethnic composition of the recruited individuals, and its comparison to the observed ethnic composition reported in the 2013 US Census.

Figure 1 CONSORT diagram (HT, health tips; PHQ, Patient Health Questionnaire). Figure 2 Demographic characteristics. (A) Percentage of recruited participants across the USA. (B) Percentage of participants within different age ranges from the recruited sample. (C) Ethnic composition of the recruited individuals, and its comparison to the observed ethnic composition reported in the 2013 US Census. Sample demographics Participants were primarily young adults (see figure 2B), although age ranged from 18 to 76, with 79% identifying as female. Fifty-eight per cent of our sample was non-Hispanic white, and an ethnicity distribution comparable to the 2013 US Census (see figure 2C). Fifty-seven per cent of our participants obtained a 4-year college degree or higher, with a mean annual income of $30–$35 000 (see table 1). Sixty-seven per cent of our participants were employed at the time of enrolment. There was a difference in age between randomised and phone-only assigned participants, with randomised participants slightly older than phone-only participants, t(954.38)=−3.22, p=0.001. However, there was no difference in gender between these groups (=0.08, p=0.77). Enrolled individuals who were single/never married reported greater symptoms of depression (t[528.17]=2.96, p=0.003). Table 1 Baseline demographics

Sample demographics Participants were primarily young adults (see figure 2B), although age ranged from 18 to 76, with 79% identifying as female. Fifty-eight per cent of our sample was non-Hispanic white, and an ethnicity distribution comparable to the 2013 US Census (see figure 2C). Fifty-seven per cent of our participants obtained a 4-year college degree or higher, with a mean annual income of $30–$35 000 (see table 1). Sixty-seven per cent of our participants were employed at the time of enrolment. There was a difference in age between randomised and phone-only assigned participants, with randomised participants slightly older than phone-only participants, t(954.38)=−3.22, p=0.001. However, there was no difference in gender between these groups (=0.08, p=0.77). Enrolled individuals who were single/never married reported greater symptoms of depression (t[528.17]=2.96, p=0.003). Table 1 Baseline demographics Variable PST EVO (iPad) HT (iPad) EVO (iPhone) HT (Android) Age 34.91 (12.33) 33.37 (10.87) 33.56 (12.27) 30.33 (11.11) 32.41 (10.26) Education <12 years 39 (18.48) 31 (14.83) 36 (17.48) 52 (21.85) 77 (32.91) College 133 (63.03) 129 (61.72) 122 (59.22) 148 (62.18) 125 (53.42) Graduate 39 (18.48) 49 (23.44) 49 (23.30) 38 (15.97) 32 (13.68) Income+ $20 000 or less 42 (35.00) 31 (35.23) 39 (31.71) 73 (53.28) 79 (58.96) $20 000–$40 000 32 (26.67) 21 (23.86) 30 (24.39) 37 (27.01) 34 (25.37) $40 000–$60 000 26 (21.67) 17 (19.32) 17 (13.82) 18 (13.14) 14 (10.45) $60 000–$80 000 6 (5.00) 9 (10.23) 18 (14.63) 7 (5.11) 4 (2.99) $80 000–$100 000 6 (5.00) 3 (3.41) 7 (5.69) 1 (0.73) 1 (0.75) $100 000+ 8 (6.67) 7 (7.95) 12 (9.76) 1 (0.73) 2 (1.49) Number of females 160 (75.83) 161 (77.03) 173 (83.98) 198 (83.19) 172 (73.50) Per cent of minority 83 (39.34) 87 (41.63) 82 (39.81) 94 (39.50) 110 (47.01) Marital status Single 118 (55.92) 112 (53.59) 118 (57.28) 164 (68.91) 150 (64.10) Married 62 (29.38) 73 (34.93) 68 (33.01) 53 (22.27) 53 (22.65) Divorced/separated/widowed 31 (14.69) 24 (11.48) 20 (9.71) 21 (8.82) 31 (13.25) Psychiatric PHQ-9 13.76 (4.91) 13.51 (5.06) 13.64 (4.90) 13.61 (5.02) 15.01 (5.23) GAD 10.36 (4.97) 9.15 (4.92) 10.39 (5.32) 10.54 (5.53) 10.38 (5.46) NIAAA 3.20 (2.54) 3.03 (2.24) 3.40 (2.31) 3.34 (2.35) 2.98 (2.43) Mania Hx 22 (14.57) 15 (11.90) 20 (14.71) 25 (16.03) 25 (15.15) Psychosis Hx 2 (1.32) 1 (0.79) 3 (2.21) 3 (1.92) 3 (1.82) Using other health apps 128 (87.67) 105 (84) 116 (86.57) 124 (79.49) 125 (77.64) Per cent of rural 5 (2.37) 3 (1.44) 7 (3.40) 9 (3.78) 5 (2.14) In treatment 84 (57.5) 64 (52.5) 74 (56.5) 84 (55.3) 81 (50.9) Mean (SD) for continuous variables; number (percentage) for categorical.

.32) 1 (0.79) 3 (2.21) 3 (1.92) 3 (1.82) Using other health apps 128 (87.67) 105 (84) 116 (86.57) 124 (79.49) 125 (77.64) Per cent of rural 5 (2.37) 3 (1.44) 7 (3.40) 9 (3.78) 5 (2.14) In treatment 84 (57.5) 64 (52.5) 74 (56.5) 84 (55.3) 81 (50.9) Mean (SD) for continuous variables; number (percentage) for categorical. HT, health tips; Hx, medical history; PHQ, Patient Health Questionnaire; PST, problem solving therapy. Clinical characteristics The sample was moderately depressed at baseline, with a PHQ-9 mean score of 13.9 (SD=5.1). There was a significant association between age and depression severity, Spearman's r=−0.11, p<0.001. There was no significant difference in depression severity among gender (t(365.85)=0.63, p=0.53) or ethnic groups, (F(6, 1091)=1.37, p=0.22). Fifty-one per cent reported comorbid anxiety, 53% reported comorbid alcohol misuse, 16% reported a history of psychosis or mania. In total, 54.5% of our sample was receiving mental health treatment for their depression. This sample mirrored the ethnic disparities in mental health service use found in the general population, with 63% of non-Hispanic white participants in treatment, and only 42% of ethnic minorities were in treatment (=28.6, p<0.001, OR=2.29). There were no significant differences in depression severity among individuals randomised to the three primary arms (F[2, 623]=0.14, p=0.87; see table 1).

d in the general population, with 63% of non-Hispanic white participants in treatment, and only 42% of ethnic minorities were in treatment (=28.6, p<0.001, OR=2.29). There were no significant differences in depression severity among individuals randomised to the three primary arms (F[2, 623]=0.14, p=0.87; see table 1). Engagement Sixty-six per cent of the sample completed the 4-week assessment, 50% completed the 8-week assessment and 41% completed the 12-week assessment (see figure 3A). There was no adherence difference by group (F(2, 241)=2.50, p=0.08) and no time by group interaction (F(3.55, 428.14)=1.93, p=0.11). We found similar adherence to the cognitive assessment tool, with neither a group (F=0.46, p=0.63) nor interaction effect present (F=0.91, p=0.42). Although lower assessment adherence was observed in the more depressed participants, younger participants, and participants with lower education, the effects sizes were small (see table 2). Table 2 Baseline demographics*

Engagement Sixty-six per cent of the sample completed the 4-week assessment, 50% completed the 8-week assessment and 41% completed the 12-week assessment (see figure 3A). There was no adherence difference by group (F(2, 241)=2.50, p=0.08) and no time by group interaction (F(3.55, 428.14)=1.93, p=0.11). We found similar adherence to the cognitive assessment tool, with neither a group (F=0.46, p=0.63) nor interaction effect present (F=0.91, p=0.42). Although lower assessment adherence was observed in the more depressed participants, younger participants, and participants with lower education, the effects sizes were small (see table 2). Table 2 Baseline demographics* Assessment adherence Adherent Non-adherent p Value Effect size† PHQ-9‡ 13.25 (5.0) 14.5 (4.9) 0.001 0.24 Age‡ 34.7 (11.4) 30.7 (10.8) <0.001 0.36 Education§ <0.01 0.25 <12 years 61 (17.09) 95 (25.54) College 206 (57.70) 211 (56.72) Graduate 90 (25.21) 66 (17.74) Gender¶ 80 (22.4) 75 (20.2) 0.46 0.11 Minority** 130 (36.4) 161 (43.3) 0.06 0.23 Income†† 0.06 0.33 $20 000 or less 81 (39.32) 91 (47.40) $20 000–$40 000 50 (24.27) 52 (27.08) $40 000–$60 000 34 (16.50) 30 (15.62) $60 000–$80 000 21 (10.19) 7 (3.65) $80 000–$100 000 6 (2.91) 6 (3.12) $100 000+ 14 (6.80) 6 (3.12) *Means and SDs unless otherwise specified. †All effect sizes converted to Cohen's d. ‡Welch t test. §Pearson χ2 test; number and percentage. ¶Pearson χ2 test; number and percentage of males. **Pearson χ2 test; number and percentage of ethnic minorities. ††Fisher's exact test; number and percentage. PHQ, Patient Health Questionnaire.

Assessment adherence Adherent Non-adherent p Value Effect size† PHQ-9‡ 13.25 (5.0) 14.5 (4.9) 0.001 0.24 Age‡ 34.7 (11.4) 30.7 (10.8) <0.001 0.36 Education§ <0.01 0.25 <12 years 61 (17.09) 95 (25.54) College 206 (57.70) 211 (56.72) Graduate 90 (25.21) 66 (17.74) Gender¶ 80 (22.4) 75 (20.2) 0.46 0.11 Minority** 130 (36.4) 161 (43.3) 0.06 0.23 Income†† 0.06 0.33 $20 000 or less 81 (39.32) 91 (47.40) $20 000–$40 000 50 (24.27) 52 (27.08) $40 000–$60 000 34 (16.50) 30 (15.62) $60 000–$80 000 21 (10.19) 7 (3.65) $80 000–$100 000 6 (2.91) 6 (3.12) $100 000+ 14 (6.80) 6 (3.12) *Means and SDs unless otherwise specified. †All effect sizes converted to Cohen's d. ‡Welch t test. §Pearson χ2 test; number and percentage. ¶Pearson χ2 test; number and percentage of males. **Pearson χ2 test; number and percentage of ethnic minorities. ††Fisher's exact test; number and percentage. PHQ, Patient Health Questionnaire. Figure 3 Intervention and assessment adherence. (A) Percentage of individuals who responded to their mood assessment during the treatment phase (first 4 weeks) and follow-up periods (weeks 8 and 12). (B) Kaplan-Meier survival estimates per study arm illustrating survival distributions of time to drop-out (last day of recorded activity) over the course of the study (84 days).

ntage of individuals who responded to their mood assessment during the treatment phase (first 4 weeks) and follow-up periods (weeks 8 and 12). (B) Kaplan-Meier survival estimates per study arm illustrating survival distributions of time to drop-out (last day of recorded activity) over the course of the study (84 days). Kaplan-Meier survival analysis was conducted to determine whether intervention assignment or any baseline demographic variables predicted drop-out status. The log-rank test revealed a significant difference between the survival distributions between groups (=19.27, p<0.001), with the EVO arm having significantly earlier time to drop-out than the PST arm (=7.45, p=0.01) or HT (=17.51, p<0.001) arms (see figure 3B). We did not find a significant difference in survival distributions for those with high versus low PHQ-9 scores (using a PHQ-9 score of 10 as a cut-point, =2.29, p=0.13). There was no significant difference in survival distributions between non-Hispanic whites and ethnic minorities (=2.13, p=0.14). Participants who received bonus pay remained in the study longer than those who did not receive a bonus (=11.82, p<0.001). Bonus pay was for assessment completion, not intervention app use.

. There was no significant difference in survival distributions between non-Hispanic whites and ethnic minorities (=2.13, p=0.14). Participants who received bonus pay remained in the study longer than those who did not receive a bonus (=11.82, p<0.001). Bonus pay was for assessment completion, not intervention app use. Cost Total study costs included participant payments ($23 320), website/enrolment portal/database development ($46 507), and salaried staff time (3; 2 student volunteers also assisted) over the 9 months the study was active ($58 917), summing to a total of $128 444. The total amount spent on for this project, including staff costs, development and β testing of the UCSF developed apps (ACE and iPST), and licensing fees for the use of the other apps (EVO and Ginger.io) was $314 264 over 2 years.

isted) over the 9 months the study was active ($58 917), summing to a total of $128 444. The total amount spent on for this project, including staff costs, development and β testing of the UCSF developed apps (ACE and iPST), and licensing fees for the use of the other apps (EVO and Ginger.io) was $314 264 over 2 years. Discussion The results from this study have a number of important implications for the future of RCTs in mental health. First, we recruited a large sample of depressed participants in a short period of time and with minimal cost and effort. Currently, the typical RCT takes 4–5 years to complete, and another 1–2 years before the outcomes from these trials are reported publically. Rapid recruitment has the potential for quickly testing intervention efficacy and effectiveness, and ultimately moving effective treatments into practice while identifying and preventing the proliferation of ineffective, even unsafe, treatments. Second, we were able to recruit a highly representative sample of the US population, without any specific cultural adaptations or targeted advertising. Remote research methods could address decades-long concerns about the generalisability of clinical findings to minority samples not typically represented in clinical research. Finally, the cost of a fully remote RCT could allow for greater distribution of dwindling clinical research and development funds from federal, foundation and industry sponsors. Investment in large-scale clinical trials is a costly endeavour, resulting in the need to focus funds on only a few research areas. Although not all mental health RCTs should be fully remote, particularly those that test hypotheses about biological or neurological processes that can only be measured with immobile devices, the methods presented here, such as automated data collection of neuropsychological processes could result in substantial savings, which in turn could be invested in a diverse research portfolio.

larly those that test hypotheses about biological or neurological processes that can only be measured with immobile devices, the methods presented here, such as automated data collection of neuropsychological processes could result in substantial savings, which in turn could be invested in a diverse research portfolio. This research method is not without its limitations. Primary among the challenges of fully remote research is the ability to keep participants engaged in the study protocol over time. Although there is an appeal to quickly recruiting and retaining large numbers of participants in an RCT, researchers and developers need to be cautious when interpreting outcomes from samples with a drop-out rate greater than 70%.30–33 However, it is important to point out here that the project was completely automated with very little contact between the participant and research team, and our retention rates were higher than in the typical internet-based RCT.34 Internet-based studies have shown that when there is more direct contact between the research team and participant through technologies such as video-over-internet protocols, retention rates are greater and less subject to bias,35 suggesting a hybrid approach may provide an optimal response.

pical internet-based RCT.34 Internet-based studies have shown that when there is more direct contact between the research team and participant through technologies such as video-over-internet protocols, retention rates are greater and less subject to bias,35 suggesting a hybrid approach may provide an optimal response. Although we experimented with two incentive models to encourage retention, we determined that participant payment was not enough to keep engagement from waning across the course of the study, as bonus pay only encouraged participants to complete their assessments, and did not engender any additional motivation to utilise the training apps. Previous work has demonstrated that externalised benefits in the form of compensation can dull motivation,36 37 indicating that the creation of internalised reward structure to enhance motivation (eg, individualised presentation of study progress, personalised encouragements) is critical for improved adherence. Conclusions Mobile technology has an important role in broadening the reach and representativeness of RCTs, while substantially reducing the time to determine intervention effectiveness and reducing study costs. Although study retention remains challenging for technology-based research, innovative methods to increase motivation and study engagement could easily address this important limitation.

d representativeness of RCTs, while substantially reducing the time to determine intervention effectiveness and reducing study costs. Although study retention remains challenging for technology-based research, innovative methods to increase motivation and study engagement could easily address this important limitation. Supplementary Material Web supplement The authors thank A Brandes-Aitken (UCSF), M Brander (UCSF) and A Bodepudi (UCSF) for assistance in data monitoring, M Gross (UCSF) and J Camire (Ekho.me) for their help in participant recruitment, J Steinmetz (Ekho.me) for database architecture, D Ziegler (UCSF) for helping with app deployment, D Albert (UCSF) for assistance in web design, L Slavik (UCSF) for accounting assistance, C Catledge (UCSF) for administrative oversight, A Piper, E Martucci, S Kellogg, J Bower, M Omerick and the entire Akili Interactive team as well as I Elson, L Kaye, S Goobich and the rest of the Ginger.io team for helping with data collection and partnering with us on this project. They also thank all the participants whose time and efforts made this work possible. Support for this research was provided by the National Institute of Mental Health (P.A.A.; R34-MH100466, T32 MH0182607, K24 MH074717). JAA and JTJ had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

efforts made this work possible. Support for this research was provided by the National Institute of Mental Health (P.A.A.; R34-MH100466, T32 MH0182607, K24 MH074717). JAA and JTJ had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Competing interests: AG is co-founder, chief science advisor and shareholder of Akili Interactive Labs, a company that develops cognitive training software. AG has a patent pending for a game-based cognitive training intervention, ‘Enhancing cognition in the presence of distraction and/or interruption’, on which the cognitive training application (PROJECT: EVO) that was used in this study was based. Ethics approval: UCSF Committee for Human Research. Provenance and peer review: Not commissioned; externally peer reviewed.

Venture capital and new medical technologies As industrialised countries develop strategies to expedite the commercial translation of biomedical discoveries1 and bring technological innovations closer to the clinic, policy initiatives that give a greater role to venture capital warrant careful examination.2 Such policies may include fiscal incentives to attract venture capital as well as the creation of funds dedicated to help health technology-based ventures grow.3–5 Drawing on the international innovation policy scholarship and a programme of research that examined the evolution of five Canadian health technology-based ventures over an 11-year period, this paper clarifies how venture capital operates and influences the kinds of technologies that make their way into healthcare systems. Our research entailed extensive document analyses (ie, annual reports, press releases, media coverage) and in-depth interviews with technology developers, capital investors, regulators and policymakers. This qualitative data set enabled our team to examine, from a health policy standpoint, the impact of business models, capital investment and economic policy on technology design processes.

ress releases, media coverage) and in-depth interviews with technology developers, capital investors, regulators and policymakers. This qualitative data set enabled our team to examine, from a health policy standpoint, the impact of business models, capital investment and economic policy on technology design processes. In this paper, our aim is to provide health services and policy researchers with the key lessons that pertain more specifically to the way current policy arrangements in systems of innovation support the development of technologies that capital investors identify as valuable, and which may not align with important health needs. While health services and policy scholars wish to encourage the creation of technologies that bring more value to healthcare, they rarely fully understand the mandate of venture capitalists and how they operate. A better understanding of how this form of financing ultimately affects healthcare systems would help health services and policy scholars play a more active role in innovation policy. Policy expectations towards venture capital In the past decades, North American and European countries actively sought to increase the size of their venture capital markets.6–10 The levels of venture capital available to Canadian life sciences companies have more than doubled from 2001 to 2010.11 In Europe, the UK enjoyed in 2009 the second largest venture capital market, accounting for 21% of all investments. In the same year, 20% of the UK £677 million of venture capital was invested in the health sector.4

enture capital available to Canadian life sciences companies have more than doubled from 2001 to 2010.11 In Europe, the UK enjoyed in 2009 the second largest venture capital market, accounting for 21% of all investments. In the same year, 20% of the UK £677 million of venture capital was invested in the health sector.4 Examining panel data of 17 European Union countries, Faria and Barbosa6 found that venture capitalists are “more willing to support innovation only after the initial and more uncertain stage of technology development has been overcome.” This tendency partly explains why the Horizon 2020 Programme protected up to €320.14 million in 2014 to help innovative firms to gain access to various types of risk financing.5 To foster the growth of technology-based ventures, venture capital typically steps in after entrepreneurs have fleshed out their core innovative idea with the financial support of governments, relatives and ‘angels’ (eg, wealthy individuals who finance entrepreneurial activities).3 As figure 1 indicates, venture capital provides both early-stage and late-stage financing, which precedes more substantive sources of capital such as public markets. Figure 1 Funding chain by stage of development and size of investment in $C (adapted from ref. 4).

Examining panel data of 17 European Union countries, Faria and Barbosa6 found that venture capitalists are “more willing to support innovation only after the initial and more uncertain stage of technology development has been overcome.” This tendency partly explains why the Horizon 2020 Programme protected up to €320.14 million in 2014 to help innovative firms to gain access to various types of risk financing.5 To foster the growth of technology-based ventures, venture capital typically steps in after entrepreneurs have fleshed out their core innovative idea with the financial support of governments, relatives and ‘angels’ (eg, wealthy individuals who finance entrepreneurial activities).3 As figure 1 indicates, venture capital provides both early-stage and late-stage financing, which precedes more substantive sources of capital such as public markets. Figure 1 Funding chain by stage of development and size of investment in $C (adapted from ref. 4). The innovation policy scholarship posits that venture capital-backed ventures are likely to outperform non-venture capital-backed ventures.6–8 The main arguments are that ‘investors can identify firms with hidden value’, that their investments operate as a ‘signal of the quality of the ventures for uninformed third parties’, and that they bring ‘external resources and competencies that would be out of reach’ without their endorsement.7 Informed by such literature, policymakers rely on ‘two pillars’ to foster the venture capital market. First, they seek to increase the ‘demand’ for venture capital by providing ‘generous subsidies’ and fiscal advantages to entrepreneurs so as to augment ‘birth rate’ of innovative firms.7 Second, they seek to increase the ‘supply’ of venture capital through ‘co-investment schemes, the launch of new government-owned venture capital companies and favourable tax treatment of capital gains’.7 Along those lines, the UK government created many funds since 2000, including the High Technology Fund, the Early Growth Funds and the Enterprise Capital Funds. Such funds played a particularly important role in early-stage financing in the following years, witnessing a peak in 2008 during which 68% of all early-stage venture capital investments were publicly backed.4

funds since 2000, including the High Technology Fund, the Early Growth Funds and the Enterprise Capital Funds. Such funds played a particularly important role in early-stage financing in the following years, witnessing a peak in 2008 during which 68% of all early-stage venture capital investments were publicly backed.4 How venture capital operates The two-pillar innovation strategy implies that public resources, either through taxation or ‘hands-off’ financing, are put to the service of venture capital, whose mandate largely conditions how capital-backed technology development processes unfold. Venture capitalists commit financial resources for a specified period of time to small privately held companies with few tangible assets and that rarely generate revenues alongside their initial research and development activities. What makes venture capital risky is the ‘illiquid’ nature of the investment during this period, meaning that the resources invested cannot be easily withdrawn.8 12 13 The window of opportunity for a ‘liquidity event’ such as acquisition by another company or an initial public offering—which provides the ability to sell shares to the public—is generally within 5 years.11 These so-called ‘exit’ events enable venture capitalists to recoup their investments and generate a return. Venture capital is very costly capital and the overarching principle is to generate the highest returns possible while knowing that most ventures fail.10 Song et al9 found that only 36% of American ventures created between 1991 and 2000 had survived after 4 years and this rate fell to 21.9% after 5 years. The returns from a subset of firms have to be much greater than average to make up for the expected failures, and yield above 20% returns for the investment portfolio as a whole.8 11 For instance, the top 25th centile of capital-backed UK firms generated returns ranging from 50% to 78% between 2003 and 2009, while the bottom 10 generated returns ranging from −14% to 0%.4

eater than average to make up for the expected failures, and yield above 20% returns for the investment portfolio as a whole.8 11 For instance, the top 25th centile of capital-backed UK firms generated returns ranging from 50% to 78% between 2003 and 2009, while the bottom 10 generated returns ranging from −14% to 0%.4 To successfully fulfil their mandate, venture capitalists generally seek to both pick winners and build them (see box 1).14 This means that they do not simply carefully choose entrepreneurs, but they also engage in ‘value-adding activities’.15 These activities include ‘coaching’ the ventures by providing the marketing and strategy support these young firms usually lack, professionalising their management by supporting the recruitment of seasoned managers, and facilitating alliances with key third parties within the industry. By having a seat at the Board of Directors of the fledgling firm, capital investors occupy an influential position from which they shape its governance (ie, advisory committees, nomination of high-level executives, partnerships) and seek to align technology developers towards their own vision.10 13 All of these value-adding activities are geared at augmenting the value of the investment, a process that entails shaping both the firm and the technology being developed.16 17 Box 1 What venture capitalists do to fulfil their mandate Picking likely winners and ‘de-risking’ deals at the outset Helping grow the ventures and intervening when progress is lacking

To successfully fulfil their mandate, venture capitalists generally seek to both pick winners and build them (see box 1).14 This means that they do not simply carefully choose entrepreneurs, but they also engage in ‘value-adding activities’.15 These activities include ‘coaching’ the ventures by providing the marketing and strategy support these young firms usually lack, professionalising their management by supporting the recruitment of seasoned managers, and facilitating alliances with key third parties within the industry. By having a seat at the Board of Directors of the fledgling firm, capital investors occupy an influential position from which they shape its governance (ie, advisory committees, nomination of high-level executives, partnerships) and seek to align technology developers towards their own vision.10 13 All of these value-adding activities are geared at augmenting the value of the investment, a process that entails shaping both the firm and the technology being developed.16 17 Box 1 What venture capitalists do to fulfil their mandate Picking likely winners and ‘de-risking’ deals at the outset Helping grow the ventures and intervening when progress is lacking Maintaining a dominant position until the liquidity event (eg, holding a seat at the Board of Directors of the ventures, enforcing a timeframe for instalments) Pushing capital-backed ventures to reach key milestones swiftly to avoid additional financing and the dilution of shares this implies Negotiating contracts and providing compensation for those who facilitate timely exit

Maintaining a dominant position until the liquidity event (eg, holding a seat at the Board of Directors of the ventures, enforcing a timeframe for instalments) Pushing capital-backed ventures to reach key milestones swiftly to avoid additional financing and the dilution of shares this implies Negotiating contracts and providing compensation for those who facilitate timely exit Venture capitalists exert control over technology design processes by setting the ‘term sheet’, which defines the milestones (eg, clinical trials, regulatory approval, sales) at which money is made available.13 This has direct implications for technology design priorities. Among the early-stage health technology companies we studied, for example, one clinically led firm had to modify the key goal of its labour monitoring system, which was to reduce unnecessary caesarean sections, and instead develop medicolegal functionalities for physician insurers who were more likely to purchase the system.18 This redesign of the system enabled their business to expedite sales and generate revenues.

trengths. We had a robust (55.1%) response rate for an email-solicited internet-based survey. This response rate itself suggests a high level of continuing engagement of hackathon participants. In addition, we dealt with potential response bias very conservatively by including non-responding teams as ‘non-progressors’. Our study also has several weaknesses. With self-reported surveys, it is possible that respondents over-reported markers of progress due to social desirability bias. However, the anonymous survey should minimise this bias. Another weakness is that the time frame of follow-up may have been short considering typical timelines of medical device development. The range of follow-up from our surveys was 1.9–40.0 months, whereas typical medical device development can range up to 10 years depending on clinical and regulatory requirements.16 In addition, projects involving drug, vaccine or entirely novel diagnostics were not pursued during these hackathons as they are considered less amenable to this format and have even longer development timelines. Finally, while it is impossible to determine how many innovations would have arisen and progressed without this hackathon model, participants reported a low likelihood that they would have proceeded without the hackathon.

ad to modify the key goal of its labour monitoring system, which was to reduce unnecessary caesarean sections, and instead develop medicolegal functionalities for physician insurers who were more likely to purchase the system.18 This redesign of the system enabled their business to expedite sales and generate revenues. As table 1 indicates, the level of congruence between a venture that seeks to bring to the clinic a new technology and the mandate of venture capital varies. The heart ablation catheter venture deployed an international cadre of clinical investigators that generated the evidence required for regulatory approval in different countries and, by offering to investors credible prospect of rapid and continued expansion, it was able to secure several rounds of capital investments. By developing a revenue-generating technology for medical specialists, this venture replicated a business model that is well established in the biomedical sector. Among the three examples, it is the home monitoring system venture where the level of congruence between the mandate of venture capital and the innovation was the lowest. Such a technology creates value for hospitals that do not generate revenues out of hospitalisations and have incentives to prevent deterioration of chronically ill patients, such as Health Maintenance Organizations(HMOs) in the USA or publicly funded integrated healthcare systems in Canada or the UK. Table 1 Contrasting examples of the level of congruence between the mandate of venture capital and health technology-based ventures (adapted from ref. 18)

As table 1 indicates, the level of congruence between a venture that seeks to bring to the clinic a new technology and the mandate of venture capital varies. The heart ablation catheter venture deployed an international cadre of clinical investigators that generated the evidence required for regulatory approval in different countries and, by offering to investors credible prospect of rapid and continued expansion, it was able to secure several rounds of capital investments. By developing a revenue-generating technology for medical specialists, this venture replicated a business model that is well established in the biomedical sector. Among the three examples, it is the home monitoring system venture where the level of congruence between the mandate of venture capital and the innovation was the lowest. Such a technology creates value for hospitals that do not generate revenues out of hospitalisations and have incentives to prevent deterioration of chronically ill patients, such as Health Maintenance Organizations(HMOs) in the USA or publicly funded integrated healthcare systems in Canada or the UK. Table 1 Contrasting examples of the level of congruence between the mandate of venture capital and health technology-based ventures (adapted from ref. 18) Heart ablation catheter Labour decision support software Home monitoring system Initial idea behind the technology, and clinical aims To develop a catheter that can identify cardiac cells causing arrhythmia and neutralise them by delivering extremely cold temperatures Cryotherapy could prove safer than the existing radiofrequency-based procedure.

ur decision support software Home monitoring system Initial idea behind the technology, and clinical aims To develop a catheter that can identify cardiac cells causing arrhythmia and neutralise them by delivering extremely cold temperatures Cryotherapy could prove safer than the existing radiofrequency-based procedure. To turn a predictive mathematical model into software that would improve obstetricians’ decisions during labour and delivery To predict labour and shoulder dystocia, reduce unnecessary Caesarean sections, and avoid birth-related injuries. To develop a computerised system to enable clinicians to support chronically ill patients from a distance To reduce unnecessary emergency visits and hospitalisations, and empower patients. Level of congruence, and key explanatory factors High An international cadre of investigators conduct clinical studies, refine early versions of the catheter, and contribute to academic marketing around the world. Market approval is first obtained in Europe to generate sales as well as clinical data required for market approval in the USA. The procedure generates revenues for physicians, and marketing channels already exist. Medium Obstetricians are not eager to use the system, and established clinical practices seem hard to change. Although market approval is obtained in the USA, the development of new marketing channels is costly. Physician insurers become the key target as purchasers (offering reduced premiums to obstetricians who accept to use the system). Low The system is developed and evaluated in collaboration with a regional hospital.

Medium Obstetricians are not eager to use the system, and established clinical practices seem hard to change. Although market approval is obtained in the USA, the development of new marketing channels is costly. Physician insurers become the key target as purchasers (offering reduced premiums to obstetricians who accept to use the system). Low The system is developed and evaluated in collaboration with a regional hospital. The co-design approach enables generating data on efficacy and responding to users’ requests, but it limits the growth of the venture. There is not a ‘single’ purchaser for a system that generates ‘distributed’ benefits (to patients, home care providers, hospitals). Among the ventures that are developing an innovation to support the provision of clinical services, those that are seen as more congruent with the goal of venture capital and less risky possess similar characteristics; their innovations address very large and reachable markets, enable physicians to generate revenues, and will ultimately be acquired by an established medical device manufacturer (for an exit to take place).18

ns as they are considered less amenable to this format and have even longer development timelines. Finally, while it is impossible to determine how many innovations would have arisen and progressed without this hackathon model, participants reported a low likelihood that they would have proceeded without the hackathon. Nevertheless, markers of progress that we identified are typical for traditional product development pathways—that is, fundraising, prototyping, patenting, trialling, company formation and licensing. In fact, with significant progress seen in <40 months—on average within 12 months—even greater progress towards commercialisation would be anticipated with longer periods of follow-up. For example, two projects that had made progress after the first hackathon have progressed to forming companies and raising significant capital (US$62.5 million and US$730 000), and one of these has already commercialised. Two other projects have been licensed to commercialisation partners. The CAMTech health hackathon model thus appears to have utility in compressing the innovation cycle.

hat are seen as more congruent with the goal of venture capital and less risky possess similar characteristics; their innovations address very large and reachable markets, enable physicians to generate revenues, and will ultimately be acquired by an established medical device manufacturer (for an exit to take place).18 The mandate of venture capitalists may, in principle, prove compatible with supporting ventures that address important health needs. But this is likely to happen by accident, not by design. Survey findings indicate that up to 85% of capital investors consider ‘not at all or somewhat important’ public health impact.19 Investors also contend that regulatory requirements decrease the ‘chance that an investment will be made in a ‘new area’’ and increase the chance an investment will be ‘made in a ‘me-too’ space’ (eg, where a slightly different technology is already implemented and marketed).19 To strike lucrative ‘homeruns’ within a predefined period of time, venture capitalists seek indeed to ‘de-risk’ the deal at the outset by enforcing stringent agreements.18 By and large, venture capitalists affect the kinds of technologies available to patients, clinicians and healthcare systems by investing in certain ventures and not in others, managing their growth and controlling the progression of their innovative products (see box 2). Box 2 How venture capital influences the technologies available to physicians, patients and healthcare systems Venture capitalists choose the sectors and innovations that are worth investing in, and increase chances of success

naging their growth and controlling the progression of their innovative products (see box 2). Box 2 How venture capital influences the technologies available to physicians, patients and healthcare systems Venture capitalists choose the sectors and innovations that are worth investing in, and increase chances of success Within the timeframe on which Returns on Investment (ROI) are estimated at the outset, the progression of capital-backed ventures is steered towards the most profitable exit Technology design priorities are influenced by the time and resource constraints that venture capitalists enforce The centrality of venture capital in innovation systems While venture capital is not designed to fulfil ‘society's most urgent public health priorities’,20 it occupies a central position in what innovation policy scholars define as ‘innovation systems’.21 Figure 2 illustrates the relationships between key milestones in the health technology development pathway and the institutions that define the ‘rules of the game’, that is, the specific requirements that players have to fulfil.22 The rules set by these institutions have both constraining and enabling effects. For instance, regulatory agencies exert control over the safety of medical devices, but by enabling market access, they also provide economic worth to technology-based ventures (as reflected in the value of their share).18 Figure 2 Institutions that enable and constrain the emergence, development and commercialisation of health technology.

The centrality of venture capital in innovation systems While venture capital is not designed to fulfil ‘society's most urgent public health priorities’,20 it occupies a central position in what innovation policy scholars define as ‘innovation systems’.21 Figure 2 illustrates the relationships between key milestones in the health technology development pathway and the institutions that define the ‘rules of the game’, that is, the specific requirements that players have to fulfil.22 The rules set by these institutions have both constraining and enabling effects. For instance, regulatory agencies exert control over the safety of medical devices, but by enabling market access, they also provide economic worth to technology-based ventures (as reflected in the value of their share).18 Figure 2 Institutions that enable and constrain the emergence, development and commercialisation of health technology. While institutional rules are often described as hindrances by investors and technology developers, such rules contribute to the stability and functioning of innovation systems: they provide incentives for innovation, supply information, reduce uncertainty, foster cooperation and make available mechanisms to handle conflicts.21 Without such rules, venture capitalists and technology developers would simply be unable to cooperate, trust each other, succeed in developing and commercialising a new medical technology, and persuade physicians and patients that their technologies are trustworthy. Overall, and as underscored by figure 2, it is innovation policy, venture capital, financial markets, and legal and corporate governance frameworks that deeply structure upstream innovation processes.16 17 22–24 Long before health policy comes into play, the way venture capital interlocks with these institutions has lasting downstream consequences for healthcare systems.

tion policy, venture capital, financial markets, and legal and corporate governance frameworks that deeply structure upstream innovation processes.16 17 22–24 Long before health policy comes into play, the way venture capital interlocks with these institutions has lasting downstream consequences for healthcare systems. One key lesson for clinical leaders and health services and policy scholars is that, despite the fact that public policies increasingly encourage venture capital-backed technological innovation in health, handling the subtleties associated with the fulfilment of valuable healthcare goals is neither part of venture capitalists’ mandate, nor of their worldview. Examining recent data from the USA, Fleming20 observed two types of shift in venture capital investments. There has been a shift away from life sciences to other sectors, such as early-stage internet and consumer-oriented start-ups, and a shift within the life sciences from early-stage to later-stage investments. The first shift is conditioned by the standpoint from which the value of the firms is assessed—that of an investment portfolio—which remains largely indifferent to the societal value of the innovation such firms may generate.24 The second shift underscores that the overarching goal of venture capital is not so much to foster the creation of innovation, but to extract economic value from innovative firms and technologies.2 6 16

stment portfolio—which remains largely indifferent to the societal value of the innovation such firms may generate.24 The second shift underscores that the overarching goal of venture capital is not so much to foster the creation of innovation, but to extract economic value from innovative firms and technologies.2 6 16 Bringing health policy considerations into innovation policy For Robinson,25 the current emphasis on more sophisticated and cost-conscious purchasing in healthcare may have the ‘potential to increase the social value of innovation’ by focusing technology developers on ‘the preferences and pocketbooks’ of their customers. Beyond their cost, we believe the value of innovations will increase only if clinical leaders and health services and policy experts contribute much more actively than they have done so far to innovation policy. Policy efforts in the past decades have been devoted to the consolidation of Health Technology Assessment (HTA) and evidence-based decision-making.26 This is not sufficient: institutional arrangements that currently prevail in systems of innovation push public policies to support the development of those technologies capital investors identify as valuable. Capital-backed technology development operates according to rules that are foreign to the fulfilment of healthcare needs, and the consideration of healthcare system-level challenges.27

in systems of innovation push public policies to support the development of those technologies capital investors identify as valuable. Capital-backed technology development operates according to rules that are foreign to the fulfilment of healthcare needs, and the consideration of healthcare system-level challenges.27 The policy assumption on which governments, capital investors and technology developers currently operate is that publicly funded research should translate into private entrepreneurial activities because technological innovation contributes to a nation's economic growth.1 2 28 This economic orientation in public policy establishes favourable conditions for venture capitalists to play an influential role in systems of innovation and shape key decisions about health needs.18 Nevertheless, if the key policy goal is to increase the relevance of innovations from a health policy standpoint, it is the knowledge and insights of health experts that should be foregrounded. Their expertise could shape the processes and criteria on which key bodies allocate important resources to health innovation, and where articulating more clearly a ‘demand’ for technologies that are valuable from a health system perspective matters (see box 3). Box 3 Decision-making bodies where the expertise of clinical leaders and health services and policy scholars could contribute to shape health innovation Research funding agencies where specific programmes dedicated to research and development in health are designed

from a health system perspective matters (see box 3). Box 3 Decision-making bodies where the expertise of clinical leaders and health services and policy scholars could contribute to shape health innovation Research funding agencies where specific programmes dedicated to research and development in health are designed Research centres of university teaching hospitals where the principles and goals of collaborative research between clinical innovators and industry are established Health technology priority setting and procurement committees where the value of new technologies is defined, sending ‘signals’ to entrepreneurs Currently, the financial speculative rules at play too easily reconcile policy goals that are distinct—health and wealth—‘without asking first what healthcare needs and challenges should be addressed’.29 Because certain types of technology are more likely than others to help tackle the ‘intractable problems’30 of healthcare systems, public policies should be equipped to promote those that address the needs of a growing elderly population, support patients who are afflicted by chronic diseases and reduce health disparities. Such knowledge can only come from clinical leaders and health services and policy scholars.

problems’30 of healthcare systems, public policies should be equipped to promote those that address the needs of a growing elderly population, support patients who are afflicted by chronic diseases and reduce health disparities. Such knowledge can only come from clinical leaders and health services and policy scholars. Contributors: PL drafted the preliminary versions of the paper. Both FAM and GD contributed substantially to its design, and to the analysis and interpretation of existing evidence. PL, FAM and GD were all engaged in a programme of research on the evolution of five Canadian technology-based ventures over an 11-year period. All authors approved the final version of the paper and are accountable for it. PL is the guarantor. Funding: This research was funded by an operating grant from the Canadian Institutes of Health Research (CIHR; #MOP-89 776). PL holds the Canada Research Chair on Health Innovations (2010–2015). Competing interests: None declared. Ethics approval: The Ethics Review Board of the University of Montreal approved the study. Provenance and peer review: Not commissioned; externally peer reviewed.

projects initiated at the hackathons had begun preclinical or clinical pilot testing. We believe these are key steps towards accelerating technical innovation for LMICs. In settings with limited resources or where financial expenditures are under increasing scrutiny, this model of innovation offers favourable returns. Medical technology development requires a specific focus on users who may be distinct from innovators and on unique considerations such as evidence of clinical effectiveness, regulatory approval and frequent ‘third-party payors’, for example, government health systems or insurances that pay on behalf of a beneficiary. A network or ‘ecosystem’ that includes skill sets necessary to commercialise high-value health products will be important to support impactful innovations. We documented several indicators of a strengthened health-focused innovation ecosystem emerging from the hackathons. Individual participants reported significantly increased confidence in their ability to address medical technology challenges in the future. Furthermore, many of these individuals would not have encountered each other without the hackathon structure. Notably, over 30% of teams reported having at least one team member who subsequently worked on other projects with people they had met at a hackathon. For example, one member on a team from the first hackathon responded that, as a mechanical engineer, he would never have met a paediatrician from sub-Saharan Africa and ultimately formed a new company together. This strengthened ecosystem helps create a structure that is better capable of solving healthcare challenges beyond the scope of hackathons.

am from the first hackathon responded that, as a mechanical engineer, he would never have met a paediatrician from sub-Saharan Africa and ultimately formed a new company together. This strengthened ecosystem helps create a structure that is better capable of solving healthcare challenges beyond the scope of hackathons. Our study has several strengths. We had a robust (55.1%) response rate for an email-solicited internet-based survey. This response rate itself suggests a high level of continuing engagement of hackathon participants. In addition, we dealt with potential response bias very conservatively by including non-responding teams as ‘non-progressors’.

Introduction Low-income and middle-income countries (LMICs) bear nearly 90% of the global burden of disease1 2 and yet have severe shortages of trained healthcare providers. Innovative medical technologies are included in one of six essential building blocks of the WHO Health System Framework3 4 and can have particular impact in LMICs. Health technology developments can be transformative.5 They can offset limitations of healthcare workforces by optimising the efficiency and effectiveness of providers. Yet, medical technology innovation largely occurs in high-income countries (HICs) and an estimated 40–70% of resulting technologies fail when implemented in LMICs.6–8 Thus, there may be significant benefit in joining efforts by innovators from LMICs and HICs to ensure relevance, feasibility and acceptability of new technologies. This could help maximise relative ‘value’ (defined as health outcomes over cost).9

mated 40–70% of resulting technologies fail when implemented in LMICs.6–8 Thus, there may be significant benefit in joining efforts by innovators from LMICs and HICs to ensure relevance, feasibility and acceptability of new technologies. This could help maximise relative ‘value’ (defined as health outcomes over cost).9 The term ‘hackathon’ combines ‘hack’—a solution reached through intense innovation—and ‘marathon’—an event of defined length and concentrated effort. Healthcare hackathons champion the process of ‘co-creation’, in which serendipitous meetings of people across geographies and disciplines such as healthcare, design, engineering and business enable diverse teams to develop potential solutions in a time-limited format. Hackathons are often incorrectly described as ‘crowdsourcing’ occurrences. First appearing in 2006, ‘crowdsourcing’ is a portmanteau of ‘crowd’ and ‘outsourcing’ to indicate a process of obtaining ideas or services from a large—usually online—community.10 In contrast, healthcare hackathons are 48-hour events in which a group of curated individuals from different backgrounds come together to drive innovation in healthcare.11 Hence, we suggest that hackathons instead represent ‘enriched crowdsourcing’.

f obtaining ideas or services from a large—usually online—community.10 In contrast, healthcare hackathons are 48-hour events in which a group of curated individuals from different backgrounds come together to drive innovation in healthcare.11 Hence, we suggest that hackathons instead represent ‘enriched crowdsourcing’. The Massachusetts Institute of Technology's (MIT) Hacking Medicine, a frequent partner of the Consortium for Affordable Medical Technologies (CAMTech), first pioneered health hackathons in 2011. Software hackathons have been conducted for decades, and few existing empirical reports indicate that up to 36% of teams formed at hackathon-style events continued to work together after 3 months.12 However, these reports do not identify additional markers of progress. In addition, unlike medical technologies, software development has fewer constraints and participants are often themselves the intended end users. In addition, there is active debate in the press and academic literature about the utility of healthcare hackathons. On the one hand, it has been argued that healthcare hackathons recruit and empower a new mix of innovators to work collectively towards novel solutions not possible within traditional academic or commercial environments.13 However, some suggest that although the hype of these events generates enthusiasm, healthcare hackathons fall short of developing lasting innovations because healthcare challenges are too complex to be addressed in a 48-hour event.14 15 Others state that hackathons actually may deter relevant innovation by creating an ‘artificial vacuum’ in which solutions are proposed.13 Ultimately, the question is whether hackathons are merely theatre—a show without tangible outcomes—or do they represent events with substantive results?

ssed in a 48-hour event.14 15 Others state that hackathons actually may deter relevant innovation by creating an ‘artificial vacuum’ in which solutions are proposed.13 Ultimately, the question is whether hackathons are merely theatre—a show without tangible outcomes—or do they represent events with substantive results? The goal of the CAMTech is to accelerate sustainable and effective medical technology innovation and enhance entrepreneurial capacity to improve health in resource-constrained settings globally. To realise the full potential of healthcare hackathons, CAMTech has developed a model in which (1) preceding priming activities focus on subsequent work, (2) 48-hour hackathons catalyse innovations and (3) a suite of posthackathon offerings stimulates the progress of innovations towards commercialisation. We evaluated outputs of events in India, Uganda and the USA to determine the impact of this hackathon model. Primary outcomes included prototypes pilot tested, funds raised, business plans created, patents filed and companies formed.

thackathon offerings stimulates the progress of innovations towards commercialisation. We evaluated outputs of events in India, Uganda and the USA to determine the impact of this hackathon model. Primary outcomes included prototypes pilot tested, funds raised, business plans created, patents filed and companies formed. Methods Hackathons From October 2012 to October 2015, CAMTech hosted 12 hackathons. Five were held in India, 3 in Uganda and 4 in the USA (3 of which were organised in collaboration with MIT's Hacking Medicine). Hackathons were advertised through academic, industry and non-governmental organisation email lists, flyers, as well as website and social media postings. All were focused on health in LMICs and other low-resource settings. Several hackathons had specific themes including reproductive, maternal, newborn and child health (RMNCH); Ebola care; diabetes; and trauma/road safety. Each hackathon was preceded by activities to maximise participant readiness including identifying healthcare challenges by engaging hackathon participants with multisector stakeholders online and at clinical summits; visiting venues coping with healthcare challenges; and engaging mentors from clinical, business, technology and design disciplines to participate (table 1). Table 1 CAMTech healthcare hackathon model Component Timeline Prehackathon 1. Advertise hackathon and solicit mentors with domain expertise Event website, social media, flyers, emails and word of mouth 1–3 months in advance 2. Curate both applicants and mentors Patients (when possible) and clinicians Engineers and designers

Methods Hackathons From October 2012 to October 2015, CAMTech hosted 12 hackathons. Five were held in India, 3 in Uganda and 4 in the USA (3 of which were organised in collaboration with MIT's Hacking Medicine). Hackathons were advertised through academic, industry and non-governmental organisation email lists, flyers, as well as website and social media postings. All were focused on health in LMICs and other low-resource settings. Several hackathons had specific themes including reproductive, maternal, newborn and child health (RMNCH); Ebola care; diabetes; and trauma/road safety. Each hackathon was preceded by activities to maximise participant readiness including identifying healthcare challenges by engaging hackathon participants with multisector stakeholders online and at clinical summits; visiting venues coping with healthcare challenges; and engaging mentors from clinical, business, technology and design disciplines to participate (table 1). Table 1 CAMTech healthcare hackathon model Component Timeline Prehackathon 1. Advertise hackathon and solicit mentors with domain expertise Event website, social media, flyers, emails and word of mouth 1–3 months in advance 2. Curate both applicants and mentors Patients (when possible) and clinicians Engineers and designers Business personnel 1 month in advance 3. Some problems (ie, pain points) posted on an event website Clinical/public health challenges solicited from experts and partner organisations 1 month in advance 4. Clinical site visits Participants tour partner healthcare facilities and hold discussions with providers

Engineers and designers Business personnel 1 month in advance 3. Some problems (ie, pain points) posted on an event website Clinical/public health challenges solicited from experts and partner organisations 1 month in advance 4. Clinical site visits Participants tour partner healthcare facilities and hold discussions with providers 1–2 days in advance 5. Clinical summits Clinical, public health and policy experts discuss challenges 1 day in advance 6. Social mixer Participants socialise to initiate cross-disciplinary relationships Evening before a hackathon Hackathon 48 hours Day 1 Introductory speeches highlighting challenges and incentive prizes; introduction to ‘hacking’ One-minute problem verbalisations (pitching pain points) Team formation and registration Hacking and mentoring AM PM Day 2 Hacking and mentoring Practise concept pitches Three-minute team concept pitches ‘After the hack’ opportunities and suggestions Hackathon awards AM PM Posthackathon 1. Innovation mentorship Tailored individual project mentorship Co-creation laboratory space for prototyping and project maturation Connection with targeted organisations/individuals having stage-specific skills needed by teams Networking, mentorship and dissemination of opportunities and information through resources such as the online CAMTech Innovation Platform Anytime Monetary awards to the most promising innovations Grant writing mentorship and workshops Dissemination/notification of additional funding sources 30–90 days posthackathon and beyond 3. Entrepreneurial expertise Facilitation of teams through existing Business Accelerators*

Networking, mentorship and dissemination of opportunities and information through resources such as the online CAMTech Innovation Platform Anytime Monetary awards to the most promising innovations Grant writing mentorship and workshops Dissemination/notification of additional funding sources 30–90 days posthackathon and beyond 3. Entrepreneurial expertise Facilitation of teams through existing Business Accelerators* Entrepreneurial group learning through 2-day Entrepreneur Bootcamps (these consist of in-person, intensive workshops with guest experts designed to enhance participant entrepreneurial skill in the medical technology space) Facilitation and funding through the CAMTech Business Accelerator Program Variable *Business Accelerator: a business programme that includes mentorship, educational components and networking that aims to grow a business rapidly. The CAMTech Accelerator Program includes the resources available on the CAMTech Innovation Platform in addition to more in-depth team coaching at frequent intervals. CAMTech, Consortium for Affordable Medical Technologies.

Variable *Business Accelerator: a business programme that includes mentorship, educational components and networking that aims to grow a business rapidly. The CAMTech Accelerator Program includes the resources available on the CAMTech Innovation Platform in addition to more in-depth team coaching at frequent intervals. CAMTech, Consortium for Affordable Medical Technologies. Individuals apply to participate with a statement of their experience and motivations for attending. These applicants are curated by organisers to permit a pool of participants with a broad mix of skills who describe compelling motivations. The 48-hour hackathons focus on framing health challenges, verbalising specific problems (pitching pain points), team formation, ideation and prototyping through hacking, and the competitive presentation of resulting concepts to judges. In the months following hackathons, CAMTech offered support to teams by facilitating clinical, technological and entrepreneurial mentorship; providing laboratory workspace and connections to production resources; enabling seed grants for promising projects; and building capacity for external fundraising (including grant writing) and engagement of investors—all with the aim of accelerating promising innovations towards commercialisation. Although prehackathon activities were in place, several posthackathon activities have become formally operationalised and/or have come ‘online’ recently. For example, the online CAMTech Innovation Platform came online on 12 June 2015.

ors—all with the aim of accelerating promising innovations towards commercialisation. Although prehackathon activities were in place, several posthackathon activities have become formally operationalised and/or have come ‘online’ recently. For example, the online CAMTech Innovation Platform came online on 12 June 2015. To determine the impact of the CAMTech healthcare hackathon model, we conducted two surveys. The first was conducted in September to October 2014 to measure outcomes of the first six hackathons. The second survey was conducted in December 2015 to January 2016 to measure outcomes from all 12 hackathons. Recruitment and participants At least one team member's email address was collected at each hackathon during team registration. A survey request consisting of an introductory email explaining the purpose and providing a link to the online survey was sent to registered email addresses for each team and to additional team members for which an email address was recorded. Two reminder emails were sent after each survey.

on during team registration. A survey request consisting of an introductory email explaining the purpose and providing a link to the online survey was sent to registered email addresses for each team and to additional team members for which an email address was recorded. Two reminder emails were sent after each survey. To be eligible, respondents had to be at least 18 years old. Since all hackathons occurred in Uganda, India or the USA, where English is the primary language, surveys were administered in English. An incentive for survey completion was offered as entry into a raffle for either a US$100 Amazon gift card or prepaid mobile phone minutes. Each survey had separate drawings for residents of India, Uganda and HICs. If respondents opted to enter these raffles, they were directed to a separate site to enter their personal information in order to maintain confidentiality.

ntry into a raffle for either a US$100 Amazon gift card or prepaid mobile phone minutes. Each survey had separate drawings for residents of India, Uganda and HICs. If respondents opted to enter these raffles, they were directed to a separate site to enter their personal information in order to maintain confidentiality. The survey The 15–20 min survey included three sections: (1) evaluation of hackathon experiences; (2) information on specific hackathon projects and (3) demographic information. Responses were confidential and reported anonymously by project code. Depending on response-driven skip patterns, 41–53 responses were possible. A respondent could provide data for more than one hackathon project if s/he was on more than one team (at the same or different hackathons). However, responses measuring a personal hackathon experience were collected only once from the initial survey response using a survey function that skipped to project-specific questions if the participant reported having responded to a survey for another project. Survey questions relating to primary outcomes are presented in online supplementary table S1. 10.1136/bmjinnov-2016-000147.supp1supplementary table Survey Questions pertaining to primary outcomes.

The survey The 15–20 min survey included three sections: (1) evaluation of hackathon experiences; (2) information on specific hackathon projects and (3) demographic information. Responses were confidential and reported anonymously by project code. Depending on response-driven skip patterns, 41–53 responses were possible. A respondent could provide data for more than one hackathon project if s/he was on more than one team (at the same or different hackathons). However, responses measuring a personal hackathon experience were collected only once from the initial survey response using a survey function that skipped to project-specific questions if the participant reported having responded to a survey for another project. Survey questions relating to primary outcomes are presented in online supplementary table S1. 10.1136/bmjinnov-2016-000147.supp1supplementary table Survey Questions pertaining to primary outcomes. Data analysis The duration of follow-up was calculated using an elapsed-time calculator from the date of each hackathon to the time-stamped response from the last team member to reply. Project-specific codes prevented double counting of project-level data. Answers for project-specific codes were counted as positive if progress afterwards, business plans, trials, patents filed, team members obtaining a job/position, or companies formed were reported by at least one respondent. If discrepant answers were received from the two surveys, answers from the later survey were used. If, within the same survey, discrepant quantitative answers were received from multiple respondents for the same project, the averages of responses were used. Such items included number of people per team, total monetary investments/awards and agreement with the statement: ‘I would have made similar progress on my project WITHOUT the hackathon’ (with answers scaled 1 ‘wouldn't have made any’ to 5 ‘would have progressed just as far’). Currencies were converted to US$ on 10 February 2016 using a standard currency converter (http://www.oanda.com/currency.converter).

ment with the statement: ‘I would have made similar progress on my project WITHOUT the hackathon’ (with answers scaled 1 ‘wouldn't have made any’ to 5 ‘would have progressed just as far’). Currencies were converted to US$ on 10 February 2016 using a standard currency converter (http://www.oanda.com/currency.converter). To correct for possible response bias, teams for which no response was received were considered to have made no progress after the hackathon. Data were analysed via SPSS V.23 (Chicago, Illinois, USA). Statistical tests included Cramer's V test, one-way analysis of variance and paired samples t-test. Statistical significance was set at p<0.05. This study protocol was reviewed by the Partners Healthcare Institutional Review Board and determined to be ‘exempt’. Results Of the 363 teams participating in the 12 CAMTech health hackathons, 356 (98.1%) had at least one functioning email address (ie, for which no ‘undeliverable’ message was received), and 196 project teams (55.1%) provided at least one survey response. From 1367 email addresses to which the survey was sent, 331 responses (24.2%) were received. There was no statistically significant difference in response rates of teams from hackathons held in India, Uganda or the USA (Cramer's V=0.09; p=0.17). Of the 331 responses, 48 were from individuals who worked on more than one CAMTech hackathon project. Hence, there were 283 responses to individual experience queries.

There was no statistically significant difference in response rates of teams from hackathons held in India, Uganda or the USA (Cramer's V=0.09; p=0.17). Of the 331 responses, 48 were from individuals who worked on more than one CAMTech hackathon project. Hence, there were 283 responses to individual experience queries. The average project team size was 4.8 people (median 5), and over 60% of teams had three or more professional disciplines represented. Respondents were predominantly young (18–25 years of age) and slightly over three-quarters were male (table 2). Table 2 CAMTech hackathon project teams Host country Projects surveyed Responses (%) India (N=5*) 192 99 (51.6) Uganda (N=3) 96 53 (55.2) USA (N=4) 68 44 (64.7) Total (N=12) 356 196 (55.1) Hackathon projects Mean Range Projects surveyed/hackathon 29.7 11–63 Project team size (people) 4.8 1–12 Disciplines/project** Per cent 1 20.0 2 18.5 ≥3 61.0 Respondents N Per cent Gender Male 256 77 Female 75 23 Age 18–25 207 62.5 26–33 75 22.7 34–40 22  6.6 41–55 19  5.7 55–65 6  1.8 ≥65 2  0.6 *N=number of hackathons. **Disciplines identified in survey: (1) medicine; (2) engineering; (3) business; (4) design; (5) life sciences and (6) other. CAMTech, Consortium for Affordable Medical Technologies. Public health and healthcare challenges identified in prehackathon activities and addressed by hackathon teams were diverse and ranged from infectious to non-communicable/chronic diseases to RMNCH (figure 1). Figure 1 Health topics addressed.

**Disciplines identified in survey: (1) medicine; (2) engineering; (3) business; (4) design; (5) life sciences and (6) other. CAMTech, Consortium for Affordable Medical Technologies. Public health and healthcare challenges identified in prehackathon activities and addressed by hackathon teams were diverse and ranged from infectious to non-communicable/chronic diseases to RMNCH (figure 1). Figure 1 Health topics addressed. Teams conceived of predominantly preventive, diagnostic or therapeutic medical devices. In total, 61.2% of projects included hardware in their innovations, with 39.8% having both hardware and software. A significant minority of projects (19.9%) included ‘process’ innovations or non-technical solutions. A similar number of projects, almost 18.9%, included software only. Over 30% of teams reported continued work over a mean of 12.2 months after hackathons, and another 10% planned to engage in future work on their projects (table 3). One-quarter (N=88, 24.7%) of projects surveyed had begun piloting including in preclinical settings (N=52, 14.6%), on healthcare workers (N=42, 11.8%) and/or involving patients (N=24; 6.7%). Notably, 22 patents were filed for projects emerging from the 12 hackathons, and 15 new companies were formed. Table 3 Posthackathon team-level responses

Over 30% of teams reported continued work over a mean of 12.2 months after hackathons, and another 10% planned to engage in future work on their projects (table 3). One-quarter (N=88, 24.7%) of projects surveyed had begun piloting including in preclinical settings (N=52, 14.6%), on healthcare workers (N=42, 11.8%) and/or involving patients (N=24; 6.7%). Notably, 22 patents were filed for projects emerging from the 12 hackathons, and 15 new companies were formed. Table 3 Posthackathon team-level responses N Per cent Mean/hackathon Continued work beyond a hackathon Yes 108 30.3 9.0 Not yet but plan to 36 10.1 3.0 Total 144 40.4 12.0 Prototype progress after a hackathon Teams reporting any progress 126 35.4 10.5 Teams reporting major progress 70 19.7 5.8 Projects initiating pilot work Any 88 24.7 7.3 Preclinical studies 52 14.6 4.3 Clinical studies with providers or healthcare workers only 42 11.8 3.5 Clinical studies with patients 24 6.7 2.0 Teams initiating a business plan 97 27.2 8.1 Patents filed 22 6.2 1.8 Companies formed 15 4.2 1.25 Financial support raised Number of awards Number of teams obtaining any award Awards 58 31 Total raised (millions of US$) Mean per hackathon (millions of US$) Median award size (US$) Amount in US$ 64.082 5.34 1800 Perceived progress without a hackathon (1=would have made no progress; 5=would have made as much if not more progress) Scale average SD All teams 1.92 0.90 Teams raising funds 1.87 0.79 Teams forming companies 2.01 0.90 Teams surveyed: 356; mean follow-up period: 12.2 months (range 1.9–40.0).

nt in US$ 64.082 5.34 1800 Perceived progress without a hackathon (1=would have made no progress; 5=would have made as much if not more progress) Scale average SD All teams 1.92 0.90 Teams raising funds 1.87 0.79 Teams forming companies 2.01 0.90 Teams surveyed: 356; mean follow-up period: 12.2 months (range 1.9–40.0). One team that founded a company raised $62.5 million for its project in 40 months following a hackathon. A member from this team reported, “The hackathon catalyzed our work…I don't think the company would have come together had we not participated in the hackathon” and also that “we now have more than 200 team members, offices and pharmacies around the country”. In total, $64.08 million was raised in funding through investment capital, awards or grants with a mean of $5.34 million per hackathon and a median of $1800 per team that raised funds. The distribution of award sizes is shown in figure 2. Even removing the largest company as an outlier, 30 teams raised funding for their continued work for a mean of $131 803 raised per hackathon. Figure 2 Distribution of funds raised. When asked about agreement with the statement, ‘I would have made similar progress on my project WITHOUT the hackathon?’ with 1=‘I wouldn't have made any progress’ to 5=‘I would have progressed just as far if not further’, the average response of team members was 1.92. The average response of teams that had received monetary awards was 1.87 and was 2.01 among teams that formed new companies with no significant differences between these three values (p=0.65).

ve made any progress’ to 5=‘I would have progressed just as far if not further’, the average response of team members was 1.92. The average response of teams that had received monetary awards was 1.87 and was 2.01 among teams that formed new companies with no significant differences between these three values (p=0.65). In addition to markers of team progress, several indicators of a strengthened innovation ecosystem emerged. Of responding teams, 45 (3.75 per hackathon) reported that at least one member had obtained a job or position due to connections made at the hackathon. One hundred and eight teams (30.3%) reported at least one member working on additional technologies with people they had met at a hackathon. Finally, responses from participants showed significantly increased confidence to address medical technology challenges in healthcare after attending a hackathon (t (282)=13.88, p<0.001; table 4). Table 4 Hackathon attitudes and impact (N=283)

In addition to markers of team progress, several indicators of a strengthened innovation ecosystem emerged. Of responding teams, 45 (3.75 per hackathon) reported that at least one member had obtained a job or position due to connections made at the hackathon. One hundred and eight teams (30.3%) reported at least one member working on additional technologies with people they had met at a hackathon. Finally, responses from participants showed significantly increased confidence to address medical technology challenges in healthcare after attending a hackathon (t (282)=13.88, p<0.001; table 4). Table 4 Hackathon attitudes and impact (N=283) N Per cent Mean N/ Hackathon Would attend again Yes 247 87.3 No 5 1.8 Unsure 31 11.0 Mean score (SD) Would recommend to a friend (1=not at all likely to 5=extremely likely) 4.7 (0.62) Mean score before (SD) Mean score after (SD) p Value Confidence to address medical technology challenges (1=not at all confident to 5=very confident) 3.25 (1.21) 4.25 (0.84) <0.001 Mean score (SD) Attending hackathon increased confidence to begin a new venture (1=not at all confident to 5=very confident) 4.22 (0.94) Discussion Despite debate on the impact and utility of healthcare hackathons, we are not aware of previously published quantitative evaluations of their outcomes. We found that 30% of teams continued work after hackathons and yielded 1.8 new patents, 1.25 new companies and $5.34 million of follow-on funding per hackathon with a mean follow-up of 12 months. Remarkably, 25% of all projects initiated at the hackathons had begun preclinical or clinical pilot testing. We believe these are key steps towards accelerating technical innovation for LMICs. In settings with limited resources or where financial expenditures are under increasing scrutiny, this model of innovation offers favourable returns.

on have progressed to forming companies and raising significant capital (US$62.5 million and US$730 000), and one of these has already commercialised. Two other projects have been licensed to commercialisation partners. The CAMTech health hackathon model thus appears to have utility in compressing the innovation cycle. Sastry and Penn13 argue that anticipating outcomes from a hackathon itself is futile. It does indeed seem that structures need to be in place to facilitate progress of good ideas beyond the typical 48 hours of hackathons. CAMTech increasingly has focused its consortium towards both priming before and nurturing nascent innovations after hackathons. The preceding activities are designed to maximise the ability of participants to focus their hackathon projects to address real-world challenges; these are followed by support to stimulate progress after hackathons. For example, CAMTech developed an innovation award programme for early stage ideas, an online community through the CAMTech Innovation Platform (http://camtechmgh.org), entrepreneur development opportunities through in-person and virtual accelerators such as the CAMTech Accelerator Program (CAP), and physical workspaces such as at the CAMTech Co-Creation Laboratory in Uganda. In response to participants' stated needs, CAMTech initiated ‘Entrepreneur Bootcamps’ to increase innovators' entrepreneurial skills and has offered grant writing workshops in India and Uganda. It may be argued that these supportive initiatives helped create the remarkable progress of the healthcare hackathons on which we report. In response to the increased numbers and stated needs of innovators, the postevent suite of offerings has been expanded over the follow-up period. It is likely that these will continue to expand over time. To aid with dissemination of this model, CAMTech has developed a standard operating procedure guide and a lessons-learnt ‘living’ document to better enable organisations to develop their own healthcare hackathons.

gs has been expanded over the follow-up period. It is likely that these will continue to expand over time. To aid with dissemination of this model, CAMTech has developed a standard operating procedure guide and a lessons-learnt ‘living’ document to better enable organisations to develop their own healthcare hackathons. These ‘enriched crowdsourcing’ events help to ensure an optimal mix of experience directed towards focused challenges. Whether similar results can be seen through replication remains to be seen. However, these 12 hackathons indicate a remarkable source of med-tech innovations. This study highlights several areas for future research to determine the impact of healthcare hackathons. Investigation into specific characteristics predictive of the greatest progress is needed. In addition, investigation into which postevent offerings are principal contributors to success would be helpful as the model matures. Further research is also warranted to determine the number and, ultimately, the value of commercialised projects over a longer time frame.

predictive of the greatest progress is needed. In addition, investigation into which postevent offerings are principal contributors to success would be helpful as the model matures. Further research is also warranted to determine the number and, ultimately, the value of commercialised projects over a longer time frame. Conclusion Healthcare is in need of value-based solutions to improve outcomes and curtail costs, particularly in LMICs. There has been debate whether healthcare hackathons represent merely an enthusiastic hype of creative participants or substantive sources of healthcare innovations. Data presented indicate that, within three different global locations, a healthcare hackathon model including preceding priming activities and targeted postevent support is a reliable source of solutions to healthcare challenges. To a great extent, these events help develop a healthcare solution ecosystem primed to solve as yet unaddressed challenges. Twitter: Follow Kristian Olson @olsonkristian1

Conclusion Healthcare is in need of value-based solutions to improve outcomes and curtail costs, particularly in LMICs. There has been debate whether healthcare hackathons represent merely an enthusiastic hype of creative participants or substantive sources of healthcare innovations. Data presented indicate that, within three different global locations, a healthcare hackathon model including preceding priming activities and targeted postevent support is a reliable source of solutions to healthcare challenges. To a great extent, these events help develop a healthcare solution ecosystem primed to solve as yet unaddressed challenges. Twitter: Follow Kristian Olson @olsonkristian1 Contributors: KRO conceptualised the study design, execution, and provided the first draft of the article, subsequent revisions, and submitted the manuscript. MW contributed to data collection, initial analysis, and research of background as well as initial drafts. PG contributed to the design and initial survey draft, and performed data collection for the first survey time period as well as draft revisions. AS, SM, SD and EB all contributed to conceptualisation of the study design and to editing of the manuscript drafts. RP and AJG contributed to the formatting of the arguments, statistical analysis and draft editing. DRB helped to conceptualise the investigation and provided guidance on the analysis and structure of the manuscript. All authors contributed to the editing of the final drafts of the manuscript and all provided approval of the final submission.

to the formatting of the arguments, statistical analysis and draft editing. DRB helped to conceptualise the investigation and provided guidance on the analysis and structure of the manuscript. All authors contributed to the editing of the final drafts of the manuscript and all provided approval of the final submission. Funding: This study was graciously funded by the Bacca Foundation (Grant Number 1200-020972). Competing interests: None declared. Ethics approval: Partners Healthcare Institutional Review Board. Provenance and peer review: Not commissioned; externally peer reviewed. Data sharing statement: KRO, MW, RP and AJG are study staff who continue to have access to the entire data set. All have completed Human Subjects Research training. These data are available only via password-protected web portals to the primary survey responses.

Introduction The development of new medicines is undertaken to improve the health and outcomes of patients. However, until recently, patient involvement in biopharmaceutical development has often been infrequent, episodic and restricted to the periphery apart from direct participation in clinical trials and postapproval activities such as disease education. But why is this? Biopharmaceutical innovation and development is complex and highly regulated with well-defined but proscriptive technical requirements for the conduct of clinical research.1 2 As would be expected, much focus has been placed on ensuring patient safety,2 appropriate measurement of clinical outcomes,2 and more recently in establishing patient confidentiality.3 All of these aspects created a regulatory-focused process,2 orientated around healthcare providers,4 efficacy and safety of products and submission of data.2 Patient participation was always recognised but in a paternalistic way, with patients being instructed and not overburdened with information.4 It is easy to understand therefore that a culture has historically existed where research has been conducted ‘on’, ‘about’ or ‘for’ patients.5 Indeed, a clear disconnect has existed between the biopharmaceutical industry and the ultimate end user, the ‘Patient’,6 with patient engagement often limited to clinical trial participation or education about a new drug and frequently using healthcare professionals as a proxy.7 Therein lies a problem.

‘for’ patients.5 Indeed, a clear disconnect has existed between the biopharmaceutical industry and the ultimate end user, the ‘Patient’,6 with patient engagement often limited to clinical trial participation or education about a new drug and frequently using healthcare professionals as a proxy.7 Therein lies a problem. Health outcomes are dependent on patient engagement,8 but the full benefits of pharmacotherapy cannot be realised when, for example, ∼50% of patients do not take their medication as prescribed in the long term.9 The causes of non-compliance are myriad and include patient factors (eg, forgetfulness, cost, inability to obtain medication), physician factors (eg, poor doctor–patient relationship, communication issues) and medication factors (eg, adverse events, complicated regimens).10 Furthermore, drug products may be developed that patients do not want and will not use.7 Clearly, development of medicines using a patient-centric approach provides an opportunity to more closely meet patient needs and therefore improve their lives in ways that patients and their families view as meaningful. In this respect, improving the lives of patients and their loved ones requires a deep understanding of their medical conditions, real-world experiences, needs and priorities.8 Therefore, it is now essential that the biopharmaceutical industry partner with patients in the discovery, research, development, distribution and access to medicines to bring about outcomes that are meaningful to them. Working with patients fosters innovation and it will ensure that the objectives of patients are met early on in the biopharmaceutical development process by incorporating their views and needs, leading to more impactful patient outcomes.8 Patients are the ultimate end users of medicines and it is they and their carers that should prioritise their needs and identify the outcomes that they desire. In this respect, patients should be given the opportunity to define their needs,8 including the value of interventions, the benefit and risk trade-offs based on their values, their desired clinical outcomes, preferences and experiences.

rers that should prioritise their needs and identify the outcomes that they desire. In this respect, patients should be given the opportunity to define their needs,8 including the value of interventions, the benefit and risk trade-offs based on their values, their desired clinical outcomes, preferences and experiences. This concept of ‘patient centricity’11 and the opportunity to include patient-centred activities in drug development has been a key topic of interest in the past few years,6 12–18 with parallel advances made with key stakeholders including industry, regulators and healthcare professionals.19–21 Patients are also demanding that drug development becomes more patient centric and that more patient-centric relationships are developed.22 23 Indeed, this is a much discussed and written about topic, but as yet, a consistent definition of patient centricity as applied to drug development has not been published, and although definitions relating to other key stakeholders (eg, healthcare providers24) have been developed, these are more specific to the delivery of healthcare by healthcare professionals and not necessarily transferable to the drug development process. Indeed, patients to date have not been asked to define how they want to engage in the scientific research and development for their medicines of the future. Research has also shown that patient centricity is a highly recognised term among life science executives in the biopharmaceutical industry, but a consistent definition is lacking.25

date have not been asked to define how they want to engage in the scientific research and development for their medicines of the future. Research has also shown that patient centricity is a highly recognised term among life science executives in the biopharmaceutical industry, but a consistent definition is lacking.25 Implementation of patient centricity within biopharmaceutical industry development programmes requires a consistent and coordinated approach,20 with a common definition and associated principles to ensure medicines deliver the optimal experience and outcome for patients and their loved ones. In this respect, it is important to ensure that the guiding principles of patient centricity are defined by patients for patients so that these can be used within industry as a point of reference and promote stronger unified commitment to patient engagement. Here, we describe the process leading to a consensus definition of patient centricity and its important principles impacting the biopharmaceutical industry as relates to the biopharmaceutical industry and the drug development process.

s a point of reference and promote stronger unified commitment to patient engagement. Here, we describe the process leading to a consensus definition of patient centricity and its important principles impacting the biopharmaceutical industry as relates to the biopharmaceutical industry and the drug development process. Methods Establishing values of importance to patients As a first step in the process to define patient centricity and its important principles, AstraZeneca conducted an extensive multifaceted research project across key stakeholders to identify issues of importance to patients, healthcare providers and payers. Key questions to be answered through the various work streams were formulated by AstraZeneca and adapted where appropriate to elicit feedback from participants or to identify appropriate existing feedback from published (eg, Ernst and Young proprietary reports) and unpublished (AstraZeneca and Ernst and Young in-house data) sources (box 1). Participants were also given the opportunity to provide unsolicited feedback. Individual patients were identified, and together with patients recruited to an online forum (CreateHealth, see below), were invited to participate in accordance with appropriate ethical guidelines. Questionnaires were used to elicit feedback during telephone interviews with individual patients (n=14) including those affiliated to patient organisations, and with patients (n=200) through the online forum (CreateHealth proprietary crowd-sourcing healthcare platform) in the form of a virtual advisory board. Patients were recruited to the open (not gated) online forum using a variety of digital recruitment tactics (eg, emails, banners, etc) across a range of social platforms (eg, Twitter, Facebook) and healthcare platforms (Patient Organisations, Health Unlocked). Further information was gathered from published and unpublished (AstraZeneca and Ernst and Young) sources. Patient feedback was further augmented with specific information offered by healthcare providers and payers (n=5–10) relating to patient needs during telephone interviews. Box 1 Key questions used to establish values of importance to patients

red from published and unpublished (AstraZeneca and Ernst and Young) sources. Patient feedback was further augmented with specific information offered by healthcare providers and payers (n=5–10) relating to patient needs during telephone interviews. Box 1 Key questions used to establish values of importance to patients How are wider consumer needs, behaviours and expectations changing? What implications do these changes have for organisations? What do patients, carers and patient organisations value when being engaged? What can we learn from providers and payers about being more patient centric?

red from published and unpublished (AstraZeneca and Ernst and Young) sources. Patient feedback was further augmented with specific information offered by healthcare providers and payers (n=5–10) relating to patient needs during telephone interviews. Box 1 Key questions used to establish values of importance to patients How are wider consumer needs, behaviours and expectations changing? What implications do these changes have for organisations? What do patients, carers and patient organisations value when being engaged? What can we learn from providers and payers about being more patient centric? Defining patient centricity and the values of importance to patients Having gained a deeper understanding of patient needs and armed with key topics of interest from this early research, AstraZeneca conducted two identical workshops to create a definition of patient centricity and the healthcare values important to patients and caregivers (referred to as patients). A total of 22 patients or caregivers were identified either directly or through patient organisations and were invited to participate in these workshops in London, UK (n=10 from across Europe), and then in Gaithersburg, USA (n=12 from across North America). Participants had been affected by a variety of illnesses (including anxiety, asthma, Asperger's syndrome, cancer (various), Crohn's disease, depression, diabetes, organ transplantation, psychiatric disease, rare disease). Some of these conditions were deliberately outside the scope of the AstraZeneca drug development strategy, as the company wished to understand the views and behaviours of a cross-section of patients, unrestricted by therapy area. All patients provided consent to participate. They were split into two groups (at each site) and asked to define patient centricity and the healthcare values most important to them. Preliminary definitions defined at both advisory boards were assimilated by the US attendees into a final version that was subsequently circulated to attendees of both advisory boards for their critique/confirmation.

(at each site) and asked to define patient centricity and the healthcare values most important to them. Preliminary definitions defined at both advisory boards were assimilated by the US attendees into a final version that was subsequently circulated to attendees of both advisory boards for their critique/confirmation. Validating the values of importance to patients The principles of importance were then tested through the PatientsLikeMe online patient network, the majority of whose contributors are located in North America. Using an online survey, patients were asked to score each of the patient-centric principles (previously defined) on a scale of 1 to 10, to select their top three principles and provide any additional qualitative feedback. Given the greater North American focus of the PatientsLikeMe exercise, the principles were also tested among patients recruited from five European countries (France, Germany, Italy, Spain, UK) through the Carenity online patient network. The same questionnaire (but translated for local use) and analysis approach was used. No participant selection procedure was employed for the online surveys.

principles were also tested among patients recruited from five European countries (France, Germany, Italy, Spain, UK) through the Carenity online patient network. The same questionnaire (but translated for local use) and analysis approach was used. No participant selection procedure was employed for the online surveys. Results Establishing values of importance to patients Patients reported that they can access a vast amount of conflicting health-related information, and that a critical part of being patient centric is to support access to easily understandable information through the channels they wish to use. Patients felt they need support to make the right decisions regarding their health, and offered that once patient needs are understood they are more open to positive influence. In this respect, the biopharmaceutical industry must rebuild trust and develop open long-term engaging relationships with patients. Access and affordability of medication and support is seen to be a key driver of patient engagement, and provision of a joined-up, system-wide approach to healthcare would provide an opportunity to improve patient experience. Patients also felt that they had to be a ‘pain in the neck’ to get the treatment they wanted and that the healthcare system should support patients to manage themselves. Payers and healthcare providers underscored that improvements in patients' health should be viewed holistically, and that improved relationships with other groups such as nurses, pharmacists and patient advocates are needed to meet patient needs more effectively. In addition, open and transparent access to product information, educational materials and clinical trial data are needed. Lack of patient compliance with treatment plans and fragmented payment systems were also seen as problematic.

s, pharmacists and patient advocates are needed to meet patient needs more effectively. In addition, open and transparent access to product information, educational materials and clinical trial data are needed. Lack of patient compliance with treatment plans and fragmented payment systems were also seen as problematic. Several key implications for the biopharmaceutical industry emerged. Efforts should be made to use communication channels of choice for the patient, and closer relationships should be developed to understand patient experiences, needs and preferences. The biopharmaceutical industry should also provide greater support for patients to achieve the best outcomes, taking a system-wide approach to patient access, including information and care. The industry and other healthcare providers should also help empower patients through the development and provision of self-management tools and services. The payer/provider perspective should also employ a more holistic approach to engaging patients, healthcare providers and payers, ensure transparency especially in relation to data, and work to improve systems and relationships in turn designed to improve health outcomes. Defining patient centricity and the values of importance to patients Definitions centred around four key areas defined as education and information, cocreation, access and transparency (table 1). Table 1 Values/issues identified as being most important to patients (in their own words)

Several key implications for the biopharmaceutical industry emerged. Efforts should be made to use communication channels of choice for the patient, and closer relationships should be developed to understand patient experiences, needs and preferences. The biopharmaceutical industry should also provide greater support for patients to achieve the best outcomes, taking a system-wide approach to patient access, including information and care. The industry and other healthcare providers should also help empower patients through the development and provision of self-management tools and services. The payer/provider perspective should also employ a more holistic approach to engaging patients, healthcare providers and payers, ensure transparency especially in relation to data, and work to improve systems and relationships in turn designed to improve health outcomes. Defining patient centricity and the values of importance to patients Definitions centred around four key areas defined as education and information, cocreation, access and transparency (table 1). Table 1 Values/issues identified as being most important to patients (in their own words) 1. Education and information Patient advocates and patient stories ‘Advocates related to each disease’. (NA) ‘Educate patients through others' experience’. (NA) ‘Support for patient education and patient “ambassadors” who can assist others’. (EU) Helping patients manage their own health and make their own decisions ‘Publish specialist (information) in conditions’. (NA) ‘Customised info/data’. (EU) ‘Support patients on treatments by providing information, insight, peer groups’. (EU) ‘Supporting people to live healthy lives and sharing resources to benefit patients’. (EU) ‘Young creative approach to awareness … Online portal for meds, lifestyle, diet, health tips’. (EU) ‘Knowing consequences to make an informed decision’. (EU) ‘Support and guidance on how I can live with my condition (it's not just the medical side that matters—it has a financial impact, impact on the family, etc)’. (EU) ‘Plain English’. (NA) ‘Plain scientific language’. (EU) ‘Patient-centred medicines leaflet developed by a Pharma company in Germany looked at what patients wanted to know, not just what the pharma medics/regulatory people had to tell them’. (EU) Resources for families, carers and communities ‘Educate—families and communities’. (NA) ‘(Don't) ignore family/caregiver concerns’. (EU) ‘Facilitating family education’. (NA) Resources for healthcare professionals ‘Create information that doctors can distribute’. (NA) ‘Champion the change in medical and organisational culture’. (EU) ‘Patient centricity is not being taught to medical students and therefore there's no incentive to learn how to be patient centric’. (EU) ‘Inadequate education of clinical staff and support personnel’. (NA) Collect patient views and provide information about side effects ‘Include patient voice—side effects’. (NA) ‘Unbiased reporting of effects of treatments’. (EU) ‘What is a tolerable side effect?’ (NA) 2. Cocreation Work with patients and other stakeholders throughout the research, development and launch of medicines ‘Open engagement with stakeholders across drug development pipeline’. (NA) ‘Clinical trial end points patient input’. (NA) ‘Shared goals and outcomes’.

ents’. (EU) ‘What is a tolerable side effect?’ (NA) 2. Cocreation Work with patients and other stakeholders throughout the research, development and launch of medicines ‘Open engagement with stakeholders across drug development pipeline’. (NA) ‘Clinical trial end points patient input’. (NA) ‘Shared goals and outcomes’. (EU) ‘Including the patient in research from the beginning (this will improve compliance too)’. (EU) Research patient's wider needs ‘Recognition of quality of life beyond the medical model’. (EU) ‘Researchers talk to patients to discover unmet need from them’. (NA) ‘Understand patient journey (ideally) begins and ends at their home and not with your product or organisations’. (EU) ‘Spending time with patients and their families to understand their experience’. (EU) Work with patients and other stakeholders to advocate policies in the interests of patients ‘Join and lobby for patient rights regarding meds (not just industry)’. (NA) ‘Encourage and support patient lobby movement’. (EU) Group of stakeholders to codesign solutions ‘Pharmaceutical companies could develop a cross-healthcare advisory group to help “cocreate” solutions instead of just sponsorship’. (NA) 3. Access Support service to help patients navigate complex health systems and issues ‘Create Public Service Agreements (PSAs) to educate’. (NA) ‘Have patient advisors 24/7’. (NA) ‘Access to right therapy and information to support decision-making’. (EU) Patient assistance programme ‘Compassion programme so no one is left out’. (NA) Flexible pricing policy ‘Make products and services affordable and have fair and honest pricing’. (NA) 4. Transparency Transparency with clinical trial data ‘Share research findings to advance overall research’. (NA) ‘Trial results reported up to trial participants’. (NA) ‘Unbiased reporting of all trials’. (EU) When developing medicines, report on the measurable patient benefit and patient safety ‘Create the safest and most efficacious drugs’. (USA) Respond to patients’ feedback—show how you have listened ‘Value and respond to feedback’. (NA) ‘Good communication. Follow-up with backup’. (NA) ‘Better medium between industry and patients’. (NA) ‘Listening and learning from patients takes new ways of working/training’. (EU) ‘Longevity of relationship not just project-focused’. (EU) Open and accountable reporting of progress ‘An annual report of AZ's patient-centred activities’. (EU) ‘Accountability from the AZ Board’. (EU) ‘Measuring (patients’) influence and impact on AZ’.

learning from patients takes new ways of working/training’. (EU) ‘Longevity of relationship not just project-focused’. (EU) Open and accountable reporting of progress ‘An annual report of AZ's patient-centred activities’. (EU) ‘Accountability from the AZ Board’. (EU) ‘Measuring (patients’) influence and impact on AZ’. (EU) ‘An action plan with measures’. (NA) ‘Accountable for actions’. (NA) ‘Socially responsible’. (NA) Values-based approach to business development ‘Balanced and ethical value proposition between profits and patients’. (NA) EU, European Union; NA, North America. Four variations of a definition were created by the workshop participants which were then combined by the North America attendees to create the most appropriate representative statement (figure 1). This was circulated to all participants and received unanimous support. Patient centricity was thus defined as: ‘Putting the patient first in an open and sustained engagement of the patient to respectfully and compassionately achieve the best experience and outcome for that person and their family’. Figure 1 Patient centricity—definition evolution.

Four variations of a definition were created by the workshop participants which were then combined by the North America attendees to create the most appropriate representative statement (figure 1). This was circulated to all participants and received unanimous support. Patient centricity was thus defined as: ‘Putting the patient first in an open and sustained engagement of the patient to respectfully and compassionately achieve the best experience and outcome for that person and their family’. Figure 1 Patient centricity—definition evolution. This definition encompassed five clear themes of importance: (1) inclusiveness; (2) sharing goals that are patient and family-centred; (3) empowering patients to take control of their own health; (4) working in a way that shows respect, compassion and openness; and (5) working in partnership (figure 1). Key words in the definition were underscored by more detailed descriptions. Thus, ‘open and sustained engagement’ reflects the need to act transparently in all interactions with patients and to maintain consistent relationships rather than those that are driven solely by project needs. ‘Respectfully’ reflects the need to foster a mutual understanding between both parties and to understand sometimes diverse viewpoints. ‘Compassionate’ underscores the need to empathise with a person, their condition and the impact it has on their lives, but at the same time recognising the valuable contribution that patients contribute in terms of healthcare decisions. ‘Best experience and outcome’ engenders the concept of supporting the whole person (rather than just the condition) through therapeutic and non-therapeutic means and understanding that all outcomes are unique for an individual. This requires an appreciation of the impact that each individual outcome has on patient's families, friends and the wider community and the need to respect patient's definitions of a desired outcome.

ndition) through therapeutic and non-therapeutic means and understanding that all outcomes are unique for an individual. This requires an appreciation of the impact that each individual outcome has on patient's families, friends and the wider community and the need to respect patient's definitions of a desired outcome. Ten principles were identified of importance to patients and caregivers for biopharmaceutical companies that centred around the previously defined key areas of education and information, cocreation, access and transparency (box 2). Box 2 Principles of importance to patients and caregivers* The company helps make sure that the people who need medicines have access to them. The company communicates transparent and unbiased information on your disease, treatment options, and available resources with care and compassion. The company provides easy-to-understand and convenient information in plain language because they understand that ‘words matter’. The company helps you gain affordable access to their medications. The company equips you to make informed choices about your healthcare and your treatment options. The company listens and responds to your feedback with respect and humility. The company partners with you to innovate and measure impact and outcomes that are important to you. The company provides access to support programmes and resources to help you improve quality of life. Everything the company does begins with an understanding of your needs and experiences. The company helps empower you to help other patients and their families. *Principles are not ranked by importance.

The company partners with you to innovate and measure impact and outcomes that are important to you. The company provides access to support programmes and resources to help you improve quality of life. Everything the company does begins with an understanding of your needs and experiences. The company helps empower you to help other patients and their families. *Principles are not ranked by importance. Validating the values of importance to patients and caregivers A total of 479 patients with cancer (n=91), respiratory, inflammation or autoimmune disease (n=246) or cardiovascular or metabolic conditions (n=142) participated in the initial validation exercise through the PatientsLikeMe network; respondents were mainly from the USA (n=320 vs 159 non-US). Patients confirmed that all ten principles were important for a biopharmaceutical company to embody, although patient empowerment received the lowest scores. When participants listed their top three principles, the access principles were cited most frequently (54% of respondents), followed by education and information principles focused on unbiased information (42%) and informed decisions principles (38%). Of the principles not listed, a small proportion of patients (4%) raised the need to ensure open communication on adverse events and side effects.

e cited most frequently (54% of respondents), followed by education and information principles focused on unbiased information (42%) and informed decisions principles (38%). Of the principles not listed, a small proportion of patients (4%) raised the need to ensure open communication on adverse events and side effects. The principles were also tested in 703 patients with cancer (n=99), respiratory, inflammation or autoimmune disease (n=192) or cardiovascular or metabolic conditions (n=412) recruited from five European countries (France n=256; Germany n=99; Italy n=158; Spain n=71; UK n=119) through the Carenity patient network. As with the PatientsLikeMe exercise, European patients confirmed that all 10 principles were important for a biopharmaceutical company to embody, although again patient empowerment received the lowest scores. When participants selected their top three principles, provision of education and information were cited most frequently (39–40% of respondents) followed by education and access (39%). Of the principles not listed, most participant suggestions related to ‘putting people first’ (25%), followed by ‘clear information on side effects’ (16%) and then ‘improve transparency/honesty’ (14%). Of note, the need for clear information on side effects was a facet highlighted by the North American and the European groups.

principles not listed, most participant suggestions related to ‘putting people first’ (25%), followed by ‘clear information on side effects’ (16%) and then ‘improve transparency/honesty’ (14%). Of note, the need for clear information on side effects was a facet highlighted by the North American and the European groups. Discussion Patient centricity is becoming a core element of medicine development and value-based healthcare management. Health outcomes are dependent on patient engagement and there is a clear need for the biopharmaceutical industry to partner with patients in the discovery, research, development, distribution and access to medicines to bring about better outcomes. Numerous patient engagement initiatives have started in the last few years, involving regulatory bodies, patient advocate groups, healthcare providers and payers, and the biopharmaceutical industry. Further, much dialogue and discussion has arisen around patient centricity but despite this process of ‘cocreation’ gaining momentum, there is still much to do. A consistent definition of patient centricity as applied to drug development has not been published, and although definitions relating to other key stakeholders (eg, healthcare providers24) have been developed, these are not transferable to the drug development process. We need a consistent and coordinated approach within the biopharmaceutical industry.

centricity as applied to drug development has not been published, and although definitions relating to other key stakeholders (eg, healthcare providers24) have been developed, these are not transferable to the drug development process. We need a consistent and coordinated approach within the biopharmaceutical industry. Against this background, a consistent definition of patient centricity and its associated principles for the biopharmaceutical industry were cocreated by patients and industry representatives from AstraZeneca through an extensive consultation exercise. These principles are consistent with those identified by other groups.26 Our call is for the biopharmaceutical companies to adopt this definition and the important principles which can then be used as a reference point for patient engagement throughout the product life cycle. By incorporating the definition and principles into the drug discovery and development process and beyond, we aspire to achieve measurable and positive improvements in patient outcomes. Indeed, AstraZeneca has already begun the process as described with incorporation of patient insight, cocreation of solutions for patients' needs and implementing methods for measuring the effectiveness of this approach. In this respect, AstraZeneca has embarked on a programme to ensure patient centricity is at the heart of medicine development and associated support/educational services. Overall, the patient-centric programme is designed to address patients' needs and to empower patients to connect patients with their providers, their community and other patients. It is intended to be sustainable and scalable, dynamic and to evolve as patient engagement continues. Ultimately, it is designed to have a measurable positive impact on patients' lives. Incorporating patient insight is an iterative process that can begin during early medicine development and continue throughout continuum of drug development. Patient insights can be used to inform disease area strategies, product development, clinical protocols, regulatory interactions and patient services.

nts' lives. Incorporating patient insight is an iterative process that can begin during early medicine development and continue throughout continuum of drug development. Patient insights can be used to inform disease area strategies, product development, clinical protocols, regulatory interactions and patient services. We employed a comprehensive process to identify the principles of importance for patient engagement and to cocreate the definition of patient centricity. Although relatively small patient groups were engaged to refine these points, validation of the principles of importance was obtained from a larger cohort. The definition of patient centricity did not undergo further validation but AstraZeneca implemented patient centricity programmes on the basis of this definition. The research was also conducted in Europe and the USA and therefore may not be transferable to other geographical regions or where there are different healthcare models.