Browse the corpus

Introduction Background Patients with suspected acute coronary syndrome are responsible for 6% of emergency department (ED) presentations.1 The majority of these patients will not receive a diagnosis of myocardial infarction,2 but clinical guidelines have recommended serial cardiac troponin-level testing to safely rule out the diagnosis, which often requires admission to the hospital.3,4 Because EDs are under increasing pressure to reduce the number of patients admitted to the hospital,5 the use of high-sensitivity troponin assays to exclude myocardial infarction earlier or at presentation may increase efficiency in the ED setting.Editor’s Capsule Summary What is already known on this topic Some emergency departments (EDs) are using rapid myocardial infarction rule-out pathways to avoid hospitalization of patients with potential acute coronary syndrome. What question this study addressed This qualitative study assessed patient impressions before and after implementation of an ED rapid rule-out pathway. Forty-nine patients were interviewed 1 week postdischarge and 5 major themes emerged across both pathways. What this study adds to our knowledge A discordance may emerge between physician relief at the absence of acute coronary syndrome according to normal troponin levels and patient concern for continued unexplained chest symptoms. Patients may also be less likely to assess their health behaviors and risk of future cardiovascular disease. How this is relevant to clinical practice

A discordance may emerge between physician relief at the absence of acute coronary syndrome according to normal troponin levels and patient concern for continued unexplained chest symptoms. Patients may also be less likely to assess their health behaviors and risk of future cardiovascular disease. How this is relevant to clinical practice Understanding patient perceptions may help physicians provide better care for ED patients with similar diagnostic trajectories.

A discordance may emerge between physician relief at the absence of acute coronary syndrome according to normal troponin levels and patient concern for continued unexplained chest symptoms. Patients may also be less likely to assess their health behaviors and risk of future cardiovascular disease. How this is relevant to clinical practice Understanding patient perceptions may help physicians provide better care for ED patients with similar diagnostic trajectories. Importance The European Society of Cardiology and the National Institute for Health and Care Excellence have endorsed early rule-out pathways based on high-sensitivity cardiac troponin assays.6,7 The Food and Drug Administration has now also approved high-sensitivity cardiac troponin assays for clinical use. A number of strategies have been proposed to identify patients at presentation or 1 to 2 hours after presentation who may be suitable for discharge directly from the ED.8, 9, 10, 11, 12, 13, 14 Although the adoption of these rule-out pathways could improve efficiency in the ED and therefore lead to major benefits for health care providers, patients will spend less time within the health care setting and may have fewer assessments from specialists and opportunities to discuss the nature of their symptoms, and as a consequence may be less likely to be reassured their symptoms are benign. Previous research into the experience of patients with acute chest pain who present to the ED has shown that they may be discharged with unanswered questions,15 feelings of uncertainty,16 and the need to feel more supported after discharge.17

s a consequence may be less likely to be reassured their symptoms are benign. Previous research into the experience of patients with acute chest pain who present to the ED has shown that they may be discharged with unanswered questions,15 feelings of uncertainty,16 and the need to feel more supported after discharge.17 Goals of This Investigation In this qualitative study embedded into a prospective clinical trial,18 we aimed to explore the experience of 2 groups of patients undergoing assessment for suspected acute coronary syndrome before and after implementation of an early rule-out pathway for myocardial infarction to identify how the assessment pathway affects the patients’ experience. The insights revealed will ensure that early rule-out pathways can be applied to patient assessment in a way that responds to patients’ needs. Materials and Methods Study Design and Setting This qualitative study was embedded into a clinical trial evaluating the safety and efficacy of an early rule-out pathway for myocardial infarction across secondary and tertiary care hospitals in Scotland.18 The qualitative component was conducted at the lead site in the ED of the Royal Infirmary of Edinburgh, a tertiary care hospital. The clinical trial was approved by the national research ethics committee and conducted in accordance with the Declaration of Helsinki. This qualitative substudy was prespecified in the trial protocol.

tive component was conducted at the lead site in the ED of the Royal Infirmary of Edinburgh, a tertiary care hospital. The clinical trial was approved by the national research ethics committee and conducted in accordance with the Declaration of Helsinki. This qualitative substudy was prespecified in the trial protocol. Selection of Participants Patients with suspected acute coronary syndrome were recruited from the ED between March 2015 and June 2017. Patients older than 18 years, for whom the attending clinician requested cardiac troponin-level testing for suspected acute coronary syndrome, and who were discharged on the basis of a negative evaluation result for myocardial infarction were eligible for inclusion. Eligibility was verified with the electronic patient record. The main trial recruited consecutive patients, but the substudy used purposive sampling to ensure representation across age and sex categories (male patients, female patients, >65 years, and ≤65 years). In a clinical scenario, the pathway is applied nonselectively to any patient presenting with symptoms suggestive of acute coronary syndrome. It was therefore thought important that a broad spectrum of patients be recruited to this study; therefore, participants were not stratified further. Sampling and recruitment occurred deliberately slowly to allow concurrent data collection and analysis.

nt presenting with symptoms suggestive of acute coronary syndrome. It was therefore thought important that a broad spectrum of patients be recruited to this study; therefore, participants were not stratified further. Sampling and recruitment occurred deliberately slowly to allow concurrent data collection and analysis. The early rule-out pathway was implemented on February 15, 2016. Before this date, patients were admitted to a medical assessment unit or cardiology ward for serial troponin-level testing to determine peak troponin level at 10 to 12 hours after symptom onset. Eligible participants were those discharged in accordance with negative serial test results for myocardial infarction. Follow-up care was at the discretion of the assessing clinician. After this date, myocardial infarction was excluded in accordance with troponin concentration at presentation or at 3 hours, enabling patients to be discharged directly from the ED.19 Follow-up care remained at the discretion of the assessing clinician. The postdischarge care of patients with the rule-out of myocardial infarction was not altered by the implementation of the early rule-out pathway. Identification of study participants according to pathway is shown in Figure 1A and B.Figure 1 Summary of chest pain pathway and identification of eligible participants. A, Before implementation of the early rule-out pathway. B, After implementation of the early rule-out pathway. GRACE, Global Registry of Acute Coronary Events; AMU, acute medical unit.

ts according to pathway is shown in Figure 1A and B.Figure 1 Summary of chest pain pathway and identification of eligible participants. A, Before implementation of the early rule-out pathway. B, After implementation of the early rule-out pathway. GRACE, Global Registry of Acute Coronary Events; AMU, acute medical unit. The early rule-out pathway was developed in a substudy of patients presenting with suspected acute coronary syndrome between June 1, 2013, and September 30, 2015. Blood samples were obtained at presentation and at 6 to 12 hours after symptoms onset for high-sensitivity cardiac troponin-level testing as part of routine clinical care. Patients provided informed consent to obtain an additional blood sample at 3 hours as previously described.19,20 The troponin assay used was the Abbott ARCHITECTSTAT high-sensitivity cardiac troponin I assay (Abbott Laboratories, Abbott Park, IL).21 The upper reference limit 99th centiles were determined in 4,590 samples from healthy individuals as 16 ng/L for women and 34 ng/L in men.22

e at 3 hours as previously described.19,20 The troponin assay used was the Abbott ARCHITECTSTAT high-sensitivity cardiac troponin I assay (Abbott Laboratories, Abbott Park, IL).21 The upper reference limit 99th centiles were determined in 4,590 samples from healthy individuals as 16 ng/L for women and 34 ng/L in men.22 Data Collection and Processing Participants were approached by A.V.F. or an ED research nurse with an information sheet while in the ED or medical assessment unit (for those recruited in the preimplementation phase). Participants were contacted by telephone at least 24 hours after discharge to discuss participation and arrange a date for interview 1 week postdischarge if appropriate to do so. After written informed consent was gained, interviews were conducted in a place of the participants’ choosing, including their own home (16 pre- and 21 postimplementation), the hospital (5 pre- and 2 postimplementation), or a private meeting room at their workplace (2 pre- and 3 postimplementation). For the majority of interviews, only the interviewer and participant were present, although a family member was present during 5 interviews.

their own home (16 pre- and 21 postimplementation), the hospital (5 pre- and 2 postimplementation), or a private meeting room at their workplace (2 pre- and 3 postimplementation). For the majority of interviews, only the interviewer and participant were present, although a family member was present during 5 interviews. A research diary was kept to document reflections after each interview, and regular debriefing with a supervisor (S.C.-B.) was used to promote researcher reflexivity.23 This also served as a decision trail to demonstrate how interpretative analysis evolved during the study and therefore increased the trustworthiness of the study.24 Interviews lasted between 18 and 88 minutes. Data collection and analysis occurred concurrently. Recruitment continued until saturation was achieved and additional interviews did not yield new insights.25

te how interpretative analysis evolved during the study and therefore increased the trustworthiness of the study.24 Interviews lasted between 18 and 88 minutes. Data collection and analysis occurred concurrently. Recruitment continued until saturation was achieved and additional interviews did not yield new insights.25 A female cardiology research nurse with experience in qualitative interviewing techniques (A.V.F.) conducted semistructured, detailed interviews, using a topic guide developed from a literature review and clinical experience of the study team (Appendix E1, available online at http://www.annemergmed.com). The interviews proceeded as a guided conversation, ensuring the same range of topics was covered yet allowing respondents flexibility in how they answered and in introducing new issues as relevant to them. A.V.F. was not involved in the clinical care of patients and was introduced to participants as a researcher. If participants questioned the interviewer directly, her identity as a nurse was revealed. Interviews all started with the same opening question: “Could you tell me what happened to take you into [the] hospital last week?” Interviews were audio recorded and transcribed verbatim by a professional transcription company. Each transcript was checked for accuracy against the audio file by A.V.F. Data are presented as quotes from transcripts, with “I” prefixing interviewer speech and “P” participant speech.

you into [the] hospital last week?” Interviews were audio recorded and transcribed verbatim by a professional transcription company. Each transcript was checked for accuracy against the audio file by A.V.F. Data are presented as quotes from transcripts, with “I” prefixing interviewer speech and “P” participant speech. Primary Data Analysis This study was guided by broadly phenomenological principles,26 aiming to uncover the meaning and relevance of experience. An interpretive approach was used to analyze data thematically,27 using abductive reasoning, which seeks to identify meaning from the accounts in iteration with previous knowledge from the field.28 This involves repeatedly reading and interpreting accounts to search for patterns of meaning in the data. Data were coded by A.V.F. under the guidance of an experienced senior qualitative researcher (S.C.-B.). Interviews were read multiple times, and relationships between codes were explored to identify themes that were derived from the data. The grouping of codes into themes was performed through discussion with S.C.-B., including review of the patient narratives and coding categories. Through this process, the themes and model were created (Figure 2). Themes arising within patient accounts from both pathways were compared to search for similarities and differences in which the chest pain pathway may be implicated. Data were managed with NVivo (version 10; QSR International, Victoria, Australia). When clear differences between care pathways were observed, the results were quantified. When themes were pervasive across both pathways, quantification of results did not add value to the analysis.Figure 2 Communication interventions aiding the development of reassurance.

n 10; QSR International, Victoria, Australia). When clear differences between care pathways were observed, the results were quantified. When themes were pervasive across both pathways, quantification of results did not add value to the analysis.Figure 2 Communication interventions aiding the development of reassurance. Results Characteristics of Study Subjects One hundred forty-three patients were approached with an information sheet; 23 participants were interviewed before and 26 participants after implementation of the early rule-out pathway. Reasons for nonparticipation are outlined in Figure 3. The mean age of men was 59 years (SD 14 years) preimplementation and 60 years (SD 15 years) postimplementation, and for women it was 58 years (SD 15 years) preimplementation and 61 years (SD 15 years) postimplementation. The median length of hospital stay was 10.5 hours (interquartile range 8.2 to 12.3 hours) and 3.4 hours (interquartile range 2.5 to 3.9 hours) before and after implementation of the early rule-out pathway, respectively. Further details of study participants are reported in Table 1.Figure 3 Reasons for nonparticipation of potential study participants. Table 1 Baseline characteristics of study population.

Results Characteristics of Study Subjects One hundred forty-three patients were approached with an information sheet; 23 participants were interviewed before and 26 participants after implementation of the early rule-out pathway. Reasons for nonparticipation are outlined in Figure 3. The mean age of men was 59 years (SD 14 years) preimplementation and 60 years (SD 15 years) postimplementation, and for women it was 58 years (SD 15 years) preimplementation and 61 years (SD 15 years) postimplementation. The median length of hospital stay was 10.5 hours (interquartile range 8.2 to 12.3 hours) and 3.4 hours (interquartile range 2.5 to 3.9 hours) before and after implementation of the early rule-out pathway, respectively. Further details of study participants are reported in Table 1.Figure 3 Reasons for nonparticipation of potential study participants. Table 1 Baseline characteristics of study population. All Preimplementation Postimplementation Study sample Participants, No. (% women) 49 (45) 23 (39) 26 (50) Women >65 y, No. (%) 12 (25) 4 (18) 8 (31) Women ≤65 y, No. (%) 10 (20) 5 (22) 5 (19) Men >65 y, No. (%) 16 (33) 7 (30) 9 (34) Men ≤65 y, No. (%) 11 (22) 7 (30) 4 (16) Men, age, y Mean (SD) 60 (14.4) 59 (4.4) 60 (14.9) Range 20–85 33–85 20–83 Women, age, y Mean (SD) 59 (14.6) 58 (15.2) 61 (15.3) Range 35–82 35–72 41–82 Medical history, No. (%) Smoking 20 (41) 11 (48) 9 (35) Diabetes mellitus 6 (12) 4 (17) 2 (8) Hypertension 22 (45) 9 (39) 13 (50) Hyperlipidemia 19 (39) 9 (39) 10 (38) Family history 25 (51) 14 (61) 11 (42) Angina 11 (22) 5 (22) 6 (23) Myocardial infarction 15 (31) 9 (39) 6 (21) Previous PCI 11 (22) 7 (30) 4 (15) Previous CABG 2 (4) 1 (4) 1 (4) Heart failure 0 0 0 Cerebrovascular disease 0 0 0 Peripheral vascular disease 1 (2) 0 1 (4) Length of stay, median (interquartile range), h —∗ 10.5 (8.2–12.3) 3.4 (2.5–3.9) PCI, Percutaneous intervention; CABG, coronary artery bypass graft.

(39) 6 (21) Previous PCI 11 (22) 7 (30) 4 (15) Previous CABG 2 (4) 1 (4) 1 (4) Heart failure 0 0 0 Cerebrovascular disease 0 0 0 Peripheral vascular disease 1 (2) 0 1 (4) Length of stay, median (interquartile range), h —∗ 10.5 (8.2–12.3) 3.4 (2.5–3.9) PCI, Percutaneous intervention; CABG, coronary artery bypass graft. ∗ Dash indicates value not calculated Main Results Five important concepts could be identified in the accounts of participants interviewed both before and after implementation of the early rule-out pathway. Where variance in response was noted between pathways, these differences have been highlighted.

(39) 6 (21) Previous PCI 11 (22) 7 (30) 4 (15) Previous CABG 2 (4) 1 (4) 1 (4) Heart failure 0 0 0 Cerebrovascular disease 0 0 0 Peripheral vascular disease 1 (2) 0 1 (4) Length of stay, median (interquartile range), h —∗ 10.5 (8.2–12.3) 3.4 (2.5–3.9) PCI, Percutaneous intervention; CABG, coronary artery bypass graft. ∗ Dash indicates value not calculated Main Results Five important concepts could be identified in the accounts of participants interviewed both before and after implementation of the early rule-out pathway. Where variance in response was noted between pathways, these differences have been highlighted. First, it was rare for a patient to make the decision to attend the hospital for assessment independently without previous contact with the health service. Participants revealed a careful consideration about their personal justification to use the health service, resulting in their seeking confirmation that their symptoms warranted professional health care assessment. Only 7 of 49 patients attended the ED without consulting a second party. Many patients (61% [30/49]) had sought advice from a general practitioner or NHS 24 (a telephone advice and triage service that runs out of hours in Scotland), with the recommendation to attend the hospital for assessment, or the even stronger message of “telephone 999” (the equivalent of 911 in the United States). The remainder of patients consulted lay networks for advice. It was also apparent in several transcripts that receiving advice from a general practitioner or NHS 24 call handler to attend the ED for assessment may actually strengthen patients’ belief that their symptoms are serious. This is illustrated through participant quotes in Table 2.Table 2 Evidence for themes of help seeking behavior, discordance, and reassurance.

that receiving advice from a general practitioner or NHS 24 call handler to attend the ED for assessment may actually strengthen patients’ belief that their symptoms are serious. This is illustrated through participant quotes in Table 2.Table 2 Evidence for themes of help seeking behavior, discordance, and reassurance. Theme Evidence Participant Help-seeking behavior I: So why NHS 24? P: That was…I don’t know. That was…to me, that’s a step before you phone 999 for an ambulance. Erm, and then you’re giving somebody else the decision to… I: Yes. P: …[I]f, though…if they think it’s serious enough, then I, I…I’m wary about taking up people’s time, et cetera…. I: Okay. P: …[I]f it’s, you know, not warranted, sort o’ thing. Participant 19, >65 y, woman (preimplementation) P: Well, it’s, it was certainly, um, again, I mean, it was one of these situations where, you know, I had to phone up and the receptionist said, “Is it an emergency,” and I said, “Well, all I can tell you is this is what’s happened….” I: Yeah. P: …[T]his has been my experience, and she said…so she made the de-decision, the receptionist made the decision…[to telephone an ambulance]. I: Right. P: …[W]hich was great because I didn’t want to be wasting the, the GP’s [general practitioner’s] time…. Participant 38, ≤65 y, man (postimplementation) I: So when you got the advice to attend A&E… P: Mm-hmm. I: …[H]ow did you feel about that? P: Oh, that…that…it…it then went from “I’ve got a pain in my chest” to “I’ve got a pain in my chest and someone thinks that my symptoms are serious.” I: Right. P: Um, so it…it escalated a wee bit. So I went from thinking, well, maybe there’s something wrong to, well, it’s maybe more than a maybe. Participant 49, >65 y, man (postimplementation) Discordance I: I’d like to explore the reason why you don’t feel you can go out. P: I just thought…I think…I just thought I got such a fright, really… I don’t know if I still thought I could see me having a heart attack…. Participant 5, >65 y, woman (preimplementation) P: At the time they told me I could go home, that felt good…. I know there’s a bigger picture. Participant 11, ≤65 y, man (preimplementation) P: I just know there is something going on. Whether it’s my heart or not I don’t know. Participant 38, ≤65 y, man (postimplementation) P: At the back of my mind you’re thinking is this my heart? Am I going to have a heart attack? Participant 48, ≤65 y, woman (postimplementation) Reassurance P: There was absolutely no damage to my heart.

re is something going on. Whether it’s my heart or not I don’t know. Participant 38, ≤65 y, man (postimplementation) P: At the back of my mind you’re thinking is this my heart? Am I going to have a heart attack? Participant 48, ≤65 y, woman (postimplementation) Reassurance P: There was absolutely no damage to my heart. Participant 2, >65 y, woman (preimplementation) P: I feel quite confident about my heart now because of the tests. I’ve no concerns about my heart. Participant 15, >65 y, man (preimplementation) P: Once I was discharged with a clean bill, I parked it. Participant 24, ≤65 y, man (postimplementation)

re is something going on. Whether it’s my heart or not I don’t know. Participant 38, ≤65 y, man (postimplementation) P: At the back of my mind you’re thinking is this my heart? Am I going to have a heart attack? Participant 48, ≤65 y, woman (postimplementation) Reassurance P: There was absolutely no damage to my heart. Participant 2, >65 y, woman (preimplementation) P: I feel quite confident about my heart now because of the tests. I’ve no concerns about my heart. Participant 15, >65 y, man (preimplementation) P: Once I was discharged with a clean bill, I parked it. Participant 24, ≤65 y, man (postimplementation) Second, and also common to participants assessed both pre- and postimplementation of the early rule-out pathway, a discordance between the objective interpretation of a troponin concentration by a clinician and the ongoing illness episode experienced by the patient was illustrated. Patient accounts revealed that, although a medical consultation may have concluded for the clinician with a negative evaluation result for acute coronary syndrome and a perception that reassurance had been provided, for some patients, their illness episode was still very much ongoing at the subsequent interview, whichever clinical pathway they had followed. A discordance was therefore found to exist between the objective interpretation of a troponin value by the clinician and the significance that result held to the patient in the context of his or her illness experience. When the pathway, driven by the high negative predictive value of a low troponin concentration, reassured the clinician that the patient did not warrant further investigation (inferred by the decision to discharge the patient), for some patients, this reassurance was not perceived despite clinical notes stating that reassurance had been given. Some patients left the hospital without a satisfactory conclusion to their illness episode, with ongoing questions about the cause of their pain and some patients still believing that their pain may have had a cardiac substrate. Participant quotes provide illustrations in Table 2.

ating that reassurance had been given. Some patients left the hospital without a satisfactory conclusion to their illness episode, with ongoing questions about the cause of their pain and some patients still believing that their pain may have had a cardiac substrate. Participant quotes provide illustrations in Table 2. Third, common to both pathways and suffusing the interviews, was the theme of “reassurance,” which appeared both implicitly and explicitly. When discordance as described above was present, reassurance appeared much more difficult to achieve. An initial code of “completed interaction,” signifying that the patient perceived that the health care encounter had come to an end, was used to conceptualize the meaning of reassurance in a clinical context. Data arising from the interviews suggest that reassurance is a process that has to be built atop certain foundations laid during the clinical assessment. Reassurance was more apparent within patient accounts if an alternative diagnosis about the cause of the chest pain was offered, if the participant was referred for outpatient investigations signaling ongoing care, or if the participant had a very low level of concern that the pain may have had a cardiac cause. For participants who did believe their pain could be cardiac in nature, the development of reassurance was influenced by 3 contributing factors: timing of information giving, patient-clinician interaction, and the development of trust. These factors all involve effective communication between the patient and clinician. Providing information about troponin testing and the possible meaning of results before actual testing appeared to relate to positive expressions of reassurance within interviews. Evidence of the clinician’s using active listening that acknowledged the patients concerns also enabled trust in the clinician to develop. When the factors of pretest information, active listening, and trust were satisfied, participant perception of closure of the acute illness episode could be observed in interviews. This model is presented in Figure 2. Participant quotes provide illustrations in Table 2.

s also enabled trust in the clinician to develop. When the factors of pretest information, active listening, and trust were satisfied, participant perception of closure of the acute illness episode could be observed in interviews. This model is presented in Figure 2. Participant quotes provide illustrations in Table 2. The standard assessment procedures carried out by hospital staff could be interpreted by patients in a manner different from that intended. The efficiency and speed with which ED staff carried out initial assessment procedures were interpreted by some participants as confirmation that their symptoms were a cause for concern. Likewise, the routine nature of repeated blood sampling for peak troponin level as required by the preimplementation pathway could also be interpreted by participants to signify a higher likelihood that their symptoms may have been due to myocardial infarction. Additionally, some participants spoke in terms of being part of a “process.” For some participants, the routine nature of the assessment process was evidence that clinicians would perform appropriate actions because they were following a protocol. For others, the protocol-driven actions were interpreted as lack of personalized care.

some participants spoke in terms of being part of a “process.” For some participants, the routine nature of the assessment process was evidence that clinicians would perform appropriate actions because they were following a protocol. For others, the protocol-driven actions were interpreted as lack of personalized care. Expressions of negativity concerning ambiguity in regard to an overnight admission to the hospital and the need to repeat a symptom history to multiple practitioners were common in interviews of participants assessed before implementation of the early rule-out pathway (78%; 18/23) but less so in those assessed after its implementation (15%; 4/26). These interpretations are illustrated through participant quotes in Table 3.Table 3 Evidence for themes of influence of hospital routines and approaches to future health.

cipants assessed before implementation of the early rule-out pathway (78%; 18/23) but less so in those assessed after its implementation (15%; 4/26). These interpretations are illustrated through participant quotes in Table 3.Table 3 Evidence for themes of influence of hospital routines and approaches to future health. Theme Evidence Participant Influence of hospital routines P: The only thing that I’ve found a bit irritating was the inconsistency of the doctors when I got in there. You’re going home. You’re staying in. You’re going home. You’re staying in. So there was 3 different doctors, told me different things. Participant 4, >65 y, woman (preimplementation) P: It was frustrating, you know, to have to tell the nurse what had happened, and then frustrating to have to tell someone else what had happened, and then a doctor what had happened, and then the consultant what had happened; you know what I mean. So there was, I was thinking, Jesus, can we not just get everybody in the room, and I'll tell you, look, here is what happened, guys. Participant 23, ≤65 y, man (preimplementation) P: And, uh, then a lady came back and then she said I needed to take an aspirin and I would need to stay in till after 12 to get another blood test, because it...she said if it was the heart and any damage had been done, this test showed up something that’s released into the blood. Um, and then I thought, oh, no. Then it was...slight panic set in, because I thought, it’s not as straightforward as I thought. What if they’ve found something? Participant 14, >65 y, man (preimplementation) P: If, if the emergency department are going through their protocols, then clearly that’s their protocols for, for that. If, however, there is, something that’s flagging up, then I think it should either be referred back to the GP [general practitioner] to take up…or sent to whoever needs to make, you know, the decisionmaker. I think not being listened to is critical. Participant 38, ≤65 y, man (postimplementation) Approaches to future health P: Now, for me, I would have said, I’m 53, history of…aged myself prematurely here…um, you know, cardiac problem history in the family, overweight, don’t smoke and things, so those are the risk factors, aren’t they? But I would have probably seen that as an opportunity to say, okay, you’ve maybe had a bit of a scare here; these are the things you should look out for if this happens again.

turely here…um, you know, cardiac problem history in the family, overweight, don’t smoke and things, so those are the risk factors, aren’t they? But I would have probably seen that as an opportunity to say, okay, you’ve maybe had a bit of a scare here; these are the things you should look out for if this happens again. Because there was none of that advice, in terms of, right, if…this is what you…so if this pain happens again, that’s okay, ’cause that’s just your frozen shoulder, but these are the, the warning signs you should maybe look out for, or these are the things you should be doing to reduce your risk of heart…problems, or even go and see your GP [general practitioner] for a general checkup…. Participant 20, ≤65 y, woman (preimplementation) P: Well, it wasn't a heart attack this time. Will there be a heart attack next time? You know, that, that's my concern. Um, I need to change some lifestyle things which I know about, and I will, I am. Um, but I need to also get to the, the root cause of the stress, and anxiety bit, which is work related. Participant 11, ≤65 y, man (preimplementation)

k this time. Will there be a heart attack next time? You know, that, that's my concern. Um, I need to change some lifestyle things which I know about, and I will, I am. Um, but I need to also get to the, the root cause of the stress, and anxiety bit, which is work related. Participant 11, ≤65 y, man (preimplementation) The final theme, and where a further difference was observed between chest pain pathways, was the way in which participants made use of their acute chest pain presentation to the hospital as an opportunity to consider their future heart health. “Approaches to future health” was an unelicited theme within interview transcripts. Participants demonstrating an awareness of future heart health did so in 3 main ways. First, they discussed their incentive to modify their lifestyle as a result of an acute chest pain admission. Second, some participants suggested their acute chest pain presentation and assessment was an appropriate opportunity for health promotion activities. Third, some participants discussed how the rule-out of myocardial infarction related to their overall heart health and their future susceptibility to heart disease. Some patients made no reference to their future health during the course of the interview. This analysis has consequently revealed 3 possible perspectives by which participants may relate to their future health status. For some participants, continuing good health was taken for granted and therefore did not have particular salience in their everyday lives. For others, the way in which they reacted to the chest pain episode varied in accordance with their position in the adult life course and their current health status. For example, their current health status because of comorbidities appeared to have dominance over the acute chest pain episode, leading to discourses of fatalism and certainty of future ill health. Other participants used the chest pain presentation, and therefore the recognition of a physical manifestation of ill health, as a trigger to appraise health behaviors and assess their future risk of cardiovascular disease. This aspect was more commonly observed in patients interviewed before implementation of the early rule-out pathway (43% [10/23] preimplementation versus 19% [5/26] postimplementation). Participant quotes provide illustrations in Table 3. Additional illustrations for all themes are shown in Table E1 (available online at http://www.annemergmed.com).

observed in patients interviewed before implementation of the early rule-out pathway (43% [10/23] preimplementation versus 19% [5/26] postimplementation). Participant quotes provide illustrations in Table 3. Additional illustrations for all themes are shown in Table E1 (available online at http://www.annemergmed.com). Limitations Our sample was limited to patients assessed in a single ED in Scotland and therefore may not represent the views of more diverse populations, particularly in regard to ethnicity. One of the main reasons for nonparticipation was prospective participants’ not returning the screening telephone call. However, despite challenges to recruitment in health-related research, recruitment rates to this study were typical of other clinical trials.29 It is possible that participants agreeing to be involved in this study were patients who were more likely to have ongoing health concerns. It is likely that this study population did not include patients who were fully reassured by the assessment process and did not view themselves as having continuing care needs or unanswered questions because such participants may well have thought that they would gain little from being part of the research process. Conducting interviews 1 week postdischarge could influence the content of participant accounts because of events occurring during the recovery period. However, many patients do not process their illness experience until after discharge from the hospital, with concerns becoming apparent only after the acute event. With patients spending less time in the hospital when assessed with the early rule-out pathway, it was believed important to capture how this period may have been influenced by implementation of the early rule-out pathway. We also acknowledge that previous illness experience could affect how participants interpreted information and events during the chest pain assessment process. For some participants, returning to the hospital for a research interview appeared to be used as a further opportunity for contact with a health care practitioner. Although we acknowledge that the study population may represent those who are more concerned about their heart, it is these very patients who require continued support. Potential biases may also exist in selection of participants and coding, although efforts were made to reduce these potential effects.

re practitioner. Although we acknowledge that the study population may represent those who are more concerned about their heart, it is these very patients who require continued support. Potential biases may also exist in selection of participants and coding, although efforts were made to reduce these potential effects. Initial identification of potential participants was performed by nurses from the ED research team, who were not involved with any other aspect of the study and were unaware of the emerging themes of the research. Participant selection could not therefore be influenced by emerging study data. Additionally, because data collection and analysis occurred concurrently, blinding to study group was not possible. All interviews were conducted by A.V.F., although interview technique was discussed and transcripts were reviewed with an experienced senior colleague (S.C.-B.). The grouping of codes into themes was performed through discussion with S.C.-B., including a review of the patient narratives and coding categories. Throughout the process of data collection and analysis, researcher reflexivity was used to discuss any possible biases.

wed with an experienced senior colleague (S.C.-B.). The grouping of codes into themes was performed through discussion with S.C.-B., including a review of the patient narratives and coding categories. Throughout the process of data collection and analysis, researcher reflexivity was used to discuss any possible biases. Discussion In this study, we explored patient experience of chest pain in 2 groups of patients assessed before and after the implementation of an early rule-out pathway for myocardial infarction. Using individual patient interviews allowed participants to describe their experiences in their own words, thereby providing data that would not be captured with quantitative data collection methods. This study adds to a prospective questionnaire study of patients admitted to a short-stay ward with symptoms of suspected acute coronary syndrome, in which closed questions were posed about the acceptability of an early discharge pathway at patient discharge.30 These methods may not capture the richness of experience, not least because many patients do not process their illness experience until after discharge from the hospital, with concerns becoming apparent only after the acute event.

were posed about the acceptability of an early discharge pathway at patient discharge.30 These methods may not capture the richness of experience, not least because many patients do not process their illness experience until after discharge from the hospital, with concerns becoming apparent only after the acute event. We report 5 major findings. First, it was common for participants to seek help from other health care sources before presentation to the ED. Second, discordance sometimes exists between the objective interpretation of troponin results by clinicians and the ongoing illness episode experienced by patients. Third, pretest information, trust in the clinician, and active listening may enhance reassurance gained from negative test results. These first 3 themes were common to participants assessed with both care pathways. Fourth, other themes appeared to relate to the specific care pathway used; routine care procedures appeared to be a source of frustration for participants requiring admission to the hospital for serial blood sampling. This frustration was not evident with the early rule-out pathway. Fifth, patients assessed with the early rule-out pathway appeared less likely to appraise their future health status; therefore, the rapid rule-out of myocardial infarction in the ED may provide less incentive for patients to use their chest pain presentation as an opportunity to address their future health.

t pathway. Fifth, patients assessed with the early rule-out pathway appeared less likely to appraise their future health status; therefore, the rapid rule-out of myocardial infarction in the ED may provide less incentive for patients to use their chest pain presentation as an opportunity to address their future health. The concepts of discordance and reassurance are linked, with reassurance being more difficult to achieve when discordance is present. Reassurance, diagnosis, explanation, and advice are the main interventions reducing suffering for patients in the ED.31 The need for reassurance, and therefore the use of mechanisms for the development of reassurance, emerged as a key theme in this study. We propose a model (Figure 2) in which communication interventions incorporate the provision of pretest information in regard to troponin testing, active listening, and acknowledgment of the patient’s illness concerns. This should lead to the development of trust between the clinician and patient and may aid patients in considering their illness episode as concluded and therefore feeling reassured.

e provision of pretest information in regard to troponin testing, active listening, and acknowledgment of the patient’s illness concerns. This should lead to the development of trust between the clinician and patient and may aid patients in considering their illness episode as concluded and therefore feeling reassured. The benefit of pretest information in promoting reassurance is supported by several randomized controlled studies.32, 33, 34 In addition, verbal information has been shown to be more effective than written information in providing reassurance from negative exercise test results for patients with chest pain.33 Many studies support the idea that, for patients, a negative test result in itself may not be reassuring.35, 36, 37, 38 Patients may be ill prepared for negative or normal findings; therefore, giving information before initial blood sampling about the meaning and subsequent care procedures related to troponin testing may aid in preparing patients for the idea of direct discharge from the ED.

n itself may not be reassuring.35, 36, 37, 38 Patients may be ill prepared for negative or normal findings; therefore, giving information before initial blood sampling about the meaning and subsequent care procedures related to troponin testing may aid in preparing patients for the idea of direct discharge from the ED. The absence of evidence of active listening was common in accounts of patients who were not reassured by the chest pain assessment process. A shared perspective between the patient and clinician on the cause of symptoms and course of action is positively associated with resolution of symptoms.39 This requires effective communication and recognition of the problem as expressed by the patient. Without this, a patient may believe himself or herself ill equipped to manage ongoing symptoms. Achieving a shared understanding can be difficult because clinicians and patients often understand health and illness through different lenses.40 This is evident by the clinician’s ability to interpret a low troponin concentration and rule-out of myocardial infarction as a conclusion to the illness episode, which can nonetheless remain very current to the patient. The patient-clinician relationship can be strengthened when patients have the opportunity to express their concerns and the clinician shows empathy with and responds to individual circumstances. When these critical points in communication break down, trust is lost and uncertainty for the patient may prevail. It appeared that patients’ trust in their clinician had to be earned through effective communication. Patients interpret reassurance in the context of their own views and illness perceptions; therefore, a key to successful reassurance is the clinician’s ability to identify and acknowledge the patient’s perspectives in regard to his or her symptoms and related concerns. When there is failure to give credence to the patient’s perspective, or the clinician’s view contradicts the patient’s view, reassurance is difficult to achieve.41 When the interventions of pretest information, active listening, and development of trust have been satisfied, patients are more likely to view their illness episode as complete.

to give credence to the patient’s perspective, or the clinician’s view contradicts the patient’s view, reassurance is difficult to achieve.41 When the interventions of pretest information, active listening, and development of trust have been satisfied, patients are more likely to view their illness episode as complete. The development of reassurance could also be hindered by patient interpretation of routine care procedures. The questionnaire study by Hancock and Carlton30 suggested that patients would gain reassurance from being admitted to the hospital, yet the evidence from the study reported here has suggested that admission to the hospital may serve to validate that symptoms may be serious, at least for some patients. Additionally, it was rare for a patient to make the decision to attend the hospital for assessment independently without previous contact with the health service. Prompts by a medical professional allow the patient to negotiate access to care and construct attendance at the ED as an appropriate action. If symptoms have been validated by a health professional in this way, then reassurance from the hospital assessment procedures may be more difficult to achieve.

he health service. Prompts by a medical professional allow the patient to negotiate access to care and construct attendance at the ED as an appropriate action. If symptoms have been validated by a health professional in this way, then reassurance from the hospital assessment procedures may be more difficult to achieve. The risk of implementing early rule-out pathways is that they may focus on process-driven care and outcomes (such as complying with the target to admit, transfer, or discharge patients within 4 hours of attendance at the ED) rather than offering a comprehensive care experience to patients. Although patients may accept that the main aim of an assessment is to exclude a serious illness, care can be considered lacking when it fails to address the patient as a whole. Patients view talking to a professional about their situation as an intervention as important as the delivery of negative test results.16 Clinical history taking represents a “process” of care in the ED. The frustration of repeating the symptom presentation story, which was highly prevalent among patients admitted to the hospital for serial troponin-level testing and appeared infrequently among those assessed with the early rule-out pathway, exemplifies further that it is the active listening by the clinician and not simply the opportunity to tell the story that is important to patients.

highly prevalent among patients admitted to the hospital for serial troponin-level testing and appeared infrequently among those assessed with the early rule-out pathway, exemplifies further that it is the active listening by the clinician and not simply the opportunity to tell the story that is important to patients. A further concern about implementing an early rule-out pathway is that the accelerated assessment process may provide less of an opportunity for patients to consider their future health than previous assessment approaches. An episode of acute chest pain can serve as a cue to action for this patient population (although we acknowledge that interview participants revealed only an intention to act). It is possible that the cue to action in the early rule-out pathway is less persuasive because symptoms are dismissed by the ED clinician much more quickly. These interview data are concordant with previous work suggesting that consideration of future health goals appears to be a reactive rather than a proactive process.42 A clinical consultation may therefore be enriched by harnessing this cue to action and providing a teachable moment to increase perception of personal risk to future ill health. The content of the consultation may shape the perceived threat of disease or belief in the benefit of lifestyle interventions for the patient. The early rule-out pathway, with its focus on the rapid rule-out of myocardial infarction, may not afford the opportunity for this interaction to develop.

to future ill health. The content of the consultation may shape the perceived threat of disease or belief in the benefit of lifestyle interventions for the patient. The early rule-out pathway, with its focus on the rapid rule-out of myocardial infarction, may not afford the opportunity for this interaction to develop. In summary, early rule-out pathways will undoubtedly be of major benefit to health care providers by decreasing unnecessary hospital admissions. Avoiding hospitalization and having fewer health care professionals involved in the assessment process were also viewed positively by patients. The successful implementation of these pathways in the ED will be aided by the addition of simple communication interventions during the chest pain assessment process. Confirmation of the absence of an acute cardiac event may not satisfy the care needs of patients. To enhance the care experience of patients presenting to the ED with symptoms of suspected acute coronary syndrome, focus must remain on the comprehensive assessment and care of the patient, and not solely on the rule-out of myocardial infarction. Supplementary Data Appendix E1 Topic guide Table E1 Additional quotes to illustrate themes. The authors acknowledge the Emergency Medicine Research Group of Edinburgh for their help with identifying potential participants. Please see page 503 for the Editor’s Capsule Summary of this article. Revised October 21, 2019 and October 28, 2019.

Supplementary Data Appendix E1 Topic guide Table E1 Additional quotes to illustrate themes. The authors acknowledge the Emergency Medicine Research Group of Edinburgh for their help with identifying potential participants. Please see page 503 for the Editor’s Capsule Summary of this article. Revised October 21, 2019 and October 28, 2019. Supervising editor: Keith A. Marill, MD, MS. Specific detailed information about possible conflict of interest for individual editors is available at https://www.annemergmed.com/editors. Author contributions: AVF, FES, NLM, and SC-B conceived the study, designed the trial, and obtained research funding. AVF undertook recruitment of participants and managed the data, including quality control, and drafted the article. AVF analyzed the data with oversight from SC-B. All authors contributed substantially to article revision. AVF takes responsibility for the paper as a whole. All authors attest to meeting the four ICMJE.org authorship criteria: (1) Substantial contributions to the conception or design of the work; or the acquisition, analysis, or interpretation of data for the work; AND (2) Drafting the work or revising it critically for important intellectual content; AND (3) Final approval of the version to be published; AND (4) Agreement to be 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.

e work or revising it critically for important intellectual content; AND (3) Final approval of the version to be published; AND (4) Agreement to be 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. Funding and support: By Annals policy, all authors are required to disclose any and all commercial, financial, and other relationships in any way related to the subject of this article as per ICMJE conflict of interest guidelines (see www.icmje.org). The authors have stated that no such relationships exist. Dr. Ferry was supported by an Edinburgh and Lothians Health Foundation grant (153-586). Dr. Lee is supported by a fellowship (FS/18/25/33454) from the British Heart Foundation. Dr. Anand is supported by a Clinical Lectureship from the Chief Scientist Office (PCL/18/05). Dr. Chapman is supported by a fellowship (FS/16/75/32533) from the British Heart Foundation. Dr. Mills is supported by the Butler Senior Research Fellowship (FS/16/14/32023) award from the British Heart Foundation. Dr. Cunningham-Burley is supported by the Wellcome Trust (209519/Z/17/Z). The trial was funded by the BHF through a project grant (PG/15/51/31596) and Research Excellence award (RE/18/5/34216). Trial registration number: NCT03005158 Readers: click on the link to go directly to a survey in which you can provide feedback to Annals on this particular article. A podcast for this article is available at www.annemergmed.com.

Introduction Background Acute respiratory failure is a common but life-threatening medical emergency, especially in elderly patients with respiratory and cardiac diseases.1-3 Inhospital noninvasive ventilation is widely used to treat acute respiratory failure that is refractory to initial medical therapy.4-12 However, the delay in providing noninvasive ventilation until arrival at the hospital may be one factor explaining why the risk of death in patients with respiratory problems increases markedly with distance traveled to the hospital.13 It has been argued that noninvasive ventilation is more likely to be effective if used early in the course of respiratory failure, before fatigue develops.14 Recent reviews have indicated that out-of-hospital noninvasive ventilation is feasible and beneficial in selected patients with acute respiratory failure.15-18 We have undertaken a meta-analysis19 suggesting that out-of-hospital continuous positive airway pressure (CPAP) has the best evidence of effectiveness. CPAP is also the most practical way of providing out-of-hospital noninvasive ventilation.Editor’s Capsule Summary What is already known on this topic Emergency medical services (EMS) often use continuous positive airway pressure (CPAP) to treat out-of-hospital acute respiratory failure. What question this study addressed Is out-of-hospital CPAP cost-effective? What this study adds to our knowledge

Introduction Background Acute respiratory failure is a common but life-threatening medical emergency, especially in elderly patients with respiratory and cardiac diseases.1-3 Inhospital noninvasive ventilation is widely used to treat acute respiratory failure that is refractory to initial medical therapy.4-12 However, the delay in providing noninvasive ventilation until arrival at the hospital may be one factor explaining why the risk of death in patients with respiratory problems increases markedly with distance traveled to the hospital.13 It has been argued that noninvasive ventilation is more likely to be effective if used early in the course of respiratory failure, before fatigue develops.14 Recent reviews have indicated that out-of-hospital noninvasive ventilation is feasible and beneficial in selected patients with acute respiratory failure.15-18 We have undertaken a meta-analysis19 suggesting that out-of-hospital continuous positive airway pressure (CPAP) has the best evidence of effectiveness. CPAP is also the most practical way of providing out-of-hospital noninvasive ventilation.Editor’s Capsule Summary What is already known on this topic Emergency medical services (EMS) often use continuous positive airway pressure (CPAP) to treat out-of-hospital acute respiratory failure. What question this study addressed Is out-of-hospital CPAP cost-effective? What this study adds to our knowledge In this cost-effectiveness analysis using a United Kingdom perspective, the incremental cost-effectiveness ratio (£20,514 [$29,720] per quality-adjusted life-year gained) was insufficient by British standards to support widespread CPAP implementation. Interpretation of cost-effectiveness may vary by country, including the US.

cost-effectiveness analysis using a United Kingdom perspective, the incremental cost-effectiveness ratio (£20,514 [$29,720] per quality-adjusted life-year gained) was insufficient by British standards to support widespread CPAP implementation. Interpretation of cost-effectiveness may vary by country, including the US. How this is relevant to clinical practice Although not directly impacting clinical practice, these findings highlight factors that should be considered in the selection and implementation of EMS therapies. Importance In the United States, the National Association of Emergency Medical Services Physicians stated that noninvasive ventilation is an important treatment modality for the out-of-hospital management of acute dyspnea.20 The recent UK Ambulance Services Clinical Practice Guidelines 201321 recommended (for the first time) the use of CPAP in the out-of-hospital environment on the basis of expert consensus. However, use of out-of-hospital CPAP in the United Kingdom remains limited in practice, probably reflecting the significant costs of establishing this treatment. The decision to establish out-of-hospital CPAP depends on weighing the benefits of improved outcomes against the additional costs incurred by establishing the service and is fundamentally an issue of cost-effectiveness.

mains limited in practice, probably reflecting the significant costs of establishing this treatment. The decision to establish out-of-hospital CPAP depends on weighing the benefits of improved outcomes against the additional costs incurred by establishing the service and is fundamentally an issue of cost-effectiveness. Goals of This Investigation We aimed to estimate the incremental cost per quality-adjusted life-year (QALY) of out-of-hospital CPAP compared with standard care and determine whether out-of-hospital CPAP should be recommended for funding according to accepted thresholds for cost-effectiveness.22

mains limited in practice, probably reflecting the significant costs of establishing this treatment. The decision to establish out-of-hospital CPAP depends on weighing the benefits of improved outcomes against the additional costs incurred by establishing the service and is fundamentally an issue of cost-effectiveness. Goals of This Investigation We aimed to estimate the incremental cost per quality-adjusted life-year (QALY) of out-of-hospital CPAP compared with standard care and determine whether out-of-hospital CPAP should be recommended for funding according to accepted thresholds for cost-effectiveness.22 Materials and Methods Theoretical Model of the Problem The cost-effectiveness of a more effective and expensive treatment, such as out-of-hospital CPAP, can be estimated by comparing the outcomes and costs associated with the treatment to an appropriate alternative, such as inhospital noninvasive ventilation. Modeling is used to estimate how better effectiveness leads to improved patient outcomes relative to the cost of the care, expressed in terms of QALYs gained by using the intervention. A QALY is equivalent to a year of life spent in full health and incorporates both quality of life and survival. If outcomes are measured as QALYs, then the incremental cost-effectiveness ratio, or cost per QALY gained, can be calculated and compared with alternative uses of health care funding. The incremental cost-effectiveness ratio can be crudely understood as the amount needed for the intervention to “buy” each additional QALY compared with standard care. In the United Kingdom, the National Institute for Health and Care Excellence typically recommends in favor of funding interventions with an incremental cost-effectiveness ratio below a threshold, widely accepted to be between £20,000 and £30,000 per QALY (ie, interventions that can be used to buy QALYs for less than £20,000 [$29,000] or £30,000 [$43,500] each) and recommends against funding interventions with an incremental cost-effectiveness ratio above these thresholds.22 Other countries use different thresholds; one of $50,000 is conventionally used in the United States. We used the National Institute for Health and Care Excellence threshold because our analysis took a UK perspective and was based on UK practice and unit costs. Appendix E1 (available online at http://www.annemergmed.com) describes the general principles of cost-effectiveness analysis.

ionally used in the United States. We used the National Institute for Health and Care Excellence threshold because our analysis took a UK perspective and was based on UK practice and unit costs. Appendix E1 (available online at http://www.annemergmed.com) describes the general principles of cost-effectiveness analysis. Study Design We developed an economic model with the statistical software program R23 (version 3.0.3) to explore the costs and health outcomes associated with the use of out-of-hospital CPAP to treat acute respiratory failure compared with standard care, and to calculate the incremental cost per QALY gained. We based the analysis on a UK health care system perspective using a lifetime horizon. We converted the results into US dollars, using the Organisation for Economic Co-operation and Development purchasing power parities rates (£1=$1.45).

ompared with standard care, and to calculate the incremental cost per QALY gained. We based the analysis on a UK health care system perspective using a lifetime horizon. We converted the results into US dollars, using the Organisation for Economic Co-operation and Development purchasing power parities rates (£1=$1.45). The structure of the model is shown in Figure 1 and the parameters used in the model are reported in Table 1. Patients in acute respiratory failure may receive either out-of-hospital CPAP or standard care, so these alternative treatments were applied to a hypothetical cohort of patients with acute respiratory failure who were eligible for noninvasive ventilation, ie, all patients receive out-of-hospital CPAP in the intervention group and standard care in the comparator group. Each patient in the model could have one of the following outcomes: hospitalized without intubation, hospitalized with intubation, or death in the ambulance or hospital. The probability of death and probability of intubation depended on whether the patient received out-of-hospital CPAP or standard care. The patients who survived accrued lifetime QALYs and health care costs according to their life expectancy. Costs were also accrued through costs of intervention (ie, out-of-hospital CPAP) and hospital treatment costs, which depended on whether the patient needed intubation. Details of each of these processes are outlined below.

nts who survived accrued lifetime QALYs and health care costs according to their life expectancy. Costs were also accrued through costs of intervention (ie, out-of-hospital CPAP) and hospital treatment costs, which depended on whether the patient needed intubation. Details of each of these processes are outlined below. We estimated the baseline risks of intubation and death (ie, for patients receiving standard care) with data from published studies.19 We modeled the mortality risk of emergency admissions with respiratory illness with a large cohort data set of 668 patients presenting with “respiratory disease” across 4 English ambulance services during a 4-year period, from 1997 to 2001.13 There were 79 deaths in 668 patients, which resulted in a mean mortality rate of 11.8%. This baseline mortality rate is similar to that in the more recent Three Interventions in Cardiogenic Pulmonary Oedema (3CPO) study.24 We modeled the risk of intubation for respiratory illness with the data from 3CPO study,24 a multicenter, open, prospective, randomized, controlled trial of 1,069 patients presenting with severe acute cardiogenic pulmonary edema at 26 emergency departments (EDs) in the UK. The study reported a mean intubation rate of 2.9%, which is similar to the intubation rates of 2.7% reported in British Thoracic Society national respiratory audit program annual report 2011/12.25

ial of 1,069 patients presenting with severe acute cardiogenic pulmonary edema at 26 emergency departments (EDs) in the UK. The study reported a mean intubation rate of 2.9%, which is similar to the intubation rates of 2.7% reported in British Thoracic Society national respiratory audit program annual report 2011/12.25 We estimated the effectiveness of out-of-hospital CPAP in reducing mortality and intubation as odds ratios (ORs) from a meta-analysis19 of 6 randomized trials, which are summarized in Table 1. The effectiveness estimates from our meta-analysis are similar to those of Williams et al18 but greater than those of Mal et al17 because the latter authors estimated effectiveness for all forms of out-of-hospital noninvasive ventilation together, whereas we estimated effectiveness of out-of-hospital CPAP and bi-level inspiratory positive airway pressure separately and used only the estimate for CPAP.

et al18 but greater than those of Mal et al17 because the latter authors estimated effectiveness for all forms of out-of-hospital noninvasive ventilation together, whereas we estimated effectiveness of out-of-hospital CPAP and bi-level inspiratory positive airway pressure separately and used only the estimate for CPAP. The patients who survived (ie, who avoided the short-term mortality risk) accrued QALYs, and these lifetime QALYs were estimated according to patients’ life expectancy and their utilities. The life expectancy of patients with acute respiratory failure and admitted to the hospital were captured from the 3CPO trial24 and parameterized as a normal distribution with a mean of 2.67 and SD of 0.16, after discussions with a clinical expert group (S.G., M.W., J.P.-A., and G.D.P.). The 3CPO study24 reported that the mean utility value (quality of life) was 0.6. The estimated QALYs for patients with acute respiratory failure were estimated by multiplying the life-years by the lifetime quality of life shown in Table 1. There was no evidence that patients who survived after receiving out-of-hospital CPAP experienced better health-related quality of life or lived longer compared with patients given standard care. We assumed that the lifetime QALYs were same for all survivors, irrespective of whether they were in standard care or the out-of-hospital CPAP arm.

hat patients who survived after receiving out-of-hospital CPAP experienced better health-related quality of life or lived longer compared with patients given standard care. We assumed that the lifetime QALYs were same for all survivors, irrespective of whether they were in standard care or the out-of-hospital CPAP arm. The costs included in the model are for of out-of-hospital CPAP, intubation, hospitalization, and lifetime care for patients. Hubble et al26 reported a mean additional 5 days in length of stay associated with intubation compared with that for patients without intubation. We assumed that the additional 5 hospital days spent by the intubated patients would be in the ICU, according to the suggestions by the clinical expert group, at a cost of £700 ($1,015) per day. We estimated lifetime costs of survivors by using the annual costs and the discounted life expectancy of patients captured from the 3CPO trial.24 The study reported that mean costs in months 4 to 6 were £1,341 ($1,944.50), which resulted in mean annual costs of £5,360 ($7,685). The costs of standard care were not included in the model because the analysis is based on incremental costs, ie, we assumed that all initial treatment costs were the same in both arms, regardless of whether the patient received out-of-hospital CPAP.

($1,944.50), which resulted in mean annual costs of £5,360 ($7,685). The costs of standard care were not included in the model because the analysis is based on incremental costs, ie, we assumed that all initial treatment costs were the same in both arms, regardless of whether the patient received out-of-hospital CPAP. We estimated the costs of out-of-hospital CPAP at an ambulance service level and converted these into a cost per patient according to a 5-year depreciation period. These costs included those for initial equipment, implementation, and ongoing maintenance. Although the costs of setting up the service are largely the same, there are substantially different estimates of incidence reported in different sources.25-29 The cost of out-of-hospital CPAP per patient in an ambulance service is estimated to be £189.93+£202,446/N ($275.50+$293,546/N), where N is the number of patients per ambulance service in a year. This information was synthesized into a mean cost of £1,212 ($1,740) per patient, according to our clinical expert input. Because there are different estimates of incidence, ranging from approximately 175 to 2,000 patients per ambulance service in a year, scenario analysis was also conducted for 3 different estimates for unit cost for performing out-of-hospital CPAP per patient (for different estimates of the eligible population), ie, a high-cost scenario with a unit cost of £1,400 ($2,030), a low-cost scenario with a unit cost of £745 ($1,080), and a lower-cost scenario with a unit cost of £300 ($435). Full details are provided in Appendix E2 (available online at http://www.annemergmed.com).

(for different estimates of the eligible population), ie, a high-cost scenario with a unit cost of £1,400 ($2,030), a low-cost scenario with a unit cost of £745 ($1,080), and a lower-cost scenario with a unit cost of £300 ($435). Full details are provided in Appendix E2 (available online at http://www.annemergmed.com). Primary Data Analysis Probabilistic analysis incorporated uncertainty in the parameter estimates to provide a measure of precision and confidence in the estimates of the mean costs and QALYs. Additionally, we calculated the probability that each strategy would be the most cost-effective at different thresholds for willingness to pay for health gain. We constructed cost-effectiveness acceptability curves by plotting the probability of each strategy’s being cost-effective against willingness to pay. Furthermore, expected value of perfect information was estimated to identify whether the expected cost of future research would be valuable. Expected value of partial perfect information and expected value of sample information techniques were also used to identify the critical areas of uncertainty in which future research would produce the most benefit. Value-of-information analyses (expected value of perfect information, expected value of partial perfect information, and expected value of sample information) provide an estimate of the monetary value of further research to reduce uncertainty and, in particular, an estimate of how much we should be prepared to pay for a trial to reduce uncertainty.

(expected value of perfect information, expected value of partial perfect information, and expected value of sample information) provide an estimate of the monetary value of further research to reduce uncertainty and, in particular, an estimate of how much we should be prepared to pay for a trial to reduce uncertainty. Results The total costs of out-of-hospital CPAP are higher than those of usual care (£16,895 versus £14,863, or $24,497 versus $21,551), but 0.099 QALYs are gained per patient treated (1.513 versus 1.414). The mean incremental cost-effectiveness ratio of out-of-hospital CPAP compared with standard care in the base case analysis is £20,514 per QALY ($29,720/QALY). It therefore costs the health service £20,514 ($29,720) to buy each additional QALY with out-of-hospital CPAP. Figure 2 shows the uncertainty associated with this estimate by plotting samples of mean incremental costs and QALYs. There is substantial uncertainty, with samples falling equally on either side of the red line, indicating the £20,000 per QALY ($29,000/QALY) threshold and a 49.5% probability of out-of-hospital CPAP’s being cost-effective at this threshold.

ed with this estimate by plotting samples of mean incremental costs and QALYs. There is substantial uncertainty, with samples falling equally on either side of the red line, indicating the £20,000 per QALY ($29,000/QALY) threshold and a 49.5% probability of out-of-hospital CPAP’s being cost-effective at this threshold. This is also observed in the cost-effectiveness analysis curve in Figure 3, which shows the proportion of model runs for which each strategy is cost-effective over a range of potential thresholds for willingness to pay. The more we are willing to pay for health gain (ie, the more we are willing to spend to buy a QALY), the more likely it is that out-of-hospital CPAP will be cost-effective, but there is substantial uncertainty between the thresholds of £20,000 per QALY ($29,000/QALY) and £30,000 per QALY ($43,500/QALY).

ingness to pay. The more we are willing to pay for health gain (ie, the more we are willing to spend to buy a QALY), the more likely it is that out-of-hospital CPAP will be cost-effective, but there is substantial uncertainty between the thresholds of £20,000 per QALY ($29,000/QALY) and £30,000 per QALY ($43,500/QALY). Scenario analysis was also conducted for 3 different estimates for unit cost for providing out-of-hospital CPAP per patient, ie, a high-cost scenario with a unit cost of £1,400 ($2,030), a low-cost scenario with a unit cost of £745 ($1,080), and a lower-cost scenario with a unit cost of £300 ($435). These estimates reflect variation in our estimates of the incidence of appropriate patients, with high estimates of incidence resulting in lower estimated costs. Table 2 compares each scenario with the base case and shows that out-of-hospital CPAP is more likely to be cost-effective (93.8% probability) if the incidence of appropriate patients is high and the resulting cost per patient low. Results from threshold analysis suggested that CPAP is unlikely to be cost-effective at £30,000 per QALY ($43,500/QALY) in an ambulance service is greater than if it costs more than £2,170 ($3,150) for out-of-hospital CPAP per patient.

he incidence of appropriate patients is high and the resulting cost per patient low. Results from threshold analysis suggested that CPAP is unlikely to be cost-effective at £30,000 per QALY ($43,500/QALY) in an ambulance service is greater than if it costs more than £2,170 ($3,150) for out-of-hospital CPAP per patient. The incidence of appropriate patients was also an important determinant of the expected value of information. The population expected value of perfect information at the threshold of £20,000 per QALY ($29,000/QALY) is £1.9 million ($2.75 million) at a low estimate of incidence of 3.5 per 100,000 population per year and £22.5 million ($32.5 million) at a higher incidence of 40.8 per 100,000 population per year. This value is defined as the maximum investment a decisionmaker would be willing to pay to eliminate all parameter uncertainty from the decision problem and reflects the amount we should be willing to pay for research to reduce uncertainty. Expected value of partial perfect information analysis suggested baseline mortality, out-of-hospital CPAP mortality OR, and costs of out-of-hospital CPAP as the key parameters driving uncertainty. The population expected value of partial perfect information for the 3 parameters together at the threshold is estimated as £1.83 million ($2.65 million) at the low incidence and £21.3 million ($30.8 million) at the higher one, both of which are very close to the population expected value of perfect information values, suggesting that most of the uncertainty in the decision problem is from these 3 parameters. The population expected value of sample information value for baseline mortality and out-of-hospital CPAP mortality OR parameters, assuming a randomized controlled trial with 100 patients in each arm, is estimated as £1.08 million ($1.56 million) at low incidence and £12.67 million ($18.37 million) at the higher one. It is cost-effective to conduct the trial to address the uncertainty if the population expected value of sample information of a proposed trial at a given sample size is greater than the costs of the trial.

stimated as £1.08 million ($1.56 million) at low incidence and £12.67 million ($18.37 million) at the higher one. It is cost-effective to conduct the trial to address the uncertainty if the population expected value of sample information of a proposed trial at a given sample size is greater than the costs of the trial. Limitations This model is generally based on robust data sources, with the estimates of effectiveness of out-of-hospital CPAP being derived from a meta-analysis of randomized trials and most cost estimates being derived from UK National Health Service reference costs. However, there are some limitations to the data. The trials in the meta-analysis were generally small and had potentially selected study populations, which may not have compared out-of-hospital CPAP with best alternative care. The model parameters were estimated from different sources, which may include different patient populations; for example, the baseline mortality risk was estimated for patients with respiratory disease, whereas the intubation risk was based on data for patients with a diagnosis of severe acute cardiogenic pulmonary edema.

model parameters were estimated from different sources, which may include different patient populations; for example, the baseline mortality risk was estimated for patients with respiratory disease, whereas the intubation risk was based on data for patients with a diagnosis of severe acute cardiogenic pulmonary edema. The perspective of the analysis was the English health service, UK cost estimates were used, key model parameters were estimated from UK sources, and the thresholds for cost-effectiveness were those used in the United Kingdom by the National Institute for Health and Care Excellence. Costs may differ in other countries; for example, the unit cost of the CPAP system may be lower in the United States. The cost-effectiveness of out-of-hospital CPAP may be more certain in health services with different parameters of willingness to pay for health gain. In particular, cost-effectiveness appears to be more certain when compared against a US threshold of $50,000 per QALY. However, our analysis used UK estimates of costs and resource use. If costs and resource use are higher in the United States, the cost-effectiveness will be less certain.

ngness to pay for health gain. In particular, cost-effectiveness appears to be more certain when compared against a US threshold of $50,000 per QALY. However, our analysis used UK estimates of costs and resource use. If costs and resource use are higher in the United States, the cost-effectiveness will be less certain. The cost per patient of providing out-of-hospital CPAP was calculated by dividing the total cost of setting up and running the service by the total number of patients treated, which means that the cost per patient was determined by the incidence of patients who were likely to benefit from out-of-hospital CPAP. We identified a number of sources for our estimate of this parameter, but these estimates varied markedly. Sensitivity analysis showed that cost per patient is an important determinant of cost-effectiveness, so an accurate estimate of the incidence of patients likely to benefit from out-of-hospital CPAP is required to accurately estimate cost-effectiveness.

ur estimate of this parameter, but these estimates varied markedly. Sensitivity analysis showed that cost per patient is an important determinant of cost-effectiveness, so an accurate estimate of the incidence of patients likely to benefit from out-of-hospital CPAP is required to accurately estimate cost-effectiveness. We have assumed in the analysis that the ambulance service has a 1-tiered response. However, some services may have different tiers of response that may allow more efficient use of equipment and trained staff. We assumed that out-of-hospital CPAP had a constant effect on mortality and intubation rate, according to estimates from meta-analysis. Effectiveness may depend on distance traveled to the hospital, being more effective in settings with long distances to the hospital. Unfortunately, distance to the hospital was not consistently collected in primary studies, so this factor could not be explored in the individual patient data meta-analysis.19

a-analysis. Effectiveness may depend on distance traveled to the hospital, being more effective in settings with long distances to the hospital. Unfortunately, distance to the hospital was not consistently collected in primary studies, so this factor could not be explored in the individual patient data meta-analysis.19 Discussion The economic analysis showed that out-of-hospital CPAP was more effective than standard care, with 0.099 QALYs gained per patient treated, but was more expensive, with an additional cost of £2,032 ($2,934) per patient treated. The incremental cost-effectiveness ratio for out-of-hospital CPAP was £20,514 per QALY ($29,720/QALY) compared with standard care, with 49.5% probability of being cost-effective at the £20,000 per QALY threshold. These findings suggest that even if the apparent effectiveness of out-of-hospital CPAP reported by recent meta-analysis17,18 were confirmed, the cost-effectiveness of this treatment is uncertain when compared with a UK cost-effectiveness threshold. It is therefore unlikely that out-of-hospital CPAP would be recommended for widespread implementation in the United Kingdom on the basis of this analysis.

PAP reported by recent meta-analysis17,18 were confirmed, the cost-effectiveness of this treatment is uncertain when compared with a UK cost-effectiveness threshold. It is therefore unlikely that out-of-hospital CPAP would be recommended for widespread implementation in the United Kingdom on the basis of this analysis. In developing the economic model, we identified marked variation between estimates from different sources of the incidence of patients likely to benefit from out-of-hospital CPAP. Sensitivity analysis showed that this parameter was an important determinant of cost-effectiveness, with the probability of out-of-hospital CPAP’s being cost-effective at the £20,000 per QALY ($29,000/QALY) threshold varying from 35.4% to 93.8%. Most of the costs of providing out-of-hospital CPAP are incurred in setting up the service. If only a small number of patients will benefit from out-of-hospital CPAP, then the cost per patient will be high and cost-effectiveness is unlikely.

at the £20,000 per QALY ($29,000/QALY) threshold varying from 35.4% to 93.8%. Most of the costs of providing out-of-hospital CPAP are incurred in setting up the service. If only a small number of patients will benefit from out-of-hospital CPAP, then the cost per patient will be high and cost-effectiveness is unlikely. We identified only 1 previous economic analysis of out-of-hospital CPAP.25 This was undertaken in the United States and estimated that out-of-hospital CPAP would save 0.75 additional lives per 1,000 patients, at a cost of $490 per life saved. This analysis had a number of limitations. Treatment effect estimates were based on trial of inhospital rather than out-of-hospital CPAP, so the analysis effectively compared CPAP with no noninvasive ventilation, rather than comparing out-of-hospital CPAP with inhospital noninvasive ventilation. Outcomes were valued as lives saved rather than QALYs, and the model used only a 1-year time horizon. One-way sensitivity analysis was performed, but the authors did not perform a probabilistic sensitivity analysis. The estimate that out-of-hospital CPAP would be used 4 times per 1,000 patients seems high compared with our estimates of patient eligibility. As a consequence, although this analysis suggested that out-of-hospital CPAP is cost-effective, it is unlikely to convince purchasers of health care.

istic sensitivity analysis. The estimate that out-of-hospital CPAP would be used 4 times per 1,000 patients seems high compared with our estimates of patient eligibility. As a consequence, although this analysis suggested that out-of-hospital CPAP is cost-effective, it is unlikely to convince purchasers of health care. Expected value of information analysis was undertaken to explore uncertainty and determine the value of further research. It showed that the value of undertaking a trial depends on the estimated incidence of eligible patients. The maximum cost at which it would be cost-effective to conduct a trial with 100 patients in each arm is only £1.08 million ($1.56 million) if there were a low estimated incidence (of 3.5 per 100,000 population per year) of eligible patients, but would be £12.67 million ($18.37 million) if there were a high estimated incidence (of 40.8 per 100,000 population per year). A more precise estimate of the incidence of eligible patients is therefore required to determine the cost-effectiveness of a future trial of out-of-hospital CPAP. Of course, the feasibility of a future trial would also depend on the incidence of eligible patients.

ed incidence (of 40.8 per 100,000 population per year). A more precise estimate of the incidence of eligible patients is therefore required to determine the cost-effectiveness of a future trial of out-of-hospital CPAP. Of course, the feasibility of a future trial would also depend on the incidence of eligible patients. Our model can be used to determine whether out-of-hospital CPAP is likely to be cost-effective in a particular system, given the estimate of the incidence of eligible patients. Our analysis indicates that out-of-hospital CPAP has uncertain cost-effectiveness. Establishing out-of-hospital CPAP as a standard treatment of acute respiratory failure will require substantial resources, and there is a relatively high probability that outcomes improvements from out-of-hospital CPAP may not be justified by the financial investment. A large pragmatic randomized trial could improve our estimates of effectiveness and cost-effectiveness, but this would also require substantial funding, with uncertain value. Our analysis suggests that the incidence of patients eligible for out-of-hospital CPAP is an important determinant of the cost-effectiveness of out-of-hospital CPAP itself and of a future trial of it. Health systems considering implementing out-of-hospital CPAP should first estimate the incidence of potentially eligible patients. If the incidence is low, then implementation of out-of-hospital CPAP is not likely to be cost-effective. If the incidence is high, then implementation or further research with a large pragmatic randomized trial may be appropriate. If incidence varies between health care systems, then out-of-hospital CPAP may be cost-effective in some systems but not others.

ation of out-of-hospital CPAP is not likely to be cost-effective. If the incidence is high, then implementation or further research with a large pragmatic randomized trial may be appropriate. If incidence varies between health care systems, then out-of-hospital CPAP may be cost-effective in some systems but not others. Appendix E1 Methods to estimate cost-effectiveness The cost-effectiveness of the different interventions was estimated with both the ICER and the net benefit approaches. Uncertainty was incorporated in the modeling by performing probabilistic sensitivity analysis. Descriptions of these terms and approaches are provided in the sections below.

estimate cost-effectiveness The cost-effectiveness of the different interventions was estimated with both the ICER and the net benefit approaches. Uncertainty was incorporated in the modeling by performing probabilistic sensitivity analysis. Descriptions of these terms and approaches are provided in the sections below. Definitions of Cost-effectiveness Terms The ICER measures the relative value of 2 strategies and is calculated as the mean incremental cost divided by the mean incremental benefits. A strategy is dominated when another strategy accrues more QALYs for less cost. Extended dominance occurs when a combination of 2 alternative strategies can produce the same QALYs as a chosen strategy but at a lower cost. Strategies that are neither dominated nor extendedly dominated constitute the cost-effectiveness frontier, and the ICER is reported for these strategies compared with the next least effective strategy. The willingness-to-pay threshold is the amount of money that the decisionmaker is willing to pay to gain 1 additional QALY. The usual threshold for decisionmaking at the United Kingdom National Institute for Care Excellence is considered to be approximately £20,000 to £30,000 per QALY. The net monetary benefit is defined as the QALYs multiplied by a value for the QALYs (eg, £20,000) minus the costs of obtaining them, that is, net monetary benefit=QALYs×λ–cost, where λ is the willingness-to-pay threshold. The net monetary benefit approach is simpler to calculate and gives equivalent findings (but requires an explicit assumption about the value of λ).

ultiplied by a value for the QALYs (eg, £20,000) minus the costs of obtaining them, that is, net monetary benefit=QALYs×λ–cost, where λ is the willingness-to-pay threshold. The net monetary benefit approach is simpler to calculate and gives equivalent findings (but requires an explicit assumption about the value of λ). Uncertainty Analysis The results presented in the following section include the effects of accounting for uncertainty in the model parameters (the costs, utilities, risks, and ORs for mortality and intubation), characterized as probability distributions. Probabilistic sensitivity analysis is undertaken whereby the model is rerun (1,000 times), each time with a different value for the risks, ORs, costs, and utilities, sampled from the probability distributions. The cost-effectiveness plane shows the incremental costs (y axis) and incremental QALYs (x axis) compared with usual care. In this chart, if a model run for a strategy had exactly the same costs and QALYs as usual care, then the “sample” for that model run would appear at the origin. Samples plotted to the right of the y axis have more QALYs than usual care and samples plotted above the x axis have more costs. Samples plotted to the right of a straight line with slope λ passing through the origin are cost-effective, whereas those plotted to the left are not. The cost-effectiveness acceptability curve shows the proportion of model runs for which each strategy is cost-effective over a range of potential willingness-to-pay thresholds (ie, λ).

to the right of a straight line with slope λ passing through the origin are cost-effective, whereas those plotted to the left are not. The cost-effectiveness acceptability curve shows the proportion of model runs for which each strategy is cost-effective over a range of potential willingness-to-pay thresholds (ie, λ). Value of Information Analysis Another measure of uncertainty is the overall expected value of perfect information (EVPI). This calculation is carried out according to the theory that the decisionmaker will choose the most cost-effective option but could acquire additional evidence to reduce the uncertainties in the decision; for example, know exactly what the ORs for mortality and hospitalizations are for each treatment. In the EVPI calculation, it can be estimated how often making the decision based on current evidence could be wrong, and also how many QALYs (and costs) would be lost by choosing the strategy that is expected to be most cost-effective, given current evidence, when in fact one of the other strategies is truly the most cost-effective. This can be multiplied by the number of patients per year and the expected lifetime of the decision to estimate the population EVPI. The interpretation of population EVPI is that if one could fund research to eliminate the uncertainty in effectiveness for all of the parameters for each strategy (eg, by a large or infinitely large clinical trial), then the value of eliminating the uncertainty through such research would be expected to be the population EVPI. This can be thought of as the maximum that the health care system should be willing to pay for additional evidence to inform the decision in the future and thus is an upper bound on the value of conducting further research, ie, if the population EVPI exceeds the expected costs of additional research, then it is potentially cost-effective to conduct further research.

ealth care system should be willing to pay for additional evidence to inform the decision in the future and thus is an upper bound on the value of conducting further research, ie, if the population EVPI exceeds the expected costs of additional research, then it is potentially cost-effective to conduct further research. Expected value of partial perfect information (EVPPI) is similar to EVPI, but instead of evaluating the uncertainty associated with all parameters it focuses on the uncertainty associated with a subset of one or more parameters, allowing the decisionmakers to conclude in which variables further research would be most beneficial. The computational time required for EVPPI is markedly more than for EVPI because the process essentially requires 2 iterations of probabilistic analyses, as standard probabilistic sensitivity analyses are undertaken for each sampled parameter value for the variable(s) under analysis. If the population EVPPI for a subset of parameters exceeds the expected costs of additional research, then it is potentially cost-effective to conduct further research to estimate those parameters.

ard probabilistic sensitivity analyses are undertaken for each sampled parameter value for the variable(s) under analysis. If the population EVPPI for a subset of parameters exceeds the expected costs of additional research, then it is potentially cost-effective to conduct further research to estimate those parameters. Expected value of sample information (EVSI) addresses the limitation of EVPI, which assumes parameters can be ascertained without uncertainty (ie, effectively assuming an infinite trial size), and seeks to provide an optimal number of patients to study within a future trial. In addition, EVSI also allows the evaluation of marginal returns associated with increased sample size formally taking uncertainty into account (eg, say that an additional 100 patients when only 500 have been recruited would be likely to provide more value than when 20,000 have been recruited). Within EVSI, the costs of the trial are compared with the benefits achieved to find the maximum expected net benefit, which would correspond with the recommended trial size. If the population EVSI of a proposed trial is greater than the costs of the trial, then it is cost-effective to conduct the trial to address the uncertainty.

, the costs of the trial are compared with the benefits achieved to find the maximum expected net benefit, which would correspond with the recommended trial size. If the population EVSI of a proposed trial is greater than the costs of the trial, then it is cost-effective to conduct the trial to address the uncertainty. Appendix E2 Costs of out-of-hospital CPAP There are a number of costs involved in providing out-of-hospital CPAP, such as initial equipment costs, implementation costs, and ongoing maintenance costs. These costs were converted into a cost per patient according to a 5-year depreciation period (ie, assuming that new out-of-hospital noninvasive ventilation equipment will be required in 5 years) and sharing the overall costs out among the number of patients who would benefit from the service during this period. Total Costs of Out-of-Hospital CPAP to the Ambulance Service The costs were often missing from the studies in literature, and thus bottom-up costing methods were used to estimate the costs of out-of-hospital CPAP. The breakdown of the costs for out-of-hospital CPAP is shown in Table E1 and is split into 3 main components:• initial costs of the out-of-hospital CPAP devices • setup/implementation costs, ie, staff training costs and service reconfiguration costs • maintenance costs of the service, ie, consumables and depreciation

Total Costs of Out-of-Hospital CPAP to the Ambulance Service The costs were often missing from the studies in literature, and thus bottom-up costing methods were used to estimate the costs of out-of-hospital CPAP. The breakdown of the costs for out-of-hospital CPAP is shown in Table E1 and is split into 3 main components:• initial costs of the out-of-hospital CPAP devices • setup/implementation costs, ie, staff training costs and service reconfiguration costs • maintenance costs of the service, ie, consumables and depreciation The costs of the out-of-hospital CPAP devices were elicited from the expert advisory group. The costs of implementation on provider organizations were estimated with bottom-up costing methods, assuming a typical ambulance service. The maintenance costs were estimated with activity-based costing for the resources spent on consumables, according to evidence from the literature.

d from the expert advisory group. The costs of implementation on provider organizations were estimated with bottom-up costing methods, assuming a typical ambulance service. The maintenance costs were estimated with activity-based costing for the resources spent on consumables, according to evidence from the literature. The out-of-hospital CPAP device can take different levels of complexity, and the cost of the device is based on this complexity. For example, the costs are different for the noninvasive positive pressure ventilation devices and CPAP/bi-level inspiratory positive airway pressure devices. Furthermore, the costs also depend on whether the devices use a cylinder or whether they are electrically or mechanically powered. The costs of the device were elicited from the expert advisory group, assuming a close-fitting face mask CPAP device with Boussignac CPAP system manufactured by Vygon as the representative of a typical CPAP system. The Boussignac Hospital CPAP kit costs £513.49 ($744.5) and contains the equipment required to deliver out-of-hospital CPAP. We assumed that each ambulance would have the equipment and 10% would need replacing during the 5-year period.

ac CPAP system manufactured by Vygon as the representative of a typical CPAP system. The Boussignac Hospital CPAP kit costs £513.49 ($744.5) and contains the equipment required to deliver out-of-hospital CPAP. We assumed that each ambulance would have the equipment and 10% would need replacing during the 5-year period. The costs of implementation on provider organizations was estimated with bottom-up costing methods, assuming a typical ambulance service, according to the mean size of National Health Service ambulance services in the United Kingdom. It was assumed that a typical ambulance service would have an average capacity of 1,500 paramedics, which was deemed sensible by the expert advisory group. They also suggested that an average of 2 days needed to be dedicated for paramedic training. The costs associated with training were estimated by multiplying this paramedic time with their daily rate according to the Personal and Social Services Research Unit (PSSRU 2012).6 The daily cost per working day was estimated as £150 ($217.5), assuming an average salary of £40,000 ($58,000), including overheads if they are in band 6/7, on suggestion by the clinical advisory group. Service reconfiguration costs were estimated as a 1-off cost of £100,000 ($145,000), and this included the cost of developing new guidelines and pathways. Installation costs were assumed to be zero because the CPAP set under consideration is a disposable system.

n band 6/7, on suggestion by the clinical advisory group. Service reconfiguration costs were estimated as a 1-off cost of £100,000 ($145,000), and this included the cost of developing new guidelines and pathways. Installation costs were assumed to be zero because the CPAP set under consideration is a disposable system. The maintenance costs were estimated by using costing for the resources spent on consumables according to information from the manufacturers that the facial mask, oxygen tubing, and valve (costing £189.93 [$275]) would need to be replaced after each use, whereas the rest of the equipment was reusable. The expert advisory group suggested that an average of an additional day halfway through the 5-year period would be required by the paramedics for updating their training.Table E1 Breakdown of out-of-hospital CPAP costs.

3 [$275]) would need to be replaced after each use, whereas the rest of the equipment was reusable. The expert advisory group suggested that an average of an additional day halfway through the 5-year period would be required by the paramedics for updating their training.Table E1 Breakdown of out-of-hospital CPAP costs. Number of Devices Source Unit Cost, £ ($) Source Total Cost, £ ($) Device costs Out-of-hospital CPAP device Number of ambulances that need the CPAP device (420) Expert advisory input 513.49 (744.50) Vygon: hospital CPAP kit 513.49×420 Assuming 10% new CPAP devices during 5-y usage (42) Expert advisory input 513.49 (744.50) Vygon: hospital CPAP kit 513.49×42 Total cost of the device 237,232 Setup/implementation costs Resource Usage Initial training 1,500 paramedics for 2 days each Expert advisory input 150 (217.50) per day Expert advisory input 450,000 (652,500) Service reconfiguration 1-off cost for reconfiguration Expert advisory input 100,000 (145,000) Total setup/implementation costs 550,000 (797,500) Maintenance costs Resource Usage Consumables Number of patients during 5 y=5×N Expert advisory input 189.93 (275) per use Vygon: facial mask, oxygen tubing, and valve 189.93×5×N (275×5×N) Ongoing training 1,500 paramedics for 1 day each Expert advisory input 150 (217.50) per day Expert advisory input 225,000 (326,500) Total maintenance costs 225,000+949.65×N (326,500+1,375×N) Total costs of out-of-hospital CPAP 1,012,232+949.65×N Total number of patients (N patients per year times depreciation period of 5 y (ie, assuming new out-of-hospital CPAP equipment will be required in 5 y) 5×N Cost of out-of-hospital CPAP per patient 189.93+202,446/N (275.50+293,546/N)

,000+949.65×N (326,500+1,375×N) Total costs of out-of-hospital CPAP 1,012,232+949.65×N Total number of patients (N patients per year times depreciation period of 5 y (ie, assuming new out-of-hospital CPAP equipment will be required in 5 y) 5×N Cost of out-of-hospital CPAP per patient 189.93+202,446/N (275.50+293,546/N) Number of Patients Receiving Out-of-Hospital CPAP in a Typical Ambulance Service The incidence of patients who will benefit from out-of-hospital CPAP is one of the key parameters in the model because the unit cost of out-of-hospital CPAP is estimated by dividing the total costs of out-of-hospital CPAP to the ambulance service by the number of patients treated. There are different estimates of incidence reported in different sources, as reported below, and these are synthesized to achieve a distribution for the costs of out-of-hospital CPAP. Spijker et al1 reported that 16 patients received out-of-hospital CPAP across 11 months in an ambulance service covering a population of 500,000, which amounts into 3.5 potentially eligible cases per 100,000/year. This study identified only patients with acute cardiogenic pulmonary edema, and many eligible patients did not receive treatment (which admittedly may reflect real life); hence, this could be an underestimate of true incidence. Similarly, Aguilar et al2 reported that 175 patients received out-of-hospital noninvasive ventilation across 22 months in an ambulance service covering a population of 1.3 million, which amounts to 7.3 potentially eligible cases per 100,000/year.

lect real life); hence, this could be an underestimate of true incidence. Similarly, Aguilar et al2 reported that 175 patients received out-of-hospital noninvasive ventilation across 22 months in an ambulance service covering a population of 1.3 million, which amounts to 7.3 potentially eligible cases per 100,000/year. Luhr et al3 estimated 77.6 cases of acute respiratory failure per 100,000/year population. Of these, 13.7% were due to chronic obstructive pulmonary disease and 9.2% were due to cardiogenic pulmonary edema (ie, cases with potential to benefit from out-of-hospital noninvasive ventilation). Thus, the incidence can be estimated as 17.8 potentially eligible cases per 100,000/year. However, they are relatively old data and include patients who developed acute respiratory failure in the hospital, so they may be an overestimate. The British Thoracic Society National Respiratory Audit Programme 2011/124 reported that 130 hospitals submitted data on 2,490 patients with noninvasive ventilation between February 1 and March 31, 2012 (ie, 2 months). This amounts to 19.15 (2,490/130) NIV patients per hospital per month, which in yearly terms is 115 per hospital. There are 168 acute hospitals in England serving a population of 53.01 million, which yields an incidence of 36.4 per 100,000 population. However, the details of the audit were not clear and may be subject to bias. Furthermore, it included patients who developed acute respiratory failure in the hospital so may be an overestimate.

68 acute hospitals in England serving a population of 53.01 million, which yields an incidence of 36.4 per 100,000 population. However, the details of the audit were not clear and may be subject to bias. Furthermore, it included patients who developed acute respiratory failure in the hospital so may be an overestimate. In the Sheffield Teaching Hospital ED, there were 255 sets of noninvasive ventilation equipment used during 1 year. This hospital serves a population of 551,800, which amounts to 46.2 potentially eligible cases per 100,000/year. However, the equipment may not actually have been used for patient care, or multiple pieces of equipment may have been used for the same patient, so this is likely to be an overestimate. Hubble et al5 estimated that there were 4 per 1,000 patients transported by ambulance who were eligible for noninvasive ventilation. In 2011 to 2012, there were 4.53 million emergency ambulance transfers to a Level I or II ED in England (population 53.01 million). If 4 per 1,000 of these were eligible, this suggests an incidence of 34.2 per 100,000 population.

per 1,000 patients transported by ambulance who were eligible for noninvasive ventilation. In 2011 to 2012, there were 4.53 million emergency ambulance transfers to a Level I or II ED in England (population 53.01 million). If 4 per 1,000 of these were eligible, this suggests an incidence of 34.2 per 100,000 population. Scenarios for Costs of Out-of-Hospital CPAP to the Ambulance Service As stated above, there are substantially different estimates of incidence reported in different sources, and they are summarized in Table E2. A typical ambulance service caters to a population of approximately 5 million, which suggests a range from approximately 175 to 2,000 patients per ambulance service in a year, depending on the estimate of the incidence. Thus, scenario analysis was conducted by estimating the unit cost for providing out-of-hospital CPAP for these different estimates of the eligible population. The unit costs for the different scenarios are presented in Table E3. This information was synthesized into an expression for the out-of-hospital CPAP costs as £1,500 to £1,000×β (2, 5), shown in Figure E1. This was chosen because our clinical experts believed that most of the samples of costs will fall between £1,400 ($2,030) and £800 ($1,160), with only a few instances in which the costs are lower than £800 ($1,160).Table E2 Scenarios for unit costs of out-of-hospital CPAP.

£1,500 to £1,000×β (2, 5), shown in Figure E1. This was chosen because our clinical experts believed that most of the samples of costs will fall between £1,400 ($2,030) and £800 ($1,160), with only a few instances in which the costs are lower than £800 ($1,160).Table E2 Scenarios for unit costs of out-of-hospital CPAP. Source Incidence of Eligible Patients per 100,000 Annual Eligible Patients in an Ambulance Service Unit Cost of Out-of-Hospital CPAP, £ ($)∗ Spijker et al1 3.5 175 1,346.76 (1,952.80) Aguilar et al2 7.3 365 744.58 (1,096.64) Luhr et al3 17.8 890 417.40 (605.20) Hubble et al5 34.2 1,700 309.02 (448.08) BTS audit4 36.1 1,800 302.40 (438.48) STH ED data 40.8 2,000 291.15 (422.15) BTS, British Thoracic Society; STH, Sheffield Teaching Hospital. ∗ Using the formula unit cost=£189.93+£202,446/N ($275.5+$293,546/N), where N is the number of patients per year. Figure E1 Histogram of the distribution of CPAP costs. High-Cost Scenario Scenario analysis was conducted with a different estimate for unit cost for performing out-of-hospital CPAP per patient (assuming 175 patients in the ambulance service as the eligible population for CPAP), a high-cost scenario with a unit cost of £1,400 ($2,030). The distribution used is normal (£1,400, £100), ie, normal ($2,030, $145).

was conducted with a different estimate for unit cost for performing out-of-hospital CPAP per patient (assuming 175 patients in the ambulance service as the eligible population for CPAP), a high-cost scenario with a unit cost of £1,400 ($2,030). The distribution used is normal (£1,400, £100), ie, normal ($2,030, $145). Low-Cost Scenario Scenario analysis was also conducted with a different estimate for unit cost for performing out-of-hospital CPAP per patient (assuming 365 patients in the ambulance service as the eligible population for CPAP), a low-cost scenario with a unit cost of £745 ($1,080). The distribution used is normal (£745, £100), ie, normal ($1,080, $145). Lower-Cost Scenario Scenario analysis was also conducted with a different estimate for unit cost for performing out-of-hospital CPAP per patient (assuming 1,700 to 2,000 patients in the ambulance service as the eligible population for CPAP), a low-cost scenario with a unit cost of £300 ($435). The distribution used is normal (£300, £50), ie, normal ($435, $72.50).Table E3 Summary of costs. Scenario Mean Value, £ ($) Distribution, £ ($) Baseline 1,212 (1,740) 1,500–1,000×β (2, 5) High cost 1,400 (2,030) Normal (1,400, 100) [normal (2,030, 145)] Low cost 745 (1,080) Normal (745, 100) [normal (1,080, 145)] Lower cost 300 (435) Normal (300, 50) [normal (435, 72.50)] Economic evaluation∗

Lower-Cost Scenario Scenario analysis was also conducted with a different estimate for unit cost for performing out-of-hospital CPAP per patient (assuming 1,700 to 2,000 patients in the ambulance service as the eligible population for CPAP), a low-cost scenario with a unit cost of £300 ($435). The distribution used is normal (£300, £50), ie, normal ($435, $72.50).Table E3 Summary of costs. Scenario Mean Value, £ ($) Distribution, £ ($) Baseline 1,212 (1,740) 1,500–1,000×β (2, 5) High cost 1,400 (2,030) Normal (1,400, 100) [normal (2,030, 145)] Low cost 745 (1,080) Normal (745, 100) [normal (1,080, 145)] Lower cost 300 (435) Normal (300, 50) [normal (435, 72.50)] Economic evaluation∗ Section/Item Item No. Recommendation Reported on Page No. Title and abstract Title 1 Identify the study as an economic evaluation or use more specific terms such as “cost-effectiveness analysis,” and describe the interventions compared. 1 Abstract 2 Provide a structured summary of objectives, perspective, setting, methods (including study design and inputs), results (including base case and uncertainty analyses), and conclusions. 1 Introduction Background and objectives 3 Provide an explicit statement of the broader context for the study. Present the study question and its relevance for health policy or practice decisions. 2 Methods Target population and subgroups 4 Describe characteristics of the base-case population and subgroups analyzed, including why they were chosen. 2 Setting and location 5 State relevant aspects of the system(s) in which the decision(s) need(s) to be made. 3 Study perspective 6 Describe the perspective of the study and relate this to the costs being evaluated. 3 Comparators 7 Describe the interventions or strategies being compared and state why they were chosen. 3 Time horizon 8 State the time horizon(s) over which costs and consequences are being evaluated and say why appropriate. 3 Discount rate 9 Report the choice of discount rate(s) used for costs and outcomes and say why appropriate. 4 Choice of health outcomes 10 Describe what outcomes were used as the measure(s) of benefit in the evaluation and their relevance for the type of analysis performed. 3 Measurement of effectiveness 11a Single-study-based estimates: Describe fully the design features of the single effectiveness study and why the single study was a sufficient source of clinical effectiveness data. 4 Measurement and valuation of preference-based outcome 12 If applicable, describe the population and methods used to elicit preferences for outcomes. NA Estimating resources and costs 13a Single-study-based economic evaluation: Describe approaches used to estimate resource use associated with the alternative interventions. Describe primary or secondary research methods for valuing each resource item in terms of its unit cost.

used to elicit preferences for outcomes. NA Estimating resources and costs 13a Single-study-based economic evaluation: Describe approaches used to estimate resource use associated with the alternative interventions. Describe primary or secondary research methods for valuing each resource item in terms of its unit cost. Describe any adjustments made to approximate to opportunity costs. 4, 5 Appendix Currency, price date, and conversion 14 Report the dates of the estimated resource quantities and unit costs. Describe methods for adjusting estimated unit costs to the year of reported costs if necessary. Describe methods for converting costs into a common currency base and the exchange rate. 3, 4, 5 Choice of model 15 Describe and give reasons for the specific type of decision analytical model used. Providing a figure to show model structure is strongly recommended. 3 Assumptions 16 Describe all structural or other assumptions underpinning the decision-analytical model. 3, 4, 5 Analytical methods 17 Describe all analytical methods supporting the evaluation. This could include methods for dealing with skewed, missing, or censored data; extrapolation methods; methods for pooling data; approaches to validate or make adjustments (such as half cycle corrections) to a model; and methods for handling population heterogeneity and uncertainty. 3, 4, 5 Appendix Results Study parameters 18 Report the values, ranges, references, and, if used, probability distributions for all parameters. Report reasons or sources for distributions used to represent uncertainty where appropriate. Providing a table to show the input values is strongly recommended. Table 1 Incremental costs and outcomes 19 For each intervention, report mean values for the main categories of estimated costs and outcomes of interest, as well as mean differences between the comparator groups. If applicable, report ICERs. Table 2 Characterizing uncertainty 20a Single-study-based economic evaluation: Describe the effects of sampling uncertainty for the estimated incremental cost and incremental effectiveness parameters, together with the impact of methodological assumptions (such as discount rate and study perspective). Figures 2, 3 Characterizing heterogeneity 21 If applicable, report differences in costs, outcomes, or cost-effectiveness that can be explained by variations between subgroups of patients with different baseline characteristics or other observed variability in effects that are not reducible by more information.

ective). Figures 2, 3 Characterizing heterogeneity 21 If applicable, report differences in costs, outcomes, or cost-effectiveness that can be explained by variations between subgroups of patients with different baseline characteristics or other observed variability in effects that are not reducible by more information. Table 2 Discussion Study findings, limitations, generalizability, and current knowledge 22 Summarize key study findings and describe how they support the conclusions reached. Discuss limitations and the generalizability of the findings and how the findings fit with current knowledge. s 6,7 Other Source of funding 23 Describe how the study was funded and the role of the funder in the identification, design, conduct, and reporting of the analysis. Describe other nonmonetary sources of support. tbc Conflicts of interest 24 Describe any potential for conflict of interest of study contributors in accordance with journal policy. In the absence of a journal policy, we recommend authors comply with International Committee of Medical Journal Editors recommendations. tbc ∗ Husereau D, Drummond M, Petrou S, et al. Consolidated Health Economic Evaluation Reporting Standards (CHEERS)—explanation and elaboration: a report of the ISPOR Health Economic Evaluations Publication Guidelines Good Reporting Practices Task Force. Value Health. 2013;16:231-250. The authors acknowledge John Stevens, PhD, and Shiije Ren, PhD, for providing meta-analysis data and Kathryn MacKellar for clerical assistance. Please see page 557 for the Editor’s Capsule Summary of this article.

Table 2 Discussion Study findings, limitations, generalizability, and current knowledge 22 Summarize key study findings and describe how they support the conclusions reached. Discuss limitations and the generalizability of the findings and how the findings fit with current knowledge. s 6,7 Other Source of funding 23 Describe how the study was funded and the role of the funder in the identification, design, conduct, and reporting of the analysis. Describe other nonmonetary sources of support. tbc Conflicts of interest 24 Describe any potential for conflict of interest of study contributors in accordance with journal policy. In the absence of a journal policy, we recommend authors comply with International Committee of Medical Journal Editors recommendations. tbc ∗ Husereau D, Drummond M, Petrou S, et al. Consolidated Health Economic Evaluation Reporting Standards (CHEERS)—explanation and elaboration: a report of the ISPOR Health Economic Evaluations Publication Guidelines Good Reporting Practices Task Force. Value Health. 2013;16:231-250. The authors acknowledge John Stevens, PhD, and Shiije Ren, PhD, for providing meta-analysis data and Kathryn MacKellar for clerical assistance. Please see page 557 for the Editor’s Capsule Summary of this article. Supervising editor: Henry E. Wang, MD, MS

Table 2 Discussion Study findings, limitations, generalizability, and current knowledge 22 Summarize key study findings and describe how they support the conclusions reached. Discuss limitations and the generalizability of the findings and how the findings fit with current knowledge. s 6,7 Other Source of funding 23 Describe how the study was funded and the role of the funder in the identification, design, conduct, and reporting of the analysis. Describe other nonmonetary sources of support. tbc Conflicts of interest 24 Describe any potential for conflict of interest of study contributors in accordance with journal policy. In the absence of a journal policy, we recommend authors comply with International Committee of Medical Journal Editors recommendations. tbc ∗ Husereau D, Drummond M, Petrou S, et al. Consolidated Health Economic Evaluation Reporting Standards (CHEERS)—explanation and elaboration: a report of the ISPOR Health Economic Evaluations Publication Guidelines Good Reporting Practices Task Force. Value Health. 2013;16:231-250. The authors acknowledge John Stevens, PhD, and Shiije Ren, PhD, for providing meta-analysis data and Kathryn MacKellar for clerical assistance. Please see page 557 for the Editor’s Capsule Summary of this article. Supervising editor: Henry E. Wang, MD, MS Author contributions: SG conceived the original idea for the study. PT and SG designed the study. PT developed the model. SG, MW, JP-A, and GDP provided clinical expertise in model development. All authors interpreted the results. PT wrote the first draft of the article. All authors contributed to redrafting and approved the final version of the article. PT takes responsibility for the paper as a whole.

the study. PT developed the model. SG, MW, JP-A, and GDP provided clinical expertise in model development. All authors interpreted the results. PT wrote the first draft of the article. All authors contributed to redrafting and approved the final version of the article. PT takes responsibility for the paper as a whole. Funding and support: By Annals policy, all authors are required to disclose any and all commercial, financial, and other relationships in any way related to the subject of this article as per ICMJE conflict of interest guidelines (see www.icmje.org). The authors have stated that no such relationships exist and provided the following details: This study was funded by the United Kingdom National Institute for Health Research Health Technology Assessment (NIHR HTA) Programme (project 11/36/09). The views expressed in this article are those of the authors and not necessarily those of the NIHR HTA Programme. Any errors are the responsibility of the authors. The funders had no role in the study design; collection, analysis, and interpretation of data; writing of the article; or decision to submit the article for publication. A feedback survey is available with each research article published on the Web at www.annemergmed.com. A podcast for this article is available at www.annemergmed.com. Figure 1 Model structure. CPAP, Continuous positive airway pressure. Figure 2 Cost-effectiveness plane for the base-case economic analysis. ICER, Incremental cost-effectiveness ratio; QALY, quality-adjusted life-years.

A feedback survey is available with each research article published on the Web at www.annemergmed.com. A podcast for this article is available at www.annemergmed.com. Figure 1 Model structure. CPAP, Continuous positive airway pressure. Figure 2 Cost-effectiveness plane for the base-case economic analysis. ICER, Incremental cost-effectiveness ratio; QALY, quality-adjusted life-years. Figure 3 Cost-effectiveness acceptability curve for the base-case economic analysis. Table 1 Summary of model parameters.∗ Parameter Mean Distribution Source Baseline risks General population mean 30-day mortality probability 0.118 Beta (79, 589) Nicholl et al13 Risk of intubation 0.029 Beta (4.45, 150) 3CPO,24 clinical opinion OR for out-of-hospital CPAP Mortality 0.43 Samples Meta-analysis19 Intubation 0.32 Samples Meta-analysis19 Life expectancy of patients Lifetime years 2.67 Normal (2.67, 0.16) 3CPO,24 clinical opinion Health-related quality of life Utility 0.6 Beta (640, 425) 3CPO,24 clinical opinion Costs, £ ($) Out-of-hospital CPAP 1,212 (1,740) 1,500–1,000×β (2, 5) Clinical input Hospitalization 2,250 (3,260) Gamma (75, 30) NHS reference costs30 Intubation 3,500 (5,075) Gamma (70, 50) NHS reference costs 2011–201230 Annual 5,360 (7,685) Gamma (53, 100) NHS reference costs 2011–201230 3CPO, Three Interventions in Cardiogenic Pulmonary Oedema24; NHS, National Health Service.

(2, 5) Clinical input Hospitalization 2,250 (3,260) Gamma (75, 30) NHS reference costs30 Intubation 3,500 (5,075) Gamma (70, 50) NHS reference costs 2011–201230 Annual 5,360 (7,685) Gamma (53, 100) NHS reference costs 2011–201230 3CPO, Three Interventions in Cardiogenic Pulmonary Oedema24; NHS, National Health Service. ∗ Beta (a,b) Distribution is a statistical distribution defined between 0 and 1; a and b parameters in the Beta distribution can be thought of as counts of the event of interest versus its complement, eg, Beta (79,589) for mortality represents 79 deaths in a population of 668 (ie, 79+589). Normal distribution is represented with mean and SD, with 95% of the values in the distribution lying between 2 SDs on either side of the mean, eg, normal (2.67, 0.16) implies that 95% of the samples lie between 2.35 and 2.99. Gamma (a,b) Distribution, where a is the shape parameter and b is the scale parameter, is typically used for skewed distributions and has a mean expected value of a×b, eg, the average value of the samples of distribution Gamma (75,30) is 2,250 (75×30). Table 2 Results for different cost scenarios.

∗ Beta (a,b) Distribution is a statistical distribution defined between 0 and 1; a and b parameters in the Beta distribution can be thought of as counts of the event of interest versus its complement, eg, Beta (79,589) for mortality represents 79 deaths in a population of 668 (ie, 79+589). Normal distribution is represented with mean and SD, with 95% of the values in the distribution lying between 2 SDs on either side of the mean, eg, normal (2.67, 0.16) implies that 95% of the samples lie between 2.35 and 2.99. Gamma (a,b) Distribution, where a is the shape parameter and b is the scale parameter, is typically used for skewed distributions and has a mean expected value of a×b, eg, the average value of the samples of distribution Gamma (75,30) is 2,250 (75×30). Table 2 Results for different cost scenarios. Scenario Type, £ ($) Standard Care Out-of-Hospital CPAP Differences Between Out-of-Hospital CPAP and Standard Care ICER (per QALY), £ ($) Probability of being Cost-effective Total Costs, £ ($) Total QALYs Total Costs, £ ($) Total QALYs Costs, £ ($) QALYs Base case 14,863 (21,551) 1.414 16,895 (24,498) 1.513 2,032 (2,946) 0.099 20,514 (29,720) 0.495 High cost, 1,400 (2,030)∗ 14,863 (21,551) 1.414 17,078 (24,763) 1.513 2,216 (3,213) 0.099 22,368 (32,434) 0.354 Low cost, 745 (1,080)∗ 14,863 (21,551) 1.414 16,421 (23,810) 1.513 1,558 (2,259) 0.099 15,728 (22,805) 0.798 Lower cost, 300 (435)∗ 14,863 (21,551) 1.414 15,977 (23,166) 1.513 1,114 (1,615) 0.099 11,248 (16,309) 0.938 ∗ Scenario analysis was conducted for different estimates for unit cost for performing out-of-hospital CPAP per patient (for different estimates of the eligible population). See Appendix E2 (available online at http://www.annemergmed.com) for more details.

significantly more acetaminophen than the indomethacin group as an adjunct for pain relief. Prednisolone supplemented with acetaminophen may be as effective as indomethacin in relieving pain. Patients taking prednisolone experienced far fewer serious or other adverse effects compared with patients taking indomethacin. Although NSAIDs have been recommended as the first-line therapy for acute gouty arthritis,25 there have been some studies on the role of steroid in the treatment of the condition. In a small preliminary study, it was suggested that a short course of oral corticosteroid therapy could be used effectively for acute gout when NSAIDs are contraindicated.16 Subsequently, in another small study of 27 patients, single intramuscular injections of betamethasone, intravenous methylprednisolone, or oral diclofenac resulted in prompt and equal improvement.26 Glucocorticoid therapy was well tolerated. The sample size was small, adverse effects were not reported, and oral corticosteroids were not used. In another small study (27 patients), the safety and effectiveness of intramuscular triamcinolone in the treatment of acute gout were also noted.27

pt and equal improvement.26 Glucocorticoid therapy was well tolerated. The sample size was small, adverse effects were not reported, and oral corticosteroids were not used. In another small study (27 patients), the safety and effectiveness of intramuscular triamcinolone in the treatment of acute gout were also noted.27 A review article by Kim et al1 reported that therapies available for managing acute gout, which included corticotropin, corticosteroids, colchicines, and NSAIDs, were associated with significant adverse events. However, in our study, we found that there were no cases of upper gastrointestinal bleeding in the group treated with prednisolone, whereas 5 (11%) patients in the indomethacin group developed upper gastrointestinal bleeding. Although none of the patients died because of the adverse effects, the occurrence of upper gastrointestinal bleeding, especially in the elderly, is a cause of significant morbidity.9, 10, 11, 12

Introduction Background Acute gouty arthritis is a crystal-induced inflammation of the joint that primarily affects middle-aged and elderly adults. It is the commonest cause of inflammatory joint disease in men older than 40 years.1 It has been estimated that the overall prevalence in the United Kingdom is 10 per 1,000, with men affected more commonly than women.2 The diagnosis of acute gout is typically made clinically by the presence of rapid onset of severe pain, swelling, and erythema of an affected joint and definitively by demonstrating the presence of negatively birefringent uric acid crystals in aspirated joint fluid. The diagnosis is often made in the absence of joint aspiration because this procedure is painful, poorly tolerated by some patients, and declined by others and is sometimes difficult or inconvenient to perform in a busy emergency department (ED). The treatment of such patients is inconsistent, and evidence-based guidelines to support clinical management are lacking. Gout is an increasingly prevalent condition worldwide and creates a heavy economic burden.1, 3 Editor’s Capsule Summary What is already known on this topic Acute gouty arthritis is typically treated with nonsteroidal antiinflammatory agents, though prednisone has also been recommended. Their relative efficacy is unknown. What question this study addressed The efficacy and adverse effects of oral prednisolone and oral indomethacin in the treatment of acute goutlike arthritis in emergency department patients. What this study adds to our knowledge

Acute gouty arthritis is typically treated with nonsteroidal antiinflammatory agents, though prednisone has also been recommended. Their relative efficacy is unknown. What question this study addressed The efficacy and adverse effects of oral prednisolone and oral indomethacin in the treatment of acute goutlike arthritis in emergency department patients. What this study adds to our knowledge In this 90-patient randomized controlled trial, the treatments exhibited similar efficacy. Adverse effects were more common in patients given indomethacin, including gastrointestinal bleeding in 11%. How this might change clinical practice A short course of prednisolone is at least as effective as indomethacin and may be safer.

roup treated with prednisolone, whereas 5 (11%) patients in the indomethacin group developed upper gastrointestinal bleeding. Although none of the patients died because of the adverse effects, the occurrence of upper gastrointestinal bleeding, especially in the elderly, is a cause of significant morbidity.9, 10, 11, 12 The pain score at rest and during activity followed a similar pattern and trend of improvement in both the prednisolone and indomethacin groups. However, patients in the prednisolone group used slightly more acetaminophen, suggesting that prednisolone alone may not be enough to relieve the pain associated with the disease. Studies that compared cyclooxygenase-2 (COX-2) inhibitors with NSAIDs found that the former had a similar comparable efficacy but better tolerability.28, 29, 30 However, COX-2 inhibitors are expensive and are less accessible to the general population. Even more important is the concern about the cardiovascular safety of COX-2 inhibitors.28, 31, 32 The use of corticosteroids in the treatment of acute gout has also been known to be effective. Corticosteroids have been used clinically in the treatment of a large variety of conditions, ranging from allergic or ectopic diseases, asthma, and connective tissue diseases to other inflammatory conditions.

In this 90-patient randomized controlled trial, the treatments exhibited similar efficacy. Adverse effects were more common in patients given indomethacin, including gastrointestinal bleeding in 11%. How this might change clinical practice A short course of prednisolone is at least as effective as indomethacin and may be safer. Importance Oral nonsteroidal antiinflammatory drugs (NSAIDs) administered in high doses are recommended as first-line agents in the management of gout but may be contraindicated because of gastrointestinal hemorrhage or renal failure.4, 5, 6, 7, 8, 9, 10, 11, 12 Oral colchicine may be as effective as NSAIDs but is limited by toxicity at higher doses.13, 14, 15 Intraarticular corticosteroids may also be administered for monoarticular disease and intramuscular corticosteroids for podagra.4 A few small controlled studies have compared oral NSAIDs with intraarticular or intramuscular steroids and suggest that NSAIDs are as effective as steroids in the treatment of acute gout.16, 17, 18 However, NSAIDs have a significant rate of adverse effects, including gastric irritation, gastrointestinal hemorrhage, and renal failure.4 Although steroids cause severe adverse events if taken at high doses for long periods, there appear to be few adverse effects if they are taken in low to moderate doses for short periods.19 The gastrointestinal adverse effects also appear to be less severe than those of NSAIDs.5 Long-term effects, such as osteoporosis and muscle wasting, are not relevant to acute gout treatment. Oral NSAIDs have not been compared to oral steroids in the management of acute gout.

low to moderate doses for short periods.19 The gastrointestinal adverse effects also appear to be less severe than those of NSAIDs.5 Long-term effects, such as osteoporosis and muscle wasting, are not relevant to acute gout treatment. Oral NSAIDs have not been compared to oral steroids in the management of acute gout. Goals of This Investigation The objective of this study was to compare the efficacy and adverse effects of oral indomethacin/acetaminophen with oral prednisolone/acetaminophen combination therapy in the treatment of patients presenting to the ED with a clinical diagnosis of acute goutlike arthritis. Materials and Methods Study Design In this prospective, randomized, double-blind, controlled study, the analgesic efficacy and adverse effects of oral prednisolone/acetaminophen and oral indomethacin/acetaminophen combination were compared in patients presenting to an ED with a clinical diagnosis of acute gouty arthritis (Figure 1). Ethical approval was obtained from the local institutional research ethics committee. Informed written consent was obtained from each patient.Figure 1 Progress of patients through randomized trial.TID, Three times a day.

ared in patients presenting to an ED with a clinical diagnosis of acute gouty arthritis (Figure 1). Ethical approval was obtained from the local institutional research ethics committee. Informed written consent was obtained from each patient.Figure 1 Progress of patients through randomized trial.TID, Three times a day. Setting The ED of the Prince of Wales Hospital, a 1,400-bed teaching hospital in the New Territories of Hong Kong, receives about 160,000 new patients per annum, of whom about 25% are admitted to the hospital. The hospital serves a population of approximately 1.5 million. Gout is responsible for 1 to 2 patient visits per day at Prince of Wales Hospital ED. Of these patients, up to 15% are admitted to the ED observation ward or a hospital ward, mainly because the patients are elderly and lack social support. Selection of Participants All patients older than 17 years, with an acute arthritis suggestive of gout, and presenting to the ED during designated periods when research staff were on duty between 9 am and 4 pm, Monday to Friday, from February 1, 2003, to June 30, 2004, were considered for enrollment. Patients were included if they had a clinical diagnosis of acute arthritis suggestive of gout, defined as the presence of pain and warmth in a joint, and presented within 3 days of the onset of pain and also had 1 or more of the following: metatarsal-phalangeal joint involvement; knee or ankle joint involvement and aspirate containing crystals; or typical gouty arthritis, with either gouty tophi present or previous joint aspiration confirming the diagnosis of gout.

and presented within 3 days of the onset of pain and also had 1 or more of the following: metatarsal-phalangeal joint involvement; knee or ankle joint involvement and aspirate containing crystals; or typical gouty arthritis, with either gouty tophi present or previous joint aspiration confirming the diagnosis of gout. Patients were excluded if there was a clinical suspicion of sepsis or other joint disease; if follow-up was impossible because of lack of transport or lack of telephone contact; if there was significant comorbidity that would interfere with assessment; and if patients had dementia, confusion, active gastrointestinal symptoms, renal insufficiency with serum creatinine level greater than 200 μmol/L, bleeding disorder, allergy to a study drug, or joint aspirate that excluded the diagnosis of gout or were taking warfarin. It is often not possible to definitively separate gout from septic arthritis on clinical grounds alone. In this study, sepsis was considered likely if the patient had a temperature greater than 38°C, chills or rigors, a wound near the affected joint, a history of immunosuppression, erythematous tracking along a lymphatic vessel or vein in the affected limb, lymphadenopathy, or a history of septic arthritis.

rounds alone. In this study, sepsis was considered likely if the patient had a temperature greater than 38°C, chills or rigors, a wound near the affected joint, a history of immunosuppression, erythematous tracking along a lymphatic vessel or vein in the affected limb, lymphadenopathy, or a history of septic arthritis. Interventions Patients were allocated with a random-number table generated from StatView for Windows, version 5.0 (Abacus Concepts, SAS Institute, Inc., Cary, NC). The random allocation sequence was implemented with numbered sealed envelopes, such that the sequence was concealed until interventions were assigned. A nurse with clinical responsibilities opened a precoded envelope with details of the drug and a randomization number. The oral preparations of indomethacin 25 mg, acetaminophen 500 mg and prednisolone 5 mg, and identical placebos were all prepacked and placed inside the envelope. Diclofenac was prepared as a solution at 25 mg/mL, and placebo (normal saline solution) as a 1-mL solution was prepared in the ED. This nurse was not involved in the administration of analgesia, the assessment of the patient, or the treatment of adverse effects. The code was only to be broken if a physician or other nurse with clinical duties was concerned about severe adverse effects. Both the research nurse with nonclinical duties and the patient were blinded to the medication.

e administration of analgesia, the assessment of the patient, or the treatment of adverse effects. The code was only to be broken if a physician or other nurse with clinical duties was concerned about severe adverse effects. Both the research nurse with nonclinical duties and the patient were blinded to the medication. In the indomethacin group, each patient initially received diclofenac (3 mL; 75 mg) intramuscularly, indomethacin 50 mg orally, acetaminophen 1 g orally, and 6 tablets of prednisolone-like placebo orally and was observed for 120 minutes. The patient was then given a 5-day prescription of indomethacin (50 mg orally every 8 hours for 2 days, followed by indomethacin 25 mg every 8 hours for another 3 days), 6 tablets of prednisolone-like placebo once a day, and acetaminophen 1 g every 6 hours as required. In the prednisolone group, each patient initially received an intramuscular placebo injection (3 mL), prednisolone 30 mg (6 times 5 mg) orally, acetaminophen 1 g (2 tablets) orally, and indomethacin-like placebo (2 tablets) orally, and was then observed for 120 minutes. The patient was then given a 5-day prescription of indomethacin-like placebo, prednisolone 30 mg orally once per day, and acetaminophen 1 g every 6 hours as required. Both acetaminophen and intramuscular injection were given in accordance with common local practice. Many patients in Hong Kong believe that symptomatic relief will be faster if an injection is administered. The physician on duty was free to give extra doses or alternative analgesic if clinically required, and this was documented.

inophen and intramuscular injection were given in accordance with common local practice. Many patients in Hong Kong believe that symptomatic relief will be faster if an injection is administered. The physician on duty was free to give extra doses or alternative analgesic if clinically required, and this was documented. Methods of Measurement Assessment included demographic data, assessment on scheduled intervals of pain scores at rest and with activity, the occurrence of adverse effects, and the time of symptom resolution. The pain scores were assessed with a visual analogue scale from 0 (absence of pain) to 10 (the most severe pain the patient has ever experienced). A 10-cm, numbered, horizontal, visual analogue pain score was used for baseline measurements (t0) and at subsequent intervals after the first injection.20 Pain scores and adverse effects were recorded every 30 minutes for 2 hours after drug administration. Patients were aware of their previous scores at all stages of recording. Research staff contacted the patient by telephone at 24 hours (or physical assessment if the patient was admitted to the observation ward), at 5 days, and also by telephone at 14 days, unless symptoms were not resolved. Patients recorded data daily for 5 days at home. The type, number, duration, and severity of adverse effects were documented. Specific questions for each adverse event were asked each day in the collection of adverse event data, eg, “Did you experience nausea today?”

Research staff contacted the patient by telephone at 24 hours (or physical assessment if the patient was admitted to the observation ward), at 5 days, and also by telephone at 14 days, unless symptoms were not resolved. Patients recorded data daily for 5 days at home. The type, number, duration, and severity of adverse effects were documented. Specific questions for each adverse event were asked each day in the collection of adverse event data, eg, “Did you experience nausea today?” The primary clinical outcome measures were pain relief, measured as changes in pain score at rest and with activity, and adverse events. Both the change in the score and absolute value were measured. Activity involved the research nurse gently moving the joint involved in a standardized manner to assess pain. Adverse events were assessed for number, duration, and severity (where applicable). Secondary outcome measures were time to complete resolution of pain, stiffness and joint swelling, supplementary acetaminophen, and relapse rate. Data collection ceased at day 14, and nonresolution of symptoms or recurrence of symptoms at this time was regarded as a treatment failure.

, and severity (where applicable). Secondary outcome measures were time to complete resolution of pain, stiffness and joint swelling, supplementary acetaminophen, and relapse rate. Data collection ceased at day 14, and nonresolution of symptoms or recurrence of symptoms at this time was regarded as a treatment failure. Primary Data Analysis Data were analyzed on an intention-to-treat basis, and all statistical analyses involved 2-tailed tests using StatView for Windows, version 5.0 Statistical Analysis Software (Abacus Concepts, SAS Institute). P<.05 was considered as statistically significant. Because pain score and time data did not conform to the Gaussian distribution, nonparametric tests were used to analyze data.21 Baseline characteristics of the 2 treatments were analyzed using the χ2 test or Mann-Whitney U test.22 A regression line indicating the change in visual analog scale pain score over time was found, and its slope was therefore a summary measure for each patient.23 The distribution of coefficients followed a normal distribution, and so the mean slope for the coefficients of each treatment group was compared and analyzed using the t test. A previous study has shown that a difference in visual analog pain scores of less than 13 mm is unlikely to be clinically relevant.24 Therefore, unless the upper limits of the confidence intervals (CIs) were less than 13 mm, we assumed that the results were inconclusive.

roup was compared and analyzed using the t test. A previous study has shown that a difference in visual analog pain scores of less than 13 mm is unlikely to be clinically relevant.24 Therefore, unless the upper limits of the confidence intervals (CIs) were less than 13 mm, we assumed that the results were inconclusive. To detect a clinically significant difference between the mean pain score of the 2 groups of 15 mm, each with an SD of 20 mm, with an estimated power of 90% and at the 5% significance level, 37 patients were required for each group. Results During the study period, 112 patients presented with a clinical diagnosis of probable acute gout, of whom 22 were excluded because of acute gastroduodenal ulcer (n=1), trauma (n=3), serum creatinine level greater than 200 nmol/L (n=3), bleeding disorders (n=3), history of adverse reactions to NSAID (n=5), and suspicion of infectious cause (n=6). The remaining 90 patients were randomized, 46 patients to the indomethacin group and 44 patients to the prednisolone group (Figure 2). Figure 2 Interventions given to patients randomized into the 2 study groups. The demographic and other baseline characteristics of the 2 groups were similar (Table 1). There was no significant difference in initial mean pain score at rest or with activity between the 2 treatment groups. Five patients in the prednisolone group and 2 patients in the indomethacin group agreed to allow joint aspiration. All were positive for urate crystals and negative for bacterial culture.Table 1 Baseline characteristics of patients between the 2 groups.

ain score at rest or with activity between the 2 treatment groups. Five patients in the prednisolone group and 2 patients in the indomethacin group agreed to allow joint aspiration. All were positive for urate crystals and negative for bacterial culture.Table 1 Baseline characteristics of patients between the 2 groups. Variable Indomethacin Group, N=46 Prednisolone Group, N=44 Age, y, mean (SD) 66 (16) 64 (15) Male patients (%) 39 (85) 35 (80) History of gout, No. (%) 45 (98) 42 (95) Duration of symptoms, d, mean (SD) 2.3 (0.6) 2.2 (1.1) Pain score at rest (pretreatment), mm, mean (SD) 15 (20.8) 24 (25.2) Pain score with activity (pretreatment), mm, mean (SD) 74 (20.3) 78 (19.7) Single joint involved, patients (%) 45 (98) 41 (93) >1 Joint involved, No. (%) 1 (2) 3 (7) Lower limb involved, No. (%) 45 (98) 39 (89) Lower and upper limb involved, No. (%) 1 (2) 0 Fever, No. (%) 8 (17) 2 (5) Presence of tophus, No. (%) 2 (4) 5 (11) Mobility, unable to bear weight, No. (%) 25 (54) 29 (66) Admission to observation ward, No. (%) 3 (7) 8 (18) Median length of stay in observation ward, h 23.6 23.4

r limb involved, No. (%) 45 (98) 39 (89) Lower and upper limb involved, No. (%) 1 (2) 0 Fever, No. (%) 8 (17) 2 (5) Presence of tophus, No. (%) 2 (4) 5 (11) Mobility, unable to bear weight, No. (%) 25 (54) 29 (66) Admission to observation ward, No. (%) 3 (7) 8 (18) Median length of stay in observation ward, h 23.6 23.4 During the ED phase, the rate of decrease in mean pain score over time both at rest and with activity (Figure 3) was similar for both groups. The mean rate of decrease in pain score at rest for indomethacin was −6.4±8.3 (SD) mm per hour and for prednisolone was −9.5±10.5 (SD) mm per hour (mean difference 3.2 mm/hour; 95% CI −0.78 to 7.14 mm/hour; P=.12). The mean rate of decrease in pain score with activity for indomethacin was −20.3±9.1 (SD) mm per hour and for prednisolone was −19.2±11.2 (SD) mm per hour (mean difference −1.0 mm/hour; 95% CI −5.34 to 3.24 mm/hour; P=.63).Figure 3 Visual analog pain score at rest and with activity during the ED phase. Pain was assessed at rest, at time (T) 0, and at 30 min, 60 min, 90 min, and 120 min. Data are presented as means (95% CIs). There were no statistically or clinically significant differences between the groups either at rest (P=.12) or with activity (P=.63).

log pain score at rest and with activity during the ED phase. Pain was assessed at rest, at time (T) 0, and at 30 min, 60 min, 90 min, and 120 min. Data are presented as means (95% CIs). There were no statistically or clinically significant differences between the groups either at rest (P=.12) or with activity (P=.63). During the follow-up phase, the rate of decrease in mean pain score over time at rest and with activity (Figure 4) was greater in the prednisolone group than the indomethacin group. The mean rate of decrease in pain score at rest for indomethacin was −0.3±0.7 (SD) mm per day and for prednisolone was −0.7±1.2 (SD) mm per day (mean difference 0.5 mm/day; 95% CI 0.03 to 0.89 mm/day; P=.04). The mean rate of decrease in pain score with activity for indomethacin was −1.7±1.6 (SD) mm per day and for prednisolone was −2.9±2.0 (SD) mm per day (mean difference 1.2 mm/day; 95% CI 0.44 to 2.00 mm/day; P=.0026). Although these differences in activity were statistically significant, at no time was the difference in mean pain score greater than 13 mm. At day 14 (the last follow-up day), the same end points were reached by the 2 groups.Figure 4 Visual analog pain score at rest and with activity during the follow-up phase. Pain was assessed at rest for days (D) 1 to 5 and 14. Data are presented as means (95% CIs). There were no statistically or clinically significant differences between the groups at any point either at rest (P=.60) or with activity (P=.0026) (see text).

og pain score at rest and with activity during the follow-up phase. Pain was assessed at rest for days (D) 1 to 5 and 14. Data are presented as means (95% CIs). There were no statistically or clinically significant differences between the groups at any point either at rest (P=.60) or with activity (P=.0026) (see text). Both treatment groups showed improvement in the joint swelling and stiffness, but there was no difference in improvement between the 2 groups. In the indomethacin group, 29 (63%) patients experienced adverse effects compared with 12 (27%) patients in the prednisolone group (P<.05; Table 2). Nausea, indigestion, epigastric pain, and dizziness were significantly more common in the indomethacin group than the prednisolone group. The most common adverse effects in the steroid group were dry mouth and dizziness.Table 2 Adverse effects reported by patients treated with indomethacin or prednisolone for the acute goutlike arthritis.⁎

epigastric pain, and dizziness were significantly more common in the indomethacin group than the prednisolone group. The most common adverse effects in the steroid group were dry mouth and dizziness.Table 2 Adverse effects reported by patients treated with indomethacin or prednisolone for the acute goutlike arthritis.⁎ Adverse Effects Indomethacin (N=46) Prednisolone (N=44) P Value Any adverse event, No. (%) 29 (63) 12 (27) .0007 Epigastric pain, of No. (%) 14 (30) 0 (0) <.0001 Other abdominal pain, No. (%) 3 (7) 0 (0) .09 Rash, No. (%) 1 (2) 3 (7) .25 Dizziness, No. (%) 9 (19) 2 (5) .03 Drowsiness, No. (%) 9 (19) 7 (16) .79 Dry mouth, No. (%) 11 (24) 9 (20) .83 Indigestion, No. (%) 14 (30) 4 (9) .02 Nausea, No. (%) 12 (26) 3 (9) .02 Vomiting, No. (%) 4 (9) 0 .05 Diarrhea, No. (%) 3 (7) 0 .09 Serious adverse effects requiring admission, No. (%) 7 (15) 0 .007 Gastrointestinal hemorrhage, No. (%) 5 (11) 0 <.05 Shortness of breath, No. (%) 1 (2) 0 .98 Chest pain, No. (%) 1 (2) 0 .98 ⁎ Percentages may not sum to 100, because of rounding.

(26) 3 (9) .02 Vomiting, No. (%) 4 (9) 0 .05 Diarrhea, No. (%) 3 (7) 0 .09 Serious adverse effects requiring admission, No. (%) 7 (15) 0 .007 Gastrointestinal hemorrhage, No. (%) 5 (11) 0 <.05 Shortness of breath, No. (%) 1 (2) 0 .98 Chest pain, No. (%) 1 (2) 0 .98 ⁎ Percentages may not sum to 100, because of rounding. Seven patients had adverse effects that were serious enough to require treatment or hospital admission; all were in the indomethacin group (P=.007). Five (11%) patients developed gastrointestinal bleeding (11%; P<.05). Their ages ranged from 62 to 84 years. Four were admitted to the hospital, and the fifth was referred to an outpatient clinic. In each case, study medication was stopped, and upper gastrointestinal endoscopy was performed; 3 patients had acute gastric ulcers, of whom 2 received adrenaline injection for active bleeding; 2 patients had gastroduodenal ulcers. No patient developed cardiovascular shock, and all ulcers healed after treatment.

linic. In each case, study medication was stopped, and upper gastrointestinal endoscopy was performed; 3 patients had acute gastric ulcers, of whom 2 received adrenaline injection for active bleeding; 2 patients had gastroduodenal ulcers. No patient developed cardiovascular shock, and all ulcers healed after treatment. All patients in the 2 groups consumed acetaminophen during the 14 days. The mean total dose of acetaminophen consumed by the prednisolone group was significantly more than that consumed by the indomethacin group (mean 10.3 g, range 1 to 21 g versus mean 6.4 g, range 1 to 21 g; P=.008). Overall, within 14 days there were 8 patients in the indomethacin group who relapsed and reattended for further treatment compared with 5 patients in the prednisolone group (P=nonsignificant). Fifteen patients in the prednisolone group required alternative medication for pain relief compared with 17 patients in the indomethacin group (P=nonsignificant).

8 patients in the indomethacin group who relapsed and reattended for further treatment compared with 5 patients in the prednisolone group (P=nonsignificant). Fifteen patients in the prednisolone group required alternative medication for pain relief compared with 17 patients in the indomethacin group (P=nonsignificant). Limitations The diagnosis was made on clinical impression, and joint aspiration was not performed on most patients. However, in routine clinical practice the majority of patients presenting with goutlike arthritis are treated clinically and without joint aspiration, unless there is a high index of suspicion of septic arthritis or atypical features. The relatively small sample size may not have allowed us to detect small differences in pain scores. The sample size was based on our estimation of a clinically significant difference. The safety aspects were not a primary aspect of the study, and as such it was not powered to evaluate safety. It is possible that with larger numbers of patients, the prednisolone group would also have had some significant adverse events. Finally, the use of placebo injections is artificial and not standard practice. The use of placebo was important to ensure complete blinding but introduces the problem of artificiality. To what extent this limits application to clinical practice is not clear.

e group would also have had some significant adverse events. Finally, the use of placebo injections is artificial and not standard practice. The use of placebo was important to ensure complete blinding but introduces the problem of artificiality. To what extent this limits application to clinical practice is not clear. Discussion This is the first double-blind randomized placebo-controlled study comparing commonly used, inexpensive, and easily available indomethacin and prednisolone in the treatment of acute goutlike arthritis. Our results indicate that the treatment of acute goutlike arthritis with commonly prescribed doses and frequencies of prednisolone and indomethacin produces similar pain relief. However, patients in the prednisolone group took significantly more acetaminophen than the indomethacin group as an adjunct for pain relief. Prednisolone supplemented with acetaminophen may be as effective as indomethacin in relieving pain. Patients taking prednisolone experienced far fewer serious or other adverse effects compared with patients taking indomethacin.

itors.28, 31, 32 The use of corticosteroids in the treatment of acute gout has also been known to be effective. Corticosteroids have been used clinically in the treatment of a large variety of conditions, ranging from allergic or ectopic diseases, asthma, and connective tissue diseases to other inflammatory conditions. Analgesic recommendations usually rank corticosteroids as a second line of treatment, whereas the NSAIDs are usually tried first.4, 25 There are several possible reasons for physicians’ reluctance to use corticosteroids. Physicians have more experience with NSAIDs in the treatment of gout. In addition, corticosteroids have received bad publicity in terms of their numerous long-term adverse effects such as Cushing syndrome, osteoporosis, diabetes mellitus, and hypertension. The recent report of avascular necrosis after use of high-dose methylprednisolone in the treatment severe acute respiratory syndrome is another vivid example.33 Our results are in line with epidemiologic reports that acute gouty arthritis most commonly affects elderly patients and that the risk of serious gastrointestinal toxicity of NSAIDs is higher in this age group.8 According to the findings in our study, we recommend that moderate doses of oral prednisolone supplemented with oral acetaminophen be considered as first-line therapy in the treatment of acute gout. Supervising editor: Richard C. Dart, MD, PhD

Our results are in line with epidemiologic reports that acute gouty arthritis most commonly affects elderly patients and that the risk of serious gastrointestinal toxicity of NSAIDs is higher in this age group.8 According to the findings in our study, we recommend that moderate doses of oral prednisolone supplemented with oral acetaminophen be considered as first-line therapy in the treatment of acute gout. Supervising editor: Richard C. Dart, MD, PhD Author contributions: CYM had the idea for the study and has overseen the entire planning, execution, analysis, and preparation of article. PAC obtained approval. He is guarantor of the work. ITFC, PAC, and THR participated in the planning, execution, and analysis. THR prepared the statistical analysis. CYM wrote the first draft of the article, and all authors have contributed to the final version. CYM takes responsibility for the paper as a whole. Funding and support: The authors report this study did not receive any outside funding or support. Available online February 5, 2007. Reprints not available from the authors.

Introduction Respiratory tract infections are common presentations among emergency department (ED) patients, some of whom may present an infectious risk. Unfortunately, definitive identification of the offending agent is generally not possible at the initial ED visit. Potential respiratory agents that the 21st century emergency physician must consider include the traditional respiratory pathogens and also emerging (eg, severe acute respiratory syndrome [SARS]), highly virulent (eg, avian influenza virus), resistant, and even bioterrorism-related agents. Thus, within the ED there is a potentially dangerous mixture of infections with serious possible public health consequences. Threat of contagion exists for uninfected patients (in the ED, the hospital at large, and the community) and health care personnel (including ED physicians). Through implementation of the most up-to-date guidelines, health care personnel can aid in minimizing respiratory infection transmission and protect patients and other hospital personnel from infection.

d patients (in the ED, the hospital at large, and the community) and health care personnel (including ED physicians). Through implementation of the most up-to-date guidelines, health care personnel can aid in minimizing respiratory infection transmission and protect patients and other hospital personnel from infection. The potential risk for respiratory infection–related morbidity and mortality is compounded in the ED because of the increasing number of immunocompromised ED patients. Populations at increased risk include organ transplant patients,1 HIV-infected patients, and postchemotherapy patients, all of whom are living longer because of improved lifesaving therapies.2 Crowded and understaffed EDs further elevate risk of contagion and possible public health disasters. Two potentially lethal infections that are transmitted by the respiratory route, which most emergency physicians are familiar with, are Neisseria meningitis (which causes meningococcemia) and Mycobacterium tuberculosis. These agents are relatively uncommon, however, in most US EDs, and as recently as 2003, the Centers for Disease Control and Prevention (CDC) reported that health care facility environments are rarely implicated in respiratory pathogen transmission (except in cases of immunocompromised patients). Case reports of transmission of SARS among hospital workers from that year resulted in heightened awareness of the need for increased attention to respiratory precautions. For example, according to Lau et al,3 44% of SARS cases (68/156 cases) at the Prince of Wales Hospital in Hong Kong occurred in hospital workers who did not take special protective measures during the SARS outbreak. Another study found that failure of providers to recognize risk, implement strict isolation measures, and diagnose disease was responsible for the majority of nosocomial cases of SARS in Hong Kong (with the vast majority of cases occurring among physicians and nurses).4 Internationally, health care worker infection has proven to account for up to 42% of SARS cases in Canada and approximately 25% of cases in Hong Kong.4, 5 These findings provide compelling data that hospital workers are at significant risk of contracting respiratory infections and establish an imperative for initiating broad-scale infection control measures.

ction has proven to account for up to 42% of SARS cases in Canada and approximately 25% of cases in Hong Kong.4, 5 These findings provide compelling data that hospital workers are at significant risk of contracting respiratory infections and establish an imperative for initiating broad-scale infection control measures. The participation of emergency physicians and nurses is critical for effective responsiveness to respiratory threats in hospitals. ED personnel represent a critical link in the chain of communication and response, along the continuum from the community to the inpatient unit. Policies should anticipate responses to the complex spectrum of possible respiratory illnesses, from highly transmissible and unexpected emerging global diseases such as SARS to yearly influenza epidemics. Lessons from the terrorist attacks on September 11, 2001, and other recent disasters emphasize the importance of integrating the public health system with both medical and mental health services, with close attention to capacity management and surge planning. Organizational systems thus require that disaster and public health planning at regional and state levels produce systems that integrate the ED (the likely focal point for patients with acute respiratory infections) with hospital and regional response plans and resources.6, 7

to capacity management and surge planning. Organizational systems thus require that disaster and public health planning at regional and state levels produce systems that integrate the ED (the likely focal point for patients with acute respiratory infections) with hospital and regional response plans and resources.6, 7 The purpose of this report is to summarize, from both the peer-reviewed literature and public health sources (eg, from the CDC), information most relevant to ED respiratory infection control. Specific and current recommendations and guidelines are provided, along with evidence supporting specific respiratory infection control measures, when available. The review is divided into 3 sections, addressing administrative, patient-related, and legal issues, with some unavoidable overlap occurring. Administrative topics include public health coordination, facility planning, and health care worker issues. The patient-related portion covers patient flow from out-of-hospital and triage to waiting room and ED treatment areas, with inclusion of a discussion of patient education and patient transport. The legal section summarizes federal and local laws pertinent to respiratory hygiene. Because SARS represents the most recent significant respiratory pathogenic threat, many of the successes and challenges about respiratory infection control reference studies from the SARS outbreak. While this research is SARS specific, lessons that may be generalizable about infection control are provided. Further, although an all-inclusive discussion about respiratory hygiene is impossible, this summary provides the most relevant and practical information for the practicing emergency physician, with specific references provided for particular topics to allow more detailed review.

le about infection control are provided. Further, although an all-inclusive discussion about respiratory hygiene is impossible, this summary provides the most relevant and practical information for the practicing emergency physician, with specific references provided for particular topics to allow more detailed review. Background According to a recent national ED-based survey, acute respiratory infections are the leading ED “illness-related” diagnosis.8 Another recent study from the pediatric literature reported that acute respiratory illnesses are the second leading category of adolescent diagnosis from ED visits among virtually every age group (except women aged 18 to 21 years).9 The significant influx of patients expected during an outbreak (such as SARS or avian influenza) would result in an even greater proportion of ED patient visits for respiratory-related complaints.

eading category of adolescent diagnosis from ED visits among virtually every age group (except women aged 18 to 21 years).9 The significant influx of patients expected during an outbreak (such as SARS or avian influenza) would result in an even greater proportion of ED patient visits for respiratory-related complaints. The CDC has developed several specific guidelines about infection control in hospitals, with the most recent updates issued in November 2004.10, 11 The recommendations are graded according to levels of supporting evidence, as defined in Figure 1. Precautionary measures are divided into standard precautions (Figure 2) to be followed in care of all patients and transmission-based precautions to be used in addition to standard precautions according to the route of pathogen transmission. Transmission-based precautions include contact precautions for agents with potential transmission by direct or indirect contact; droplet precautions for agents with potential transmission by coughing, sneezing, talking, or performance of procedures (Figure 3); and airborne precautions for agents with potential transmission by dissemination of either airborne droplet nuclei or evaporated droplets that remain suspended in the air for long periods (Figure 4). Airborne transmission is relevant for small infectious particles that are 5 μm or smaller.Figure 1 Rating categories applying to Figure 2, Figure 3, Figure 4.

th potential transmission by dissemination of either airborne droplet nuclei or evaporated droplets that remain suspended in the air for long periods (Figure 4). Airborne transmission is relevant for small infectious particles that are 5 μm or smaller.Figure 1 Rating categories applying to Figure 2, Figure 3, Figure 4. Figure 2 Summary and level of supporting evidence for standard precautions (see Figure 1 for definitions of levels of grading); available at: http://www.cdc.gov/ncidod/hip/isolat/std_prec_excerpt.htm. Figure 3 Summary and level of supporting evidence for droplet precautions (see Figure 1 for definitions of levels of grading); available at: http://www.cdc.gov/ncidod/hip/isolat/droplet_prec_excerpt.htm. Figure 4 Summary and level of supporting evidence for airborne precautions (see Figure 1 for definitions of levels of grading); available at: http://www.cdc.gov/ncidod/hip/isolat/airborne_prec_excerpt.htm.

Figure 3 Summary and level of supporting evidence for droplet precautions (see Figure 1 for definitions of levels of grading); available at: http://www.cdc.gov/ncidod/hip/isolat/droplet_prec_excerpt.htm. Figure 4 Summary and level of supporting evidence for airborne precautions (see Figure 1 for definitions of levels of grading); available at: http://www.cdc.gov/ncidod/hip/isolat/airborne_prec_excerpt.htm. Administrative issues Administrative issues surrounding respiratory hygiene apply to the entire health care facility. Emergency physicians should take a lead role in development and implementation of policies because the ED serves as the initial entry point for many patients. Policies to address routine respiratory pathogens (eg, tuberculosis [TB] and influenza A), emerging pathogens (eg, SARS or avian influenza), and bioterrorist agents are necessary. The CDC has provided detailed recommendations about health care facility response preparedness for a SARS outbreak (available online at http://www.cdc.gov/ncidod/sars/guidance/C/recommended.htm). Although these may not all be generalized to every new respiratory threat, the principles described in the reference can guide institutional preparation for any large-scale respiratory pathogen threat. Similar readiness plans for bioterrorism preparedness have been devised and published conjointly by the CDC and the Association of Professionals in Infection Control and Epidemiology.12

threat, the principles described in the reference can guide institutional preparation for any large-scale respiratory pathogen threat. Similar readiness plans for bioterrorism preparedness have been devised and published conjointly by the CDC and the Association of Professionals in Infection Control and Epidemiology.12 Public health coordination Surge Capacity Plans All health care facilities should have policies and procedures in place for respiratory infection control practice with specific operational plans for handling a large influx of potentially infectious patients in the event of a significant outbreak.13 When patient influx exceeds institutional capacity, plans should designate alternative triage and treatment areas either outdoors or in other nearby large-capacity facilities.14 Although plans may designate patient care areas that exceed hospital capacity, staffing issues may limit the ability to actually use these areas in a real event. Community isolation and treatment facilities may also be activated; a prototype for such a facility was developed by the CDC for SARS. In general, community facilities will likely house and treat patients with milder disease, with the public health department coordinating these procedures and venues. Community isolation facilities (eg, motels, hotels) should have rooms that are equipped with private bathrooms, as well as receptacles to dispose of soiled linen and contaminated waste. Personnel who work at the facility should also have N-95 respirators available, as well as disposable gowns and gloves. In general, patients at these facilities will be expected to care for themselves.15 Other hospital infection-control procedures may involve cohorting potentially infectious patients (if isolation beds are not available), as well as rapidly discharging appropriate inpatients and canceling elective procedures to alleviate strain on hospital resources.

acilities will be expected to care for themselves.15 Other hospital infection-control procedures may involve cohorting potentially infectious patients (if isolation beds are not available), as well as rapidly discharging appropriate inpatients and canceling elective procedures to alleviate strain on hospital resources. Preemptive planning and knowledge of health care facility (and ED) resource availability are critical components of preparedness. Lack of resource planning was cited as a significant factor that contributed to the spread of SARS in Southeast Asia and Toronto.16 Unfortunately, few recent data exist describing the availability of isolation facilities in US EDs. A 1995 study found that less than 20% of EDs had negative-pressure isolation rooms.17 In a recent press release from November 2005, American College of Emergency Physicians leaders warned that there is an urgent need for increased ED and hospital planning, specifically citing lack of adequate surge capacity, isolation facilities, and staff to treat the large increase in the number of patients that may result from an influenza pandemic.18 Depending on the circumstances of the outbreak, public health officials may recommend keeping suspected noncritical infectious patients at home. The CDC’s guidelines for home isolation for SARS and pandemic influenza serve as prototypes.19, 20 Alternatively, specific health care facilities may be designated as referral centers for suspected cases.15 Although the public health department will ultimately be responsible for coordinating implementation of these types of large-scale overcapacity plans, emergency physicians need to understand the types of options available. ED physician participation in policy development will be critical in providing practical guidance for ED patient care and operations.

t will ultimately be responsible for coordinating implementation of these types of large-scale overcapacity plans, emergency physicians need to understand the types of options available. ED physician participation in policy development will be critical in providing practical guidance for ED patient care and operations. Contact Tracing Policies to support rapid identification of patients with suspected respiratory infections that have serious public health consequences (eg, SARS, avian influenza) should include mechanisms for definitive diagnostic testing and immediate reporting to the local health department. The hospital laboratory should be advised to take appropriate precautions with specimens and coordinate specialized testing with local or state health department laboratories.

(eg, SARS, avian influenza) should include mechanisms for definitive diagnostic testing and immediate reporting to the local health department. The hospital laboratory should be advised to take appropriate precautions with specimens and coordinate specialized testing with local or state health department laboratories. In a suspected outbreak with potential epidemic risk, procedures for contact tracing must be instituted. Effective communications mechanisms between EDs and health departments are required to allow contact tracing of potentially exposed patients, visitors, and health care workers who live in the community. Contact tracing involves either active or passive monitoring. Active monitoring consists of direct public health contact (telephone or in person), for example, once a day for exposed persons to assess for symptoms and address any needs. Passive monitoring relies on the affected person’s contacting the health authorities if symptoms develop. Methods of monitoring depend on the exposure risk and capacity of the public health infrastructure. Regardless of the type of monitoring recommended, all individuals in contact with a potentially infectious person need to be advised of symptoms and what to do if symptoms develop. Additionally, persons with high-risk exposures may require activity restrictions. Although the public health department would be responsible for the contact tracing process, emergency physicians need to understand these basic principles because they will likely be called on to work closely with public health departments and provide information about persons who are infected or exposed while in the ED. Telephone numbers for the local health department should be readily available in all EDs.

emergency physicians need to understand these basic principles because they will likely be called on to work closely with public health departments and provide information about persons who are infected or exposed while in the ED. Telephone numbers for the local health department should be readily available in all EDs. Communication Policies should include clear designations of specific persons within the hospital who are responsible for communication with public health officials (eg, hospital infection control officer) and dissemination of up-to-date information to health care staff (eg, hospital chief executive officer). Policies need to include processes for initiating communication with key public health officials after hours and on weekends and guidance about when communication should be initiated. Potential community contacts should be identified in advance and be capable of effectively communicating needs and concerns of the public.6

icies need to include processes for initiating communication with key public health officials after hours and on weekends and guidance about when communication should be initiated. Potential community contacts should be identified in advance and be capable of effectively communicating needs and concerns of the public.6 Facility planning Infection Containment Technology Although proper patient care is the main priority within the ED, the burden of protecting uninfected individuals from communicable illnesses is critical for minimizing spread of disease and the influx of new cases. The “hierarchy of control technologies” consists of (in order of effectiveness) engineering controls, administrative and work practice controls, and use of personal protective equipment. Consistent application of these principles demonstrated success in limiting TB resurgence more than a decade ago and, more recently, the spread of SARS.21 Understanding the hierarchy allows comprehensive planning, clear implementation, and appropriate local adaptations. The most effective practices from each category should be implemented according to characteristics of the responsible agent. For instances in which the infectious agent is unknown, the most restrictive isolation methods available should be instituted. Emergency physicians’ preparedness thus requires understanding of institutional resource availability and capacity and early initiation of infectious disease or public health consultation if a new outbreak is suspected or institutional capacity is at risk of being overwhelmed. Problems with limited isolation resources in the ED or inpatient setting are usually best addressed in the short term by use of cohorting strategies.

apacity and early initiation of infectious disease or public health consultation if a new outbreak is suspected or institutional capacity is at risk of being overwhelmed. Problems with limited isolation resources in the ED or inpatient setting are usually best addressed in the short term by use of cohorting strategies. Engineering Controls Engineering controls provide passive protection for health care workers, visitors, and patients. Measures include use of isolation rooms (including negative pressure), filtration devices, and physical separation (eg, closing doors or cohorting). Figure 5 summarizes recommendations (and supporting level of evidence) published by the National Institute for Occupational Safety and Health that are specific for maintaining airborne infection isolation rooms.Figure 5 Recommendations about air handling systems in health care facilities from Guidelines for environmental infection control in health-care facilities: recommendations of CDC and the Health Care Infection Control Practices Advisory Committee. MMWR Morb Mortal Wkly Rep. 2003;52:5-13,n0.rr10. Available at: http://www.cdc.gov/MWWR/preview/MMWRhtml/rr5210a1.htm.

ystems in health care facilities from Guidelines for environmental infection control in health-care facilities: recommendations of CDC and the Health Care Infection Control Practices Advisory Committee. MMWR Morb Mortal Wkly Rep. 2003;52:5-13,n0.rr10. Available at: http://www.cdc.gov/MWWR/preview/MMWRhtml/rr5210a1.htm. Negative pressure isolation systems prevent contaminated air from traveling to other areas of the ED or hospital, which is the most efficient method for early containment of infectious respiratory pathogens because airflow from either single rooms or small units can be controlled. However, when the organism load is extremely high, negative-pressure units may not be 100% effective, because they leave live pathogen in the air or on surfaces.22, 23, 24, 25 Increased efficacy can be realized by supplementing negative-pressure isolation systems with a high efficiency particulate air (HEPA) filtration system. HEPA filtration systems supplement negative-pressure systems, removing fungi and bacteria greater than 0.1 μm from the atmosphere. These can be installed in ventilation ducts but are also available as portable units. Addition of ultraviolet lights allows killing of spores and active organisms. All HEPA filters must be properly installed and maintained according to the manufacturer’s instructions to ensure satisfactory decontamination.25, 26

phere. These can be installed in ventilation ducts but are also available as portable units. Addition of ultraviolet lights allows killing of spores and active organisms. All HEPA filters must be properly installed and maintained according to the manufacturer’s instructions to ensure satisfactory decontamination.25, 26 Closing doors and cohorting of patients are recommended if no proper isolation room is available. Such methods proved effective in Hong Kong in early 2003, when SARS patients were cohorted into 3 separate observation wards, with no subsequent secondary transmission reported.27, 28 Unfortunately, if not done properly, cohorting in open wards may contribute to increased infection, as was seen in the early Toronto SARS experience.27 Thus, it should be recognized that although possibly beneficial as an adjunctive measure when resources are scarce, physical separation and cohorting do not guarantee protection. Accordingly, health care workers should use proper infection controls when visiting patients in rooms, including droplet precautions and, if indicated, personal Occupation Safety and Health Administration (OSHA)–approved respirators.

n resources are scarce, physical separation and cohorting do not guarantee protection. Accordingly, health care workers should use proper infection controls when visiting patients in rooms, including droplet precautions and, if indicated, personal Occupation Safety and Health Administration (OSHA)–approved respirators. Administrative and Work Practice Controls Administration of effective infection containment requires written policies and procedures and is the “second tier” in the hierarchy of infection control. Operational policies should include explicit criteria for suspecting disease, restricting contact with patients suspected of having infection, controlling transport and high-risk procedures, quarantining of patients and contacts, contact tracing, implementing methods for disinfection, and monitoring of isolation procedures.21 Procedural policies should address need for supplemental staff, education and training for health care workers, medical surveillance of exposed health care workers, and communication with public health officials and the general public.21 Rapid implementation of these policies is the key to infection control in an outbreak scenario and proved critically important in controlling SARS.29 Although SARS specific, the CDC’s checklist for SARS Preparedness in Health Care Facilities16 provides a prototype of the types of policies and procedures that should be considered in the event of any bioterrorism-related or emerging communicable respiratory outbreak.

proved critically important in controlling SARS.29 Although SARS specific, the CDC’s checklist for SARS Preparedness in Health Care Facilities16 provides a prototype of the types of policies and procedures that should be considered in the event of any bioterrorism-related or emerging communicable respiratory outbreak. In the aftermath of an infectious outbreak, postevent analysis should be conducted to determine which management efforts were effective and which were not from the hospital’s and community’s perspective. Representatives of all affected departments and organizations should be included, and appropriate revisions should be incorporated into policies. Health care worker surveillance should also attend to posttraumatic stress assessment and treatment, as indicated.30

were not from the hospital’s and community’s perspective. Representatives of all affected departments and organizations should be included, and appropriate revisions should be incorporated into policies. Health care worker surveillance should also attend to posttraumatic stress assessment and treatment, as indicated.30 Personal Protective Equipment Personal protective equipment, including gloves, gown, masks, and respirators, provides barrier protection, preventing skin and mucous membrane exposures. Although these resources offer protection to individuals by reducing likelihood of direct contact, they are categorized as the third hierarchy level because they do not eliminate the pathogen and may have limited effectiveness because of problems such as variable health care worker adherence, potential for equipment failure, and inadequate equipment availability. The 2 CDC- and OSHA-approved personal air filtration systems are N95 masks and powered air-purifying respirators. N95 masks are simple and inexpensive but require individual fit-testing. Powered air-purifying respirators offer the advantage of nearly universal fit but are far more expensive and cumbersome to use. Both devices require medical evaluation and clearance for safe use. Personal protective equipment should be used by all health care workers in outbreak settings when patients with potentially contagious respiratory infections are treated.

vantage of nearly universal fit but are far more expensive and cumbersome to use. Both devices require medical evaluation and clearance for safe use. Personal protective equipment should be used by all health care workers in outbreak settings when patients with potentially contagious respiratory infections are treated. Health care workers Education Proper education of health care workers about respiratory hygiene practice is critical for effective infection control. Written policies and procedures for education and training of health care workers should be developed at the institutional level. Education topics should include infection control precautions, criteria for suspecting disease at first contact and methods for restricting contact with patients suspected of having infection, limiting and controlling patient transport, and minimizing exposure during high-risk procedures. Additional educational topics for ED providers and administrators include criteria and procedures for quarantining of contacts, protocols for disinfection and for monitoring isolation, and methods for maintaining medical surveillance of exposed health care workers.21

and minimizing exposure during high-risk procedures. Additional educational topics for ED providers and administrators include criteria and procedures for quarantining of contacts, protocols for disinfection and for monitoring isolation, and methods for maintaining medical surveillance of exposed health care workers.21 Staffing Issues Providing adequate hospital staffing is important in any disaster, and personnel issues particular to infectious disasters must be given consideration in developing hospital and ED plans. All health care facilities should have policies and procedures for mobilizing and reassigning staff to more critical areas in the event of a disaster. Because health care staff may be reluctant to come to work if they believe they are at risk of contracting an infectious illness, it is critical that the facility planning measures be reviewed in advance, with contingencies and backups in place. Health care workers (particularly those working in front-line ED settings) should also be given priority for receiving vaccines or prophylactic antimicrobials, when appropriate. Offering additional incentives to staff to come to work may also be required in certain situations.

, with contingencies and backups in place. Health care workers (particularly those working in front-line ED settings) should also be given priority for receiving vaccines or prophylactic antimicrobials, when appropriate. Offering additional incentives to staff to come to work may also be required in certain situations. Infectious outbreaks create the additional problem that health care workers themselves may become ill. Plans for respiratory outbreaks should include regular evaluation of health care workers for infectious signs or symptoms, criteria for removing health care workers from patient care, and criteria for quarantine (either at home or in the workplace). Health care workers’ desire for a workplace quarantine option was demonstrated during the SARS outbreak, in which individuals did not want to subject family members to an increased risk of infection.

ia for removing health care workers from patient care, and criteria for quarantine (either at home or in the workplace). Health care workers’ desire for a workplace quarantine option was demonstrated during the SARS outbreak, in which individuals did not want to subject family members to an increased risk of infection. Vaccination and Chemoprophylaxis Recommendation Although the CDC provides recommendations for influenza vaccination among health care workers,31 there are no uniform recommendations for health care worker vaccination for all potential respiratory pathogens. In light of this, the influenza recommendations not only serve to guide planning for annual influenza epidemics but also may provide a template for other vaccine-preventable pathogens. Research has demonstrated that influenza vaccination of health care workers contributes to a substantial decrease in patient mortality,32 which has led some experts to call for mandatory vaccination of health care workers.33 The CDC provides specific recommendations about when to provide chemoprophylaxis for influenza,34 which may be used as a template and adapted to other pathogens when guidelines are developed for new and emerging pathogens for which vaccines are available. Vaccinations plans for certain agents (eg, anthrax and smallpox) are controversial. Currently, preexposure anthrax vaccine is not recommended for health care workers. After the terrorist attacks of 2001, the US government developed a smallpox vaccination plan that included “formation of smallpox response teams” at each institution. Emergency physician volunteers participated as critical members of the team. Although controversy still exists in the emergency medicine community about these recommendations,35 they remain. Because the threat of a true smallpox event remains low, however, routine vaccination for all health care workers for smallpox is not recommended by the Advisory Committee on Immunization Practice.36 Facilities should create a priority list for employee smallpox vaccination in the event of an outbreak, and emergency staff should be included.

true smallpox event remains low, however, routine vaccination for all health care workers for smallpox is not recommended by the Advisory Committee on Immunization Practice.36 Facilities should create a priority list for employee smallpox vaccination in the event of an outbreak, and emergency staff should be included. Patient issues Concerns about the potential spread of respiratory pathogens begin at the point of entry into the health care system and continue to the inpatient setting. Emergency physicians need to be aware of the potential for infection, illness, and transmissibility in a variety of potentially high-risk environments, including (1) emergency medical services (EMS) and triage settings (in which historical and clinical information may be limited and risk underestimated), (2) during performance of “high-risk” invasive airway procedures, and (3) during patient transport to the various inpatient units throughout the hospital.

vironments, including (1) emergency medical services (EMS) and triage settings (in which historical and clinical information may be limited and risk underestimated), (2) during performance of “high-risk” invasive airway procedures, and (3) during patient transport to the various inpatient units throughout the hospital. EMS issues The CDC provides specific recommendations for EMS transport of SARS patients.37 Although specific EMS recommendations do not exist for each of the transmissible respiratory threats, the general principles outlined in the SARS directives are applicable to the transport of any patient with a suspected serious and contagious life-threatening respiratory infection37 and include the following: (1) potentially contagious patients should be transported with as few EMS personnel as possible, (2) family members should not be allowed to ride with patients in the ambulance, (3) EMS personnel traveling with a patient suspected of having infection should wear proper personal protective equipment, including isolation gown, double gloves, facemask, and N95 or higher-grade respirator (eg, N99, 100, a powered air-purifying respirator), (4) patients should wear a surgical mask if feasible and, if not, use tissues to cover their mouth or nose during coughing or sneezing; and (5) patients should be transported in a vehicle that has separate ventilation systems and compartments for patient and driver, whenever possible. Finally, advanced ED notification is advised to facilitate prearrival planning to limit exposure of other individuals. EMS personal protective equipment should be handled as medical waste, and EMS vehicles should be decontaminated before transporting another patient.38

rtments for patient and driver, whenever possible. Finally, advanced ED notification is advised to facilitate prearrival planning to limit exposure of other individuals. EMS personal protective equipment should be handled as medical waste, and EMS vehicles should be decontaminated before transporting another patient.38 ED triage and waiting room The importance of implementing effective triage and ED-based diagnostic strategies is underscored by experience with highly transmissible respiratory infections such as TB and SARS. Several hospital- and ED-based studies provide data that demonstrate that lack of either provider education or adherence to institutional guidelines or inadequate diagnostic evaluation of patients at risk results in increased risk of disease transmission.3, 5, 16, 17, 39 Underscoring this is the findings from one epidemiologic outbreak of SARS in Toronto that found that 36% of new infections in the hospital occurred in health care workers, with the highest rates in those working in EDs and ICUs.5

ion of patients at risk results in increased risk of disease transmission.3, 5, 16, 17, 39 Underscoring this is the findings from one epidemiologic outbreak of SARS in Toronto that found that 36% of new infections in the hospital occurred in health care workers, with the highest rates in those working in EDs and ICUs.5 Both the World Health Organization and the CDC provide general recommendations for handling of patients with suspected respiratory infections that include having triage staff adhere to proper hand hygiene procedures and donning face masks and eye protection.40, 41 If SARS or TB is suspected, health care workers in EDs should don an N-95, 99, or 100 respirator.42 The degree of vigilance that should be applied to screening for respiratory infections depends on the current risk level, with the most up-to-date regional risk information based on surveillance data provided on a CDC Web site.43, 44 For example, there are 3 basic risk levels that apply to SARS: (1) no current SARS transmission anywhere in the world, (2) active SARS transmission in limited geographic areas, and (3) SARS transmission within the community in which one is practicing. In the absence of person-to-person transmission of SARS worldwide, the goal of domestic surveillance is to maximize early detection of cases while minimizing unnecessary laboratory testing and social disruption. In the absence of known transmission worldwide, the overall likelihood that a person in the United States with fever and respiratory symptoms will have SARS is exceedingly low.45 If SARS transmission is present in limited geographic areas, screening should focus on identifying persons with possible geographic exposures. When person-to-person SARS transmission is present in the community, everyone with fever or respiratory symptoms should be screened for SARS.

ill have SARS is exceedingly low.45 If SARS transmission is present in limited geographic areas, screening should focus on identifying persons with possible geographic exposures. When person-to-person SARS transmission is present in the community, everyone with fever or respiratory symptoms should be screened for SARS. In an outbreak scenario (eg, SARS, avian influenza, or TB), explicit written criteria should be provided to triage personnel to allow rapid isolation of patients who may be harboring a highly contagious infection. The CDC has issued specific screening tools to be used for rapid detection and isolation of possible SARS patients, depending on the absence or presence of person-to-person transmission in the world (Figure 6, Figure 7). Various similar ED-based triage guidelines for specific agents (eg, TB, influenza and avian influenza)13, 17, 34 that include use of early radiography have been developed, and the CDC Web site (available online at http://www.cdc.gov) should be consulted for the most up-to-date recommendations, as well as the current threat level of SARS.44 From the ED perspective, development of decision guidelines may be based on the characteristics of the epidemic and may require development and modification in real time. An excellent example is provided by an ED in Singapore, in which a triage tool, developed throughout a 1-year period, yielded a false-negative rate for SARS case identification of 0.28%.46, 47 Figure 6 CDC algorithm for evaluation and treatment of patients requiring hospitalization for radiographically confirmed pneumonia in the absence of person-to-person transmission of SARS-Coronavirus. Available at: www.cdc.gov/ncidod/sars/clinicalguidanceframe1.htm.

tive rate for SARS case identification of 0.28%.46, 47 Figure 6 CDC algorithm for evaluation and treatment of patients requiring hospitalization for radiographically confirmed pneumonia in the absence of person-to-person transmission of SARS-Coronavirus. Available at: www.cdc.gov/ncidod/sars/clinicalguidanceframe1.htm. Figure 7 CDC algorithm for management of fever or respiratory symptoms when SARS-CoV person-to-person transmission is occurring. Available at: http.www.cdc.gov/ncidod/sars/clinicalguidanceframe2.htm. The CDC recommends that tissues and masks be made readily available for all symptomatic patients who enter the ED or hospital doors (to cover their mouths and noses) and that sinks or handwashing stations be accessible for all patients in waiting rooms and triage areas.13 During periods of increased respiratory infections (eg, influenza season), separation of symptomatic and asymptomatic patients in waiting rooms and triage areas is advised, and surgical masks should be distributed to all patients with active respiratory symptoms. When it is not feasible to set up separate waiting areas in the ED, symptomatic patients should be encouraged to sit at least 3 feet away from other patients in the waiting room. According to the CDC, this practice is supported by level IB evidence.

hould be distributed to all patients with active respiratory symptoms. When it is not feasible to set up separate waiting areas in the ED, symptomatic patients should be encouraged to sit at least 3 feet away from other patients in the waiting room. According to the CDC, this practice is supported by level IB evidence. Patient education The CDC recommends that visual education be provided at all patient entrances to EDs during periods of heightened respiratory alert. Visual alerts (including signs, pamphlets, and other general education measures about respiratory hygiene) are proven measures that can decrease disease transmission.13 It is recommended that visual alerts be present in several languages (depending on the region of the country and population served) and be provided at an appropriate reading level to allow for comprehension by the majority of the population. Content of educational material should include a general description of standard respiratory hygiene methods, including handwashing, use of disposable tissues for covering mouth and nose, and staying at least 3 feet away from persons with symptoms.

ing level to allow for comprehension by the majority of the population. Content of educational material should include a general description of standard respiratory hygiene methods, including handwashing, use of disposable tissues for covering mouth and nose, and staying at least 3 feet away from persons with symptoms. ED treatment area Although proper patient care is the main priority within the ED, protecting uninfected patients from communicable illnesses is also important. Early isolation decreases the likelihood of person-to-person transmission. Patients with a suspected but unidentified communicable respiratory infection should be placed in an environment with the highest level of protection available until definitive identification of the offending pathogen can be made or the possibility of a public health threat can be safely ruled out.

person transmission. Patients with a suspected but unidentified communicable respiratory infection should be placed in an environment with the highest level of protection available until definitive identification of the offending pathogen can be made or the possibility of a public health threat can be safely ruled out. Laboratory diagnosis of respiratory contagious pathogens represents a critical step in decisionmaking about the need for isolation, treatment, and disposition. Unfortunately, from the standpoint of the emergency physician, most current criterion-standard laboratory assays rely on serologic or culture methodologies often requiring days to weeks for definitive reporting. Even when alternative nonculture-based methodologies are available (eg, Acid-fast bacillus smear results for TB), reliable confirmation requires multiple sample procurement during a period of several days. For this reason, decisionmaking about patient care relies on clinical suspicion, which includes current knowledge of the community likelihood of a respiratory infectious event, risk status of the patient, and patient presenting signs and symptoms, which are often nonspecific. As described under “ED Triage and Waiting Room,” clinical guidelines may be used as well for assistance.

Laboratory diagnosis of respiratory contagious pathogens represents a critical step in decisionmaking about the need for isolation, treatment, and disposition. Unfortunately, from the standpoint of the emergency physician, most current criterion-standard laboratory assays rely on serologic or culture methodologies often requiring days to weeks for definitive reporting. Even when alternative nonculture-based methodologies are available (eg, Acid-fast bacillus smear results for TB), reliable confirmation requires multiple sample procurement during a period of several days. For this reason, decisionmaking about patient care relies on clinical suspicion, which includes current knowledge of the community likelihood of a respiratory infectious event, risk status of the patient, and patient presenting signs and symptoms, which are often nonspecific. As described under “ED Triage and Waiting Room,” clinical guidelines may be used as well for assistance. Rapid diagnostic assays for contagious respiratory pathogens hold great promise with regard to assisting ED physicians in treatment of patients with suspected respiratory contagious pathogens. Although significant molecular advancements have recently been made in design and evaluation of rapid molecular-based methods, most notably using polymerase chain reaction techniques, few have reached the status of standard of care for point-of-care use. Rapid diagnostic assays for influenza are available, but none has adequate sensitivity or specificity to allow recommendation for definitive care in ED settings.48

pid molecular-based methods, most notably using polymerase chain reaction techniques, few have reached the status of standard of care for point-of-care use. Rapid diagnostic assays for influenza are available, but none has adequate sensitivity or specificity to allow recommendation for definitive care in ED settings.48 High-Risk Airway Procedures Interventional airway procedures in the ED (including use of nebulized therapy and endotracheal intubation) increase risk for airborne transmission of disease because they result in release of high pathogen loads.49 Although most procedures can be done in the ED, the US Department of Health and Human Services recommends that in outbreak settings, aerosol-generating procedures (eg, nebulized medications or bilevel positive air pressure) be avoided as much as possible.50 When essential for patient care, health care workers involved in these procedures should use N95 respirators or powered air-purifying respirators, along with gloves and gowns. After the procedure is completed, personal protective equipment should be removed and safely discarded to avoid contaminating the health care worker or the environment.51 Specific detailed recommendations about intubation suggest that added measures be taken to reduce unnecessary exposure to health care workers, including reducing the number of health care workers present and adequately sedating or paralyzing the patient to reduce the possibility of a cough.52 All high-risk procedures should be performed only by highly experienced staff.

uggest that added measures be taken to reduce unnecessary exposure to health care workers, including reducing the number of health care workers present and adequately sedating or paralyzing the patient to reduce the possibility of a cough.52 All high-risk procedures should be performed only by highly experienced staff. Transport Issues It is recommended that patient transport and movement from the room be limited to essential purposes only. When transport out of the room occurs, masks should be worn by the patient to reduce the opportunity for transmission to patients and staff and reduce environmental contamination. Further, health care workers in the area to which the patient is to be taken should be notified in advance.53 Legal issues There is a forceful interplay between the health and well-being of the public in general and an individual’s rights, which is set within a complex and often confusing legal field.54 This affects emergency practitioners and health care facilities concerning respiratory infections primarily in 2 ways: (1) through the need to notify appropriate public health authorities of reportable infectious diseases, and (2) through the requirement to isolate ill patients and quarantine sick contacts.

al field.54 This affects emergency practitioners and health care facilities concerning respiratory infections primarily in 2 ways: (1) through the need to notify appropriate public health authorities of reportable infectious diseases, and (2) through the requirement to isolate ill patients and quarantine sick contacts. The CDC is recognized as the lead federal agency for protecting the health of the public and has various federal responsibilities in this regard, including investigations of unusual diseases and federal quarantine authority.55 According to Title 42 United States Code Section 264, the surgeon general, with the approval of the secretary, is authorized to make and enforce regulations to prevent the introduction, transmission, or spread of communicable diseases. However, the current legal framework of public health oversight and response in the United States is a complex mix of state and federal laws. Thus, the specific requirements for any practitioner, ED, or hospital vary according to the local and state laws.56, 57 The federal government has oversight of importation of infectious diseases and overall quarantine authority, but the individual states generally have the primary authority and responsibility of responding to public health problems within their jurisdiction, such as investigating a cluster of TB cases and isolating infectious individuals. States also have the responsibility of addressing their own public health emergencies.

y, but the individual states generally have the primary authority and responsibility of responding to public health problems within their jurisdiction, such as investigating a cluster of TB cases and isolating infectious individuals. States also have the responsibility of addressing their own public health emergencies. The interface between law, medicine, and public health requires the balancing of many potentially competing interests, especially individual human rights versus the need to protect the public’s health. There is significant background and legal precedent on this topic.58 It is best for institutions to have an existing relationship with local or state public health officials to ensure ongoing bidirectional communication in times of urgency or emergency. As in any emergency, adequate preparedness, coupled with clear communication, allows for coordinated response. Disease Reporting The list of reportable diseases is established by each state or territory, though the CDC has recommended specific case definitions for infectious conditions that could fall under public health surveillance.57 Timeliness and mechanism for reporting also vary for different diseases. For example, a case of smallpox requires an immediate telephone call, whereas cases of gonorrhea may be reported in a weekly written report. Although this reporting activity may be mandated, it raises important legal and ethical issues about the balance between the duty to report and an individual’s right to privacy.

. For example, a case of smallpox requires an immediate telephone call, whereas cases of gonorrhea may be reported in a weekly written report. Although this reporting activity may be mandated, it raises important legal and ethical issues about the balance between the duty to report and an individual’s right to privacy. Quarantine and Isolation The surgeon general is responsible for controlling, directing, and managing all United States quarantine stations, which includes isolation for people who are ill and quarantine for people exposed but not ill. In April 2003, SARS was added to the list of diseases for which quarantine is authorized (other diseases included are cholera, diphtheria, TB, plague, smallpox, yellow fever, and viral hemorrhagic fever). A lesson from SARS quarantines in Singapore is the capacity of a highly contagious infection to cause a rapid pandemic. The implications of quarantining a population or individuals for the length of the incubation time (or the length of the illness if patient is infected) are numerous. Quarantining a large population involves significant commitment of resources. To overcome the legal obstacles of a major quarantine, a plan must be in place well in advance of an outbreak. As part of the public health infrastructure, ED health care workers may be called on to participate in various infection containment strategies, including quarantining of individuals or vaccinating large segments of the population.

obstacles of a major quarantine, a plan must be in place well in advance of an outbreak. As part of the public health infrastructure, ED health care workers may be called on to participate in various infection containment strategies, including quarantining of individuals or vaccinating large segments of the population. OSHA One other important legal aspect relates to occupational safety. OSHA has a number of rules and regulations designed to protect the health and safety of health care workers. OSHA’s jurisdiction includes all health care facilities. Health care workers in EDs should be aware that rules and regulations related to respiratory hygiene are legally mandated and must be implemented in hospitals in accordance with current guidelines, as described elsewhere. Future research The increasing likelihood that a highly contagious respiratory outbreak such as pandemic influenza will be seen soon, coupled with recognition of the presence of significant gaps between experimental and theoretic advances in both technologic and methodologic approaches to infection control (versus true ED preparedness), has created the need for further research.

gious respiratory outbreak such as pandemic influenza will be seen soon, coupled with recognition of the presence of significant gaps between experimental and theoretic advances in both technologic and methodologic approaches to infection control (versus true ED preparedness), has created the need for further research. Rapid point-of-care diagnostics hold great potential for improving triage, treatment, and disposition planning. Future research will need to bridge the divide between the numerous point-of-care assays that are under development and the need to have a reliable, easy-to-use test that is adequately sensitive and specific for clinical decisionmaking. Although such development will likely take several years of investigation, 2 such diagnostics in early phases of development include a polymerase chain reaction–based respiratory pathogen panel59 and a mass spectrophotometry platform that can rapidly evaluate polymerase chain reaction products to identify any potential new emerging threat.60

nt will likely take several years of investigation, 2 such diagnostics in early phases of development include a polymerase chain reaction–based respiratory pathogen panel59 and a mass spectrophotometry platform that can rapidly evaluate polymerase chain reaction products to identify any potential new emerging threat.60 There are also multiple practical issues related to ED evaluation requiring study, including development of more effective clinical decision guidelines for isolation and diagnosis and determination of the impact and best practice methods for care in ambient settings. The effectiveness of ED air filtration techniques also remains unclear, and educational research in this area is required. The numerous ethical, legal, and practical challenges associated with isolation and quarantining of patients will also require further study, with emphasis on ED-specific questions such as the role of EDs in care of “routine” emergencies, development of ED surge capacity, and optimization of methods for coordination of EDs with the public health sector. One other area of research that is gaining increased attention and has particular relevance for EDs involves surveillance methods for tracking respiratory illnesses. Current approaches that involve ED-based researchers include syndromic surveillance based on ED complaints,61 evaluation of the efficacy of increased diagnostic testing in EDs,62 and tracking of ED prescriptions.63 These new areas of research will likely grow rapidly as the threat of respiratory infections becomes more prevalent.

roaches that involve ED-based researchers include syndromic surveillance based on ED complaints,61 evaluation of the efficacy of increased diagnostic testing in EDs,62 and tracking of ED prescriptions.63 These new areas of research will likely grow rapidly as the threat of respiratory infections becomes more prevalent. Summary This review serves as a brief synopsis of the issues surrounding respiratory hygiene as they relate to the ED. Protecting patients and staff is a difficult task in the ED because cases of contagious respiratory infections are often not immediately identifiable. This report focuses on the development of appropriate policies relating to patients with potential transmissible respiratory pathogens. Education of key individuals, along with rapid dissemination of accurate information, is necessary to support these policies and will be instrumental in ensuring effective implementation. Emergency physicians will continue to be pivotal in the development of these policies by maintaining active administrative and leadership positions in hospitals, and advancing understanding of the critical role they play in the early identification, treatment, and containment of these potentially lethal respiratory pathogens.25 The ACEP Public Health Committee is grateful to Ms. Margaret Montgomery and Ms. Julie Dill for their time and assistance in preparation of this manuscript. Supervising editors: David A. Talan, MD; Michael L. Callaham, MD Funding and support: The authors report this study did not receive any outside funding or support.

Summary This review serves as a brief synopsis of the issues surrounding respiratory hygiene as they relate to the ED. Protecting patients and staff is a difficult task in the ED because cases of contagious respiratory infections are often not immediately identifiable. This report focuses on the development of appropriate policies relating to patients with potential transmissible respiratory pathogens. Education of key individuals, along with rapid dissemination of accurate information, is necessary to support these policies and will be instrumental in ensuring effective implementation. Emergency physicians will continue to be pivotal in the development of these policies by maintaining active administrative and leadership positions in hospitals, and advancing understanding of the critical role they play in the early identification, treatment, and containment of these potentially lethal respiratory pathogens.25 The ACEP Public Health Committee is grateful to Ms. Margaret Montgomery and Ms. Julie Dill for their time and assistance in preparation of this manuscript. Supervising editors: David A. Talan, MD; Michael L. Callaham, MD Funding and support: The authors report this study did not receive any outside funding or support. Publication dates: Available online August 23, 2006. Reprints not available from the authors.

Introduction Emergency medical services (EMS) professionals respond daily to victims of serious illness or injury. They often encounter patients infected with dangerous communicable diseases, including methicillin-resistant Staphylococcus aureus, Clostridium difficile, Neisseria meningitis, hepatitis B and C, HIV, and multidrug-resistant tuberculosis. To be prepared to safely manage such patients, paramedics and emergency medical technicians are trained, as part of their core curriculum, to implement standard and transmission-based precautions to prevent exposure to themselves or others. Across the United States, EMS transports of such patients occur regularly. However, the transport of the first patient with confirmed Ebola virus disease to the United States (Atlanta, GA), the transport of the first US-diagnosed Ebola virus disease case (Dallas, TX), and the first transmission of Ebola virus disease to health care workers in the United States understandably raised anxiety in the EMS community about the appropriate education and training, policies, and procedures, as well as supervision required, to be best prepared for transporting patients with serious communicable disease in the United States.

bola virus disease to health care workers in the United States understandably raised anxiety in the EMS community about the appropriate education and training, policies, and procedures, as well as supervision required, to be best prepared for transporting patients with serious communicable disease in the United States. EMS and other US health care professionals lack familiarity with Ebola virus disease,1 which is understandable. Although sporadic human outbreaks of Ebola virus disease have occurred since the illness was first recorded in 1976, the disease has been limited to the geographic area of its animal host, in rural parts of Africa. Although there have been reported cases of similar viral hemorrhagic fevers exported to other countries,2, 3 the occurrences have been few. Lapses in standard and transmission-based infection control practice, an essential foundation for the safe transport and management of patients with Ebola virus disease, are well documented.4, 5, 6, 7, 8 EMS crews do not consistently comply with such basic infection control practices as hand hygiene.4 More broadly, the compliance of EMS providers with standard infection control precautions and equipment disinfection has been described as suboptimal.5 Unannounced collections of environmental samples from ambulances have grown methicillin-resistant S aureus and other multidrug-resistant organisms.6, 7, 8

and hygiene.4 More broadly, the compliance of EMS providers with standard infection control precautions and equipment disinfection has been described as suboptimal.5 Unannounced collections of environmental samples from ambulances have grown methicillin-resistant S aureus and other multidrug-resistant organisms.6, 7, 8 In 2002, Emory University partnered with the Centers for Disease Control and Prevention to develop a capability for the evaluation and management of individuals with serious communicable disease. This partnership included Grady EMS for the development of a team with the requisite competencies necessary to transport and manage these patients in the out-of-hospital setting. In 2005, the University of Nebraska developed isolation capability for serious communicable diseases and initiated partnerships with EMS professionals to provide safe transport and management of patients admitted to the facility. The objectives of these hospital and out-of-hospital collaborations were to close education, training, and practice gaps to best facilitate the care for patients with serious communicable disease while ensuring the safety of the medics and the general public through meticulous implementation of infection control practices as recommended by CDC. The description of practices implemented by EMS teams in these communities for the transport of patients with confirmed Ebola virus disease is shared so that others might more readily implement these practices, policies, and procedures as applicable to their mission requirements and system design. A discussion about evaluation and transport of patients with relevant travel history and development of illness is also included.

h confirmed Ebola virus disease is shared so that others might more readily implement these practices, policies, and procedures as applicable to their mission requirements and system design. A discussion about evaluation and transport of patients with relevant travel history and development of illness is also included. Education and Training Team medics are provided education about serious communicable pathogens, their modes of transmission, the availability of vaccines, pre- and postexposure prophylaxis and treatment. The education most importantly emphasizes strict adherence to standard and transmission-based infection control practices (Figure 1 ). We believe that it is important for paramedics to not only understand what to do but also why. Understanding the nature of the illnesses they confront helps providers overcome apprehension and fear so they can render safe and effective care.Figure 1 Education and training. PPE, Personal protective equipment.

igure 1 ). We believe that it is important for paramedics to not only understand what to do but also why. Understanding the nature of the illnesses they confront helps providers overcome apprehension and fear so they can render safe and effective care.Figure 1 Education and training. PPE, Personal protective equipment. After initial instruction, team medics complete competency-based training, with special attention to the proper donning and doffing of a variety of personal protective equipment in the presence of a trained observer. Careful attention is paid to compliance because a seemingly minor lapse in technique can put health care workers at risk for infecting themselves or spreading dangerous pathogens to others. For example, a cluster of severe acute respiratory syndrome cases among protected health care workers in 2003 raised questions about health care worker familiarity with how best to remove personal protective equipment (PPE) without contaminating themselves.9 Developing and maintaining a high level of competency with the use of PPE is vital for workforce safety.10

atory syndrome cases among protected health care workers in 2003 raised questions about health care worker familiarity with how best to remove personal protective equipment (PPE) without contaminating themselves.9 Developing and maintaining a high level of competency with the use of PPE is vital for workforce safety.10 Vehicle Preparation Team medics learn to prepare the ambulance in advance to facilitate decontamination and disinfection after patient contact (Figure 2 ). They do this by separating the driver compartment from the passenger compartment such that the driver compartment is always considered clean. Medics who have made patient contact do not return to the driver compartment. The ventilation system in the driver compartment is turned on, with the fan set on high and in such a way as to not allow recirculation of air. The intent is to create a positive-pressure environment in the driver compartment to guard against aerosol. Although Ebola is known to be transmitted by direct contact with blood or infectious bodily fluids or droplets and not by aerosol, this simple technique is implemented as an extra measure of caution should an aerosol-producing procedure be required in the patient compartment. CDC has previously recommended this technique for the ground transportation of patients who pose a risk of disease transmission by aerosol.11 Figure 2 Ambulance preparation.

this simple technique is implemented as an extra measure of caution should an aerosol-producing procedure be required in the patient compartment. CDC has previously recommended this technique for the ground transportation of patients who pose a risk of disease transmission by aerosol.11 Figure 2 Ambulance preparation. Medics also envelop the interior of the patient compartment with impervious barriers to prevent contamination of surfaces difficult to clean and disinfect. This is especially important for patients who pose a high risk for sharing infectious bodily fluid through active bleeding, vomiting, or diarrhea. The stretcher is similarly protected with an impervious sheet (Figure 3, Figure 4 ).Figure 3 Grady EMS draping technique using One Tuff disposable impermeable sheeting 4′×15′. Figure 4 UNMC draping technique using 6-mil impermeable plastic. A full complement of medical gear is available but protected behind the impervious drapes. Medical gear can be made more readily accessible by sealing in a clear plastic bag within the patient compartment to facilitate rapid access if required. This gear may include a bag-valve-mask device, airway adjuncts, and equipment to facilitate intravenous access and volume resuscitation. A leakproof container should be available for collection of emesis.

accessible by sealing in a clear plastic bag within the patient compartment to facilitate rapid access if required. This gear may include a bag-valve-mask device, airway adjuncts, and equipment to facilitate intravenous access and volume resuscitation. A leakproof container should be available for collection of emesis. Patient Preparation Patients who can cooperate may be asked to wear an impervious suit to prevent exposure to sites of cutaneous bleeding, or an undergarment capable of collecting large volumes of diarrhea. Should the patient not be able to don an impervious suit, an impervious covering may be used for the same purpose: to limit contamination of environmental surfaces and minimize the potential for paramedic exposure. To limit spread of droplets to environmental surfaces inside and outside the ambulance, patients are asked to wear a surgical mask if tolerated (Figure 5, Figure 6 ). If a patient is vomiting and cannot tolerate a surgical mask, a leakproof container should be provided for the patient to assist in the collection of emesis and to limit contamination of environmental surfaces.Figure 5 Patient enveloped in impervious sheet. Figure 6 Patient preparation.

Patient Preparation Patients who can cooperate may be asked to wear an impervious suit to prevent exposure to sites of cutaneous bleeding, or an undergarment capable of collecting large volumes of diarrhea. Should the patient not be able to don an impervious suit, an impervious covering may be used for the same purpose: to limit contamination of environmental surfaces and minimize the potential for paramedic exposure. To limit spread of droplets to environmental surfaces inside and outside the ambulance, patients are asked to wear a surgical mask if tolerated (Figure 5, Figure 6 ). If a patient is vomiting and cannot tolerate a surgical mask, a leakproof container should be provided for the patient to assist in the collection of emesis and to limit contamination of environmental surfaces.Figure 5 Patient enveloped in impervious sheet. Figure 6 Patient preparation. Health Care Worker PPE Personal protective equipment must be donned, in the presence of a trained observer, to protect the medic from exposure to blood or infectious bodily fluid. CDC recommends contact and droplet precautions, to include aerosol protection should an aerosol-producing procedure be required. The transport team met the CDC’s standard by wearing a footed Tyvek (DuPont, Wilmington, DE) suit, gloves, and a hooded, powered, air-purifying respirator for eye and mucous membrane protection. We recognized that goggles and a surgical mask technically meet the requirement for mucous membrane protection from droplets and that the hooded, powered, air-purifying respirator was not strictly required. During years of operational experience in our work environment, however, the team found the hooded, powered, air-purifying respirator cooler and more comfortable to use, especially on an extended mission. The device also averted any issues with eyewear fogging and protected health care providers from inadvertent touching of the face. The Tyvek suit and hooded, powered, air-purifying respirator also afford excellent splash protection, a very important consideration when managing a patient in an ambulance compartment where vomit, diarrhea, and blood pose a serious risk of exposure. Our patients were all several days into the course of their Ebola virus disease illness, a time when episodic emesis and profound diarrhea are reported.12 And should the patient suddenly require an aerosol-producing procedure, such as open suctioning to prevent aspiration or emergency intubation, the medic is already afforded the appropriate respiratory protection. A properly fitted air-purifying respirator has also been successfully used. Other PPE ensembles that prevent exposure to skin and mucous membranes will also be effective if used properly. Selection of PPE should be in accordance with CDC guidelines and accommodate the condition of the patient, the anticipated mission requirements, and the work environment and be in concert with the requisite competencies to properly use the equipment.

to skin and mucous membranes will also be effective if used properly. Selection of PPE should be in accordance with CDC guidelines and accommodate the condition of the patient, the anticipated mission requirements, and the work environment and be in concert with the requisite competencies to properly use the equipment. Work Practices Policies and procedures are implemented to prevent exposure to blood and infectious bodily fluids (Figure 7 ).Figure 7 Personnel roles and protective posture. HCW, Health care worker.

to skin and mucous membranes will also be effective if used properly. Selection of PPE should be in accordance with CDC guidelines and accommodate the condition of the patient, the anticipated mission requirements, and the work environment and be in concert with the requisite competencies to properly use the equipment. Work Practices Policies and procedures are implemented to prevent exposure to blood and infectious bodily fluids (Figure 7 ).Figure 7 Personnel roles and protective posture. HCW, Health care worker. The transport team consists of 2 trained medics, a team leader, and an EMS physician. Patient contact is limited to the least number of persons required to provide care for the patient. If a patient is ambulatory, it may be possible for only 1 medic to make patient contact, leaving the second medic free to provide support and drive the ambulance. If the patient is nonambulatory or in need of more active management, both medics will make patient contact, and the supervisor will don PPE and move to drive the ambulance, with the EMS physician following. The team leader and EMS physician support the mission by managing communications and logistics with involved agencies (which may include law enforcement, airport operations, public health, and emergency management), managing safety, and providing clinical decisionmaking and direct medical control when required. This frees the treating medic to focus on patient assessment and adherence to sound infection control practice. Communications are facilitated by hands-free push-to-talk radios, which are worn inside the impervious suit, to protect against exposure to blood and bodily fluids. When communications are not encrypted, the transport team is careful to protect privileged health information.

herence to sound infection control practice. Communications are facilitated by hands-free push-to-talk radios, which are worn inside the impervious suit, to protect against exposure to blood and bodily fluids. When communications are not encrypted, the transport team is careful to protect privileged health information. Standard clinical care guidelines are modified to enhance safety by avoiding aerosol-producing procedures when possible and prohibiting sharps in a moving vehicle. The care team is capable of providing advanced life support and will administer supportive therapy as indicated by the patient’s condition and as guided by the EMS physician. Every precaution will be taken to provide for the safety of the care team, which may include stopping the vehicle if needed to ensure no breach in infection control processes. Although written guidelines are helpful, direct medical control, provided by the EMS physician or medical director on scene, can provide decision support.

aution will be taken to provide for the safety of the care team, which may include stopping the vehicle if needed to ensure no breach in infection control processes. Although written guidelines are helpful, direct medical control, provided by the EMS physician or medical director on scene, can provide decision support. Transition of the patient to the care of the isolation unit team is a function of robust planning and exercise. Patients with confirmed Ebola virus disease are brought directly into the isolation unit. A patient report and expected time of arrival are provided in advance for the waiting care team. Routes to the isolation unit are selected to be most direct and to limit any risk of exposure to environmental surfaces. Ambulatory patients may be walked into the isolation unit to facilitate the most direct route. Nonambulatory patients are conveyed by stretcher, using the most direct route available. Selected routes are secured to prevent unintended contact with unprotected staff, patients, or visitors. The isolation unit care team is waiting to receive the patient in the PPE ensemble appropriate for the patient’s condition. Hospital environmental services personnel are trained to disinfect environmental surfaces in the transit area that may have become contaminated.

tact with unprotected staff, patients, or visitors. The isolation unit care team is waiting to receive the patient in the PPE ensemble appropriate for the patient’s condition. Hospital environmental services personnel are trained to disinfect environmental surfaces in the transit area that may have become contaminated. Decontamination, Disinfection of the Ambulance, and Doffing of PPE Isolating the driver compartment and the application of impermeable barriers to keep interior surfaces of the ambulance clean facilitate decontamination and disinfection of the ambulance. All waste is double bagged and clearly marked a biohazard. The interior of the ambulance, the stretcher, any exposed equipment, and all exterior surfaces of the waste bags are disinfected with an Environmental Protection Agency-registered hospital-grade disinfectant appropriate for the suspected or known pathogen.13 A disinfectant effective against nonenveloped viruses, such as norovirus, will be effective against Ebola. Surfaces are disinfected by wiping; sprays are not used to avoid inadvertent splash or droplet generation and to prevent uncontrolled running of fluid into crevices. Special attention is given to ensuring the appropriate contact time for the selected disinfectant agent. To limit creation of multiple waste streams, all waste produced by the transport is managed by the hospital isolation unit (Figure 8 ).Figure 8 Decontamination and disinfection of ambulance.

of fluid into crevices. Special attention is given to ensuring the appropriate contact time for the selected disinfectant agent. To limit creation of multiple waste streams, all waste produced by the transport is managed by the hospital isolation unit (Figure 8 ).Figure 8 Decontamination and disinfection of ambulance. The team leader and EMS physician observe and supervise the proper disinfection of the ambulance, the collection of infectious waste, and the doffing of PPE to facilitate safety and ensure no violation of technique or breach in protocol. The doffing process may include wiping the exterior of the impermeable suit and hood with disinfectant wipes before removal to limit the likelihood of inadvertent exposure to infectious fluid. Any member of the team who is physically assisting with the removal of PPE is also, at a minimum, in protective garments that protect exposure of skin, clothing, and mucous membranes (Figure 9 ). Any breach of infection control procedure is immediately managed in accordance with CDC guidance. Any exposed skin is immediately washed with cleansing or antiseptic solution, mucous membranes are copiously irrigated, and supervisor and public health authority are immediately notified to facilitate rapid evaluation, consideration of postexposure prophylaxis options, and monitoring. Even in the absence of a recognized exposure, the transport team is monitored for subjective illness and fever for the duration of the incubation period (21 days in the case of Ebola) to ensure that any developing illness is recognized and swiftly evaluated.Figure 9 Doffing of PPE.

prophylaxis options, and monitoring. Even in the absence of a recognized exposure, the transport team is monitored for subjective illness and fever for the duration of the incubation period (21 days in the case of Ebola) to ensure that any developing illness is recognized and swiftly evaluated.Figure 9 Doffing of PPE. The transport of patients with confirmed Ebola virus disease to isolation units at Emory University Hospital and University of Nebraska Medical Center was guided by years of training and exercise. In our view, a team with the requisite competencies is best suited for scheduled transport of patients with confirmed or high risk of having Ebola virus disease. This is particularly relevant, given the active screening of travelers returning from Ebola-affected countries who may develop signs and symptoms of disease, necessitating their transportation for further evaluation and management. The procedures we describe are applicable to transport of patients with confirmed Ebola virus disease or considered at high risk of having the disease. This does not describe every traveler returning from an Ebola-affected country who develops signs of illness.

The transport of patients with confirmed Ebola virus disease to isolation units at Emory University Hospital and University of Nebraska Medical Center was guided by years of training and exercise. In our view, a team with the requisite competencies is best suited for scheduled transport of patients with confirmed or high risk of having Ebola virus disease. This is particularly relevant, given the active screening of travelers returning from Ebola-affected countries who may develop signs and symptoms of disease, necessitating their transportation for further evaluation and management. The procedures we describe are applicable to transport of patients with confirmed Ebola virus disease or considered at high risk of having the disease. This does not describe every traveler returning from an Ebola-affected country who develops signs of illness. Transport of Patients Under Investigation For Having Ebola Virus Disease A person under investigation is an individual who develops signs and symptoms of illness and also has some epidemiologic risk of exposure to Ebola virus disease in the preceding 21 days. The authors also have experience transporting patients who meet criteria for persons under investigation, and in these cases, a patient evaluation assists in selecting the appropriate procedures and PPE ensemble to provide for the safety of the transport team. With the assistance of public health officials, a history is obtained to determine what risk the patient has for exposure to Ebola virus disease.14 The risk may be high, as in the case of a needle stick from a needle contaminated with the blood of a confirmed Ebola patient, or low, as in the case of a patient who has traveled in a country with widespread Ebola activity, but without a known exposure. This epidemiologic assessment is then complemented by an assessment of the patient’s clinical condition and the risk that the health care worker could be exposed to bodily fluids. If the patient simply has fever, has no evidence of vomiting or diarrhea, and is lucid and conversational (the “dry” patient), the medic is considered to be at less risk for exposure to bodily fluids. In contrast, the patient with active vomiting and diarrhea (the “wet” patient) poses a greater risk of transmission to the medic simply because there is greater risk of exposure to infectious bodily fluid. For patients who pose a limited risk for exposure to bodily fluids, the selected PPE is less onerous. Medics apply 2 pairs of gloves, an impermeable gown, and a face shield or surgical mask and goggles for protection of mucous membranes. For patients posing a greater risk of exposure, a more protective PPE ensemble, as previously described, is recommended.15 The selection of PPE for the transporting EMS team is determined by assessment of patient condition and risk of exposure to bodily fluids.

gical mask and goggles for protection of mucous membranes. For patients posing a greater risk of exposure, a more protective PPE ensemble, as previously described, is recommended.15 The selection of PPE for the transporting EMS team is determined by assessment of patient condition and risk of exposure to bodily fluids. For the transport of patients with suspected Ebola virus disease, it may also be useful to recognize that those with the disease become more contagious as the course of their disease progresses.12 The patient in the later phase of disease, in which vomiting, diarrhea, and bleeding are present, is more likely to transmit the disease than the patient with fever alone early in the course of the illness (see Figure 10 for relevant considerations).16 Figure 10 Considerations when managing a patient with suspected or confirmed Ebola virus disease.

of disease, in which vomiting, diarrhea, and bleeding are present, is more likely to transmit the disease than the patient with fever alone early in the course of the illness (see Figure 10 for relevant considerations).16 Figure 10 Considerations when managing a patient with suspected or confirmed Ebola virus disease. The direction of travel of individuals from Ebola-affected countries to designated ports of entry in the United States and the active monitoring of returned travelers by public health officials has greatly eased the challenges faced by EMS personnel for management of persons under investigation. Fever and other signs of illness in recently returned travelers can now be quickly recognized and patients quickly referred to appropriate health centers by appropriate conveyance for evaluation of their illness. It may further reassure medics to know that travelers from Ebola-affected countries who develop generalized signs of illness are still more likely to have malaria, acute diarrheal illness, or other infectious disease than they are to have Ebola virus disease.17 While this indicates a need for timely patient evaluation, it should also put in context the risk posed to EMS by these patients. To date, more than 1,000 travelers returned from Ebola-affected countries have been monitored in Georgia, with fewer than 30 requiring evaluation as persons under investigation. No cases of Ebola virus disease have been diagnosed in this cohort.18, 19, 20, 21

also put in context the risk posed to EMS by these patients. To date, more than 1,000 travelers returned from Ebola-affected countries have been monitored in Georgia, with fewer than 30 requiring evaluation as persons under investigation. No cases of Ebola virus disease have been diagnosed in this cohort.18, 19, 20, 21 Conclusion EMS responders in every community must be prepared to respond to a caller with fever and a relevant travel history. Education and training are essential for implementation of effective standard and transmission-based infection control practice. They are also essential for EMS professionals to have the confidence to be comfortable in their work environment. Guidance for effective screening of patient contacts by 911 public safety answering points and EMS providers, as well as measures to facilitate appropriate infection control, continues to evolve.15

ractice. They are also essential for EMS professionals to have the confidence to be comfortable in their work environment. Guidance for effective screening of patient contacts by 911 public safety answering points and EMS providers, as well as measures to facilitate appropriate infection control, continues to evolve.15 The foundation of safe care for patients with confirmed or suspected serious communicable disease is effective infection control practice. This is not achieved by simply donning a particular PPE ensemble. It requires implementation of appropriate administrative policies, work practices, and environmental controls, accompanied by focused education, training, and supervision. CDC publishes guidelines to assist agencies with this task, but it inevitably evolves as risk points are identified and best practices emerge about the management of Ebola virus disease and other serious communicable diseases in the US health system. Although formal guidance may change, recognition of these accepted principles may help guide the EMS community to be best prepared to meet this ever-present challenge. Supervising editor: Gregory J. Moran, MD Funding and support: By Annals policy, all authors are required to disclose any and all commercial, financial, and other relationships in any way related to the subject of this article as per ICMJE conflict of interest guidelines (see www.icmje.org). The authors have stated that no such relationships exist. A podcast for this article is available at www.annemergmed.com.

Introduction The emergence of a novel strain of H1N1 influenza (“2009 H1N1”) in North America in spring 2009 caught public health authorities by surprise. Because many people were susceptible and a vaccine would not be available for many months, experts predicted that the virus could rapidly infect 30% to 50% of the population. In this article, we describe how a large number of clinical and public health experts came together to rapidly create, refine, and deploy the Strategy for Off-Site Rapid Triage (SORT), a set of clinical algorithms designed to help minimally trained health care workers and laypeople make informed decisions about care-seeking in the setting of a hypothetical influenza pandemic. Ultimately, SORT was adopted, with minor modifications, by the Centers for Disease Control and Prevention (CDC) and offered free to the public through interactive Web sites supported by the US Department of Health and Human Services (http://www.Flu.gov) and the Microsoft Corporation (http://www.H1N1ResponseCenter.com). The rapid emergence and spread of 2009 H1N1 precluded prospective evaluations of SORT before its deployment. Now that the immediate crisis is past, prospective studies are needed to affirm SORT's safety and utility and to further explore the feasibility of using the Web and other bidirectional technologies to reduce surge and collect real-time epidemiologic information about rapidly emerging public health threats.

deployment. Now that the immediate crisis is past, prospective studies are needed to affirm SORT's safety and utility and to further explore the feasibility of using the Web and other bidirectional technologies to reduce surge and collect real-time epidemiologic information about rapidly emerging public health threats. Background The 2009 H1N1 influenza pandemic is the first the world has faced in 40 years.1 Four previous flu pandemics have been studied in detail: 1889 to 1892, 1918 to 1920, 1957 to 1960, and 1968 to 1970. In each, localized, often mild outbreaks were followed some months later by massive waves of illness and mortality, often in younger age groups.2 Virulence varied widely; the 1918 to 1920 pandemic (also caused by an H1N1 strain) was by far the worst, killing 20 to 40 million people worldwide. The 1968 to 1970 pandemic was the mildest of the 4. But in each instance, the toll of deaths and illness substantially exceeded that of a typical flu season.3 Because pandemic monitoring was primarily focused on Asia, public health officials were caught by surprise when a novel strain of H1N1 influenza suddenly emerged in the Western Hemisphere at the end of a regular flu season. Initially, the high number of reported fatalities relative to hospitalized cases in Mexico suggested that the virus was highly lethal. This prompted Mexican health authorities to close schools, cancel public events, and impose other community disease mitigation strategies in hopes of limiting the outbreak's effects.

n. Initially, the high number of reported fatalities relative to hospitalized cases in Mexico suggested that the virus was highly lethal. This prompted Mexican health authorities to close schools, cancel public events, and impose other community disease mitigation strategies in hopes of limiting the outbreak's effects. As 2009 H1N1 quickly spread across the United States, epidemiologists determined that the virus was not particularly virulent.4 Unlike most influenza strains, it tended to spare the elderly, perhaps because many older Americans have some degree of preexisting immunity to H1N1 viruses. But the virus greatly affected children and young adults. Many who died in the first wave of illness had underlying medical problems, but some were previously healthy individuals.5, 6 Word of these deaths provoked widespread anxiety and prompted many to seek immediate medical attention, regardless of the nature or severity of their symptoms. In August 2009, the President's Council of Advisors on Science and Technology (PCAST) reported on preparations for 2009 H1N1 influenza.7 Although it was generally understood by that point that the virus was less dangerous than first feared, PCAST concluded that it still represented a serious public health threat. “The issue” PCAST warned, “is not that the [2009 H1N1] virus is more deadly than other flu strains, but rather that it is likely to infect more people than usual because it is a new strain against which few people have immunity. This could mean that doctors' offices and hospitals may get filled to capacity.”7

h threat. “The issue” PCAST warned, “is not that the [2009 H1N1] virus is more deadly than other flu strains, but rather that it is likely to infect more people than usual because it is a new strain against which few people have immunity. This could mean that doctors' offices and hospitals may get filled to capacity.”7 The PCAST predicted that 2009 H1N1 could infect 30% to 50% of the population and produce symptoms in 20% to 40%. If even half that number sought medical attention, the pandemic would send 30 to 60 million Americans into an already heavily burdened emergency care system.8 Early experience appeared to justify the council's concern. Between May 15 and June 15, more than 44,000 people with influenza-like illness had visited New York City emergency departments (EDs) compared with 4,267 the year before. Total ED visits increased by 32%, from 325,135 the previous year to more than 428,000.9 In some communities, ED use surged by 50% to 100%.10 Because the second wave of influenza pandemics is typically larger than the first, officials worried at the time that a sustained surge of ED visits could worsen already dangerous waits for emergency care,11 overload EDs, and promote iatrogenic transmission of influenza from individuals with flu to medically fragile patients nearby.

cond wave of influenza pandemics is typically larger than the first, officials worried at the time that a sustained surge of ED visits could worsen already dangerous waits for emergency care,11 overload EDs, and promote iatrogenic transmission of influenza from individuals with flu to medically fragile patients nearby. The Origin of SORT Four months before 2009 H1N1 emerged, the Emory University Department of Emergency Medicine and Emory Office of Critical Event Preparedness and Response in Atlanta, GA, convened an interdisciplinary expert group to devise a clinical algorithm for use in future influenza pandemics (Figure). Their goal was to create a simple but accurate tool that could help minimally trained health care workers screen large numbers of patients with influenza-like illness. The meeting was sponsored by the Georgia Division of Public Health as part of its pandemic influenza preparedness activity.12 Supplemental funding was provided by the Robert W. Woodruff Foundation and later by the de Beaumont Foundation. The group named the product of their deliberations Strategy for Offsite Rapid Triage, or SORT.Figure Participants* in the SORT development meeting: December 10 to 11, 2008.

influenza preparedness activity.12 Supplemental funding was provided by the Robert W. Woodruff Foundation and later by the de Beaumont Foundation. The group named the product of their deliberations Strategy for Offsite Rapid Triage, or SORT.Figure Participants* in the SORT development meeting: December 10 to 11, 2008. Before the meeting, 3 group members (A.P.I., B.D., T.C.B.; see acknowledgements) conducted an adjudicated literature search to identify clinical decision rules that are designed to identify patients who are severely ill with pneumonia or other conditions. Candidate clinical decision rules were presented to the full group. After a period of discussion, all agreed that the United Kingdom's CRB-65 (4 components of the score: Confusion, Respiratory rate greater than 24 breaths per minute, systolic Blood pressure less than 100 millimeters of mercury and age greater than or equal to 65 years) score was best suited to anchor the proposed influenza-like illness algorithm. The CRB-65 score was selected because it uses simple signs and symptoms, does not require diagnostic testing, and accurately quantifies illness severity in patients with pneumonia.13

ters of mercury and age greater than or equal to 65 years) score was best suited to anchor the proposed influenza-like illness algorithm. The CRB-65 score was selected because it uses simple signs and symptoms, does not require diagnostic testing, and accurately quantifies illness severity in patients with pneumonia.13 The group then developed an efficient, 3-step process to assess patients with influenza-like illness. In the first step, patients are screened to determine whether they meet CDC criteria for influenza-like illness. Those who do proceed to the second step, an assessment of illness severity using questions adopted from the CRB-65 score. Patients with influenza-like illness who have a CRB-65 score of 0 (suggesting relatively mild illness) move on to the third step, a short series of questions designed to determine whether they have a health condition that increases their risk of developing severe complications of influenza.

opted from the CRB-65 score. Patients with influenza-like illness who have a CRB-65 score of 0 (suggesting relatively mild illness) move on to the third step, a short series of questions designed to determine whether they have a health condition that increases their risk of developing severe complications of influenza. According to the patient's answers, SORT assigns a level of risk and recommends a specific action. Patients with “high-risk” influenza-like illness—in the group's first iteration of the algorithm, those with a CRB-65 score of 3 or more—would be sent directly to an ED. “Intermediate-risk” patients—CRB score of 1 or 2 or comorbid conditions that increase their risk of complications—would be advised to contact their physician or seek care in a walk-in clinic because early administration of antiviral medication might reduce the chance of complications. “Low-risk” patients—those with mild disease (CRB-65=0) and no comorbid conditions—would be advised to convalesce at home. SORT was initially envisioned for use by minimally trained health care workers at off-site flu assessment stations and walk-in clinics. But the development group quickly realized that a slightly modified version—one that substitutes symptoms for measured respiratory rate and blood pressure—could be used by call centers or even self-administered through an interactive Web site. Ultimately, both versions were included in the group's work product (Figure E1, available online at http://www.annemergmed.com).

ightly modified version—one that substitutes symptoms for measured respiratory rate and blood pressure—could be used by call centers or even self-administered through an interactive Web site. Ultimately, both versions were included in the group's work product (Figure E1, available online at http://www.annemergmed.com). Refinement and Validation When 2009 H1N1 influenza emerged 4 months later, we reexamined SORT 1.0 in light of what was being learned about the new virus.14, 15 According to early reports from the CDC and the World Health Association (WHO), we added several risk factors to our list: pregnant, morbidly obese (body mass index >40%), asthma, cystic fibrosis, coronary vascular disease, and neuromuscular or neurocognitive disease. We also decreased the threshold for being categorized as high risk from a CRB-65 score of 3 to 2 (Figure E2, available online at http://www.annemergmed.com). To assess the utility and safety of the revised algorithm, colleagues in New York City and Los Angeles agreed to retrospectively assess how SORT 2.0 would have performed had it been used to screen patients with influenza-like illness in their EDs earlier that spring. Because a primary objective of SORT is to identify patients with influenza-like illness who can safely recover at home, we were particularly interested in assessing its negative predictive value. These audits did not require individual identifiers, sought no sensitive information, involved no therapeutic interventions, and analyzed data in aggregate, so they were determined to be exempt by Emory's institutional review board.

er at home, we were particularly interested in assessing its negative predictive value. These audits did not require individual identifiers, sought no sensitive information, involved no therapeutic interventions, and analyzed data in aggregate, so they were determined to be exempt by Emory's institutional review board. Colleagues at the Montefiore ED reported that SORT 2.0 would have classified 60 of 102 influenza-like illness visits (60%) as low risk. Five of the low-risk patients were hospitalized, mainly for social reasons. (J. Gallagher, personal communication, August 11, 2009). At the New York University Langone Medical Center ED, SORT 2.0 classified 40 of 87 patients with influenza-like illness (47%) as low risk. One low-risk patient was admitted. Although he was afebrile, had normal vital signs, and had normal oxygenation, he was found to have pneumonia and a WBC count of 1,500 (I. Portelli and S. Smith, personal communication, August 18, 2009). At UCLA/Olive View Medical Center, 2 of 36 patients with low-risk influenza-like illness were hospitalized. Both had intractable vomiting (L. Baraff, personal communication, August 20, 2009).

on, he was found to have pneumonia and a WBC count of 1,500 (I. Portelli and S. Smith, personal communication, August 18, 2009). At UCLA/Olive View Medical Center, 2 of 36 patients with low-risk influenza-like illness were hospitalized. Both had intractable vomiting (L. Baraff, personal communication, August 20, 2009). According to these observations and additional epidemiologic data from the CDC and WHO, we added intractable vomiting, chest pain, and recrudescence as high-risk symptoms. To minimize the risk of false-negative results, we recalibrated the algorithm to send patients with even one high-risk symptom to the ED (Figure E3, available online at http://www.annemergmed.com). The Kaiser Permanente Colorado Institute for Health Research agreed to perform a retrospective assessment, using their health system's computerized records, to determine how well SORT 3.0 would have performed had it been used to screen patients with influenza-like illness. Between April 1 and June 30, 2009, 2,758 outpatients with influenza-like illness visited the Kaiser Permanente Colorado health system. SORT 3.0 categorized 1,540 of these encounters (56%) as low risk. During the next 2 weeks, 7 low-risk patients were hospitalized, but only 2 had problems that were related to the index visit (negative predictive value 99.9%). Intermediate-risk patients were much more likely to be admitted within 2 weeks than low-risk patients (odds ratio 11.9; 95% confidence interval 5.29 to 26.9) (D. Magid, personal communication, August 23, 2009).

hospitalized, but only 2 had problems that were related to the index visit (negative predictive value 99.9%). Intermediate-risk patients were much more likely to be admitted within 2 weeks than low-risk patients (odds ratio 11.9; 95% confidence interval 5.29 to 26.9) (D. Magid, personal communication, August 23, 2009). Buoyed by these findings, we developed a demonstration Web site with branching logic to depict how patients could use SORT to self-assess their need for care. To ensure that the site was comprehensible to laypeople, we asked experts in health literacy at our institution to translate SORT's clinical terms into plain language.16 More than 100 lay volunteers of widely varying age, race, and socioeconomic status reviewed draft text and offered suggestions on how to make the content understandable and actionable. Some had an influenza-like illness when they participated; others had recently recovered from the flu. On September 3 to 4, 2009, we presented draft adult and pediatric SORT algorithms and our demonstration Web site at a hastily convened Institute of Medicine workshop titled “Assessing the Severity of Influenza-Like Illnesses: Clinical Algorithms to Inform and Empower Health Care Professionals and the Public.”17 The event, which was sponsored by UnitedHealth Group, attracted national leaders from academia, major clinical societies, public health, law, government, and private industry. Feedback was highly favorable.

Influenza-Like Illnesses: Clinical Algorithms to Inform and Empower Health Care Professionals and the Public.”17 The event, which was sponsored by UnitedHealth Group, attracted national leaders from academia, major clinical societies, public health, law, government, and private industry. Feedback was highly favorable. Deployment From that point forward, the pace of events accelerated. On October 2, 2009, the CDC adopted a slightly modified version of SORT 3.0 and posted it on the agency's Web site at http://cdc.gov/h1n1flu/clinicians/pdf/adultalgorithm.pdf. In an accompanying disclaimer, the CDC stated that the algorithm was intended for use “by physicians and those working under their supervision.” It was also limited to patients older than 18 years. Five days later, US Department of Health and Human Services (HHS) secretary Katherine Sibelius announced the posting of an H1N1 self-evaluation application at http://www.Flu.gov. It closely adheres to the CDC's adult algorithm and used many of the terms and phrases we devised for our demonstration Web site. It is intended for use by adults older than 18 years. The same day (October 7, 2009), Microsoft Corporation unveiled its own flu self-assessment application at http://www.H1N1ResponseCenter.com. Like Flu.gov's application, Microsoft's site closely adheres to the CDC's adult algorithm and uses health-literate language licensed, at no charge, from Emory University. Both HHS and Microsoft encouraged health departments, nongovernmental organizations, private health plans, employers, and other organizations to link to their Web sites free. Many chose to do so.

losely adheres to the CDC's adult algorithm and uses health-literate language licensed, at no charge, from Emory University. Both HHS and Microsoft encouraged health departments, nongovernmental organizations, private health plans, employers, and other organizations to link to their Web sites free. Many chose to do so. Three major specialty societies endorsed elements of the effort. The American College of Emergency Physicians was the first major group to endorse SORT, on September 9 (N. Jouriles, personal communication). The American College of Physicians followed suit on October 19 (J. Stubbs, personal communication). The American Academy of Pediatrics worked directly with the CDC on a pediatric algorithm, which was posted on the CDC's Web site on October 16 (http://www.cdc.gov/h1n1flu/clinicians/pdf/childalgorithm.pdf). As soon as this pediatric algorithm was posted, we began drafting health-literate content to offer the guidance directly to the public through the Web. Unfortunately, the American Academy of Pediatrics opposed this effort because the algorithm was not prospectively validated. Concerns were also expressed that an interactive pediatric Web site might discourage some parents from contacting their child's medical provider.

guidance directly to the public through the Web. Unfortunately, the American Academy of Pediatrics opposed this effort because the algorithm was not prospectively validated. Concerns were also expressed that an interactive pediatric Web site might discourage some parents from contacting their child's medical provider. Notwithstanding this disappointment, the overall effort to create, test, and deploy SORT was highly collaborative from beginning to end. Numerous organizations and individuals gave freely of their time and expertise (see acknowledgements). Recognizing the urgency of the effort, Emory's Office of Technology Transfer readily licensed the technology, at no charge, to any vendor who agreed to provide it free. Initial Experience Between October 7, 2009, and February 24, 2010, Flu.gov recorded 721,906 total page views, 320,333 visits to Flu.gov/evaluation (the opening page of the self-evaluation site), and 230,761 completed evaluations to flu.gov/evaluation/index2.html (A. Roszak, personal communication). To reassure the public that the federal government would respect each user's privacy, HHS did not retain data on site visitors. As a consequence, we have no additional information.

page of the self-evaluation site), and 230,761 completed evaluations to flu.gov/evaluation/index2.html (A. Roszak, personal communication). To reassure the public that the federal government would respect each user's privacy, HHS did not retain data on site visitors. As a consequence, we have no additional information. Between October 5 and December 13, 2009, Microsoft's Web site, http://www.H1N1ResponseCenter.com, was visited 1.6 million times. Of the 442,000 visitors (28%) who completed a self-assessment, slightly less than half (N=202,000) chose to share anonymous data with the site. Preliminary analysis indicates that 37% of these visitors provided answers that categorized them as high risk and 13% were too young to receive guidance. The other half either did not meet influenza-like illness criteria or were assessed as not requiring ED treatment. Microsoft did not identify visitors who used the site multiple times, so it is possible that some individuals repeatedly entered positive replies. Limitations Despite our best efforts, SORT was not deployed until early October 2009, when the pandemic's peak had nearly passed. The CDC did not actively promote the static algorithms posted on its Web site. HHS and Microsoft did not aggressively market their interactive, self-assessment Web sites. News coverage did, however, generate some “free media.”

, SORT was not deployed until early October 2009, when the pandemic's peak had nearly passed. The CDC did not actively promote the static algorithms posted on its Web site. HHS and Microsoft did not aggressively market their interactive, self-assessment Web sites. News coverage did, however, generate some “free media.” Beyond counting Web hits, we cannot determine SORT's effect on care-seeking and patient safety. Although we received no reports of adverse events associated with use of SORT, we cannot exclude the possibility that some patients delayed care and experienced harm as a result. It is also possible that SORT was so conservative that it encouraged some patients to make an unnecessary trip to the ED. Prospective evaluations, including criterion-standard comparisons and follow-up studies, are needed to fully assess SORT's utility and safety.

patients delayed care and experienced harm as a result. It is also possible that SORT was so conservative that it encouraged some patients to make an unnecessary trip to the ED. Prospective evaluations, including criterion-standard comparisons and follow-up studies, are needed to fully assess SORT's utility and safety. Discussion As this article goes to press, 2009 H1N1 appears to be the mildest flu pandemic on record, at least in the United States. The CDC estimates that between April 2009 and January 16, 2010, between 8,330 and 17,160 deaths related to 2009 H1N1 occurred. The midlevel estimate is 11,690 deaths.18 Although any death is tragic, this toll is less than that associated with a typical year of seasonal flu, much less a pandemic year.2, 3, 4 2009 H1N1 did, however, cause considerable morbidity. According to the CDC, between 183,000 and 378,000 H1N1-related hospitalizations occurred during the same interval. The midlevel estimate is 257,000 hospitalizations. The CDC projects that between 41 and 84 million Americans were infected with 2009 H1N1.17 The midlevel estimate is 57 million, close to the lower boundary of 60 million predicted by the PCAST in August 2009.7 Influenza is a notoriously unpredictable virus. 2009 H1N1 may remain relatively mild or it could mutate in the coming months to become more virulent. Continued vigilance is warranted.

Discussion As this article goes to press, 2009 H1N1 appears to be the mildest flu pandemic on record, at least in the United States. The CDC estimates that between April 2009 and January 16, 2010, between 8,330 and 17,160 deaths related to 2009 H1N1 occurred. The midlevel estimate is 11,690 deaths.18 Although any death is tragic, this toll is less than that associated with a typical year of seasonal flu, much less a pandemic year.2, 3, 4 2009 H1N1 did, however, cause considerable morbidity. According to the CDC, between 183,000 and 378,000 H1N1-related hospitalizations occurred during the same interval. The midlevel estimate is 257,000 hospitalizations. The CDC projects that between 41 and 84 million Americans were infected with 2009 H1N1.17 The midlevel estimate is 57 million, close to the lower boundary of 60 million predicted by the PCAST in August 2009.7 Influenza is a notoriously unpredictable virus. 2009 H1N1 may remain relatively mild or it could mutate in the coming months to become more virulent. Continued vigilance is warranted. Our experience illustrates both the opportunities and challenges of Web-based self-evaluation to limit a health system surge. In less than 12 months, a large, diverse, and highly motivated group of experts representing multiple disciplines took SORT from a hypothetical concept to reality. With the help of the US government and one of America's top technology companies, 2 static algorithms and 2 interactive Web sites were deployed nationwide. Originally envisioned to help health care workers make informed decisions about where to direct patients for treatment of influenza-like illness, SORT rapidly evolved into a decision support tool for the general public. We hoped to safely and responsibly reduce a surge of visits to EDs and other health care facilities while encouraging those who truly need care to receive it.

ormed decisions about where to direct patients for treatment of influenza-like illness, SORT rapidly evolved into a decision support tool for the general public. We hoped to safely and responsibly reduce a surge of visits to EDs and other health care facilities while encouraging those who truly need care to receive it. The Web sites we helped devise were used approximately 650,000 times. We have no way to determine how many times the CDC's adult and pediatric algorithms were used by clinicians and call centers. No adverse events were reported. Microsoft's data suggests that their Web site may have prevented as many as 100,000 ED visits, although the true total is probably less. Because HHS did not record data on visitors to Flu.gov, we cannot estimate the effect of their self-assessment tool. Several challenges remain. Because of the rapid emergence of 2009 H1N1, we were unable to follow the lengthy timetable normally required to methodically develop and validate a new clinical decision rule. We attempted to compensate for this fact by anchoring SORT in a previously validated clinical decision rule.

The Web sites we helped devise were used approximately 650,000 times. We have no way to determine how many times the CDC's adult and pediatric algorithms were used by clinicians and call centers. No adverse events were reported. Microsoft's data suggests that their Web site may have prevented as many as 100,000 ED visits, although the true total is probably less. Because HHS did not record data on visitors to Flu.gov, we cannot estimate the effect of their self-assessment tool. Several challenges remain. Because of the rapid emergence of 2009 H1N1, we were unable to follow the lengthy timetable normally required to methodically develop and validate a new clinical decision rule. We attempted to compensate for this fact by anchoring SORT in a previously validated clinical decision rule. Some clinicians may question the wisdom of providing decision support directly to the public. The legal and political risks of misclassifying severely ill patients compelled us to make each generation of SORT more conservative than the last. The resulting increase in sensitivity came, no doubt, at the expense of specificity, which probably limited SORT's effect on surge. Microsoft's data suggest that their current version overtriages patients to some degree. Prospective research is needed to determine whether SORT and other Web-based tools can be made more precise without compromising safety.

came, no doubt, at the expense of specificity, which probably limited SORT's effect on surge. Microsoft's data suggest that their current version overtriages patients to some degree. Prospective research is needed to determine whether SORT and other Web-based tools can be made more precise without compromising safety. Many individuals and organizations contributed time and effort to this project. However, future efforts to address the threat of health system surge should not depend on good will alone. Public health preparedness warrants the same level of societal commitment afforded to vaccine development and other disease control efforts. Now that the immediate crisis is past, SORT should be prospectively evaluated to confirm its safety and effectiveness. Given 2009 H1N1's effect on children, it will be particularly important to assess SORT in pediatric age groups. One way to do this is to ask the parents of children with flu-like symptoms to complete a Web-based assessment immediately before their child is examined by a clinician. The 2 assessments can then be compared, with the clinical examination as the criterion standard. Existing research networks, such as the CDC's EMERGEncy ID Net19 or the Health Resources and Services Administration Pediatric Emergency Care Applied Research Network,20 are well suited to this task.

xamined by a clinician. The 2 assessments can then be compared, with the clinical examination as the criterion standard. Existing research networks, such as the CDC's EMERGEncy ID Net19 or the Health Resources and Services Administration Pediatric Emergency Care Applied Research Network,20 are well suited to this task. SORT is but one example of how information technology may be used to empower patients to make prudent decisions about their health. Although SORT is designed to assess patients with influenza-like illness, the 3-step approach it uses (screening, severity assessment, associated risk factors) may be used to evaluate many illnesses. SORT-like algorithms for selected public health threats such as severe acute respiratory syndrome could be even be prepared and evaluated in advance and deployed if needed. This method could help reassure a nervous public, particularly in the early phases of an outbreak when many people otherwise rush to the nearest ED.

ORT-like algorithms for selected public health threats such as severe acute respiratory syndrome could be even be prepared and evaluated in advance and deployed if needed. This method could help reassure a nervous public, particularly in the early phases of an outbreak when many people otherwise rush to the nearest ED. With additional refinement, Web-based decision-support tools such as SORT may be used to collect important epidemiologic information about disease incidence and severity in nonhospitalized individuals. Information of this type is vital to quickly characterize a new disease's attack rate and virulence. Initial assessment of the 2009 H1N1 outbreak in Mexico was hindered by lack of knowledge about the number and distribution of nonhospitalized cases. As a result, Mexican officials did not know whether they were dealing with the early stages of a highly lethal outbreak (what was initially feared) or an already established epidemic with relatively few deaths (the picture that subsequently emerged in the United States). Someday, bidirectional technologies such as the World Wide Web and short messaging service21 may be used to generate “epidemic science in real time.”22

lethal outbreak (what was initially feared) or an already established epidemic with relatively few deaths (the picture that subsequently emerged in the United States). Someday, bidirectional technologies such as the World Wide Web and short messaging service21 may be used to generate “epidemic science in real time.”22 Conclusion Advances in biomedical science have given us new vaccines, antimicrobials, and diagnostic tests. Information technology has progressed at an equal, if not greater, pace, but it has not been exploited to the same degree. SORT is but one example of how the Web and other bidirectional technologies may be used to inform and empower the public to make prudent health care decisions. In the future, information technology may revolutionize public health preparedness by providing new and powerful ways to detect, contain, and mitigate public health threats. Appendix Figure E1 SORT version 1.0. Figure E2 SORT version 2.0. Figure E3 SORT version 3.0.

Conclusion Advances in biomedical science have given us new vaccines, antimicrobials, and diagnostic tests. Information technology has progressed at an equal, if not greater, pace, but it has not been exploited to the same degree. SORT is but one example of how the Web and other bidirectional technologies may be used to inform and empower the public to make prudent health care decisions. In the future, information technology may revolutionize public health preparedness by providing new and powerful ways to detect, contain, and mitigate public health threats. Appendix Figure E1 SORT version 1.0. Figure E2 SORT version 2.0. Figure E3 SORT version 3.0. Numerous individuals and organizations contributed to this effort. Although their listing here does not constitute their personal endorsement of this article or every aspect of the project, we are nonetheless indebted to them for their contributions and wisdom. Contributors include, but are not limited to, Larry Baraff, MD, T. Christopher Bond, PhD, Holley Butkovich, Lorenzo DiFrancesco, MD, Barry Diner, MD, Alexandra Drane, Harvey Fineberg, MD, John Gallagher, MD, Lewis Goldfrank, MD, Lynn R. Goldman, PhD, Michael Handrigan, Kara Jacobsen, MA, John A. Jernigan, MD, Nicholas Jouriles, MD, Rebecca Katz, PhD, Priscilla Keith, JD, Naghma Khan, MD, Steve Krug, MD, Adam Landman, MD, Hal C. Lawrence, MD, Linda Lawrence, MD, Weimin Lu, MBA, Nicole Lurie, MD, Gene W. Matthews, MD, Meg McCoy, JD, John A. Mitas, MD, Arnold S. Monto, MD, Robert Nadolski, Kristine Nash-Wong, Andy Pope, PhD, Ian Portelli, MD, Lynne Richardson, MD, Andrew Roszak, JD, EMT-P, Jeffrey Runge, MD, Judith Salerno, PhD, Harold Simon, MD, Joseph Stubbs, MD, Todd Taylor, MD, Reed Tuckson, MD, Silas Smith, MD, and Charlotte Yeh, MD.

cCoy, JD, John A. Mitas, MD, Arnold S. Monto, MD, Robert Nadolski, Kristine Nash-Wong, Andy Pope, PhD, Ian Portelli, MD, Lynne Richardson, MD, Andrew Roszak, JD, EMT-P, Jeffrey Runge, MD, Judith Salerno, PhD, Harold Simon, MD, Joseph Stubbs, MD, Todd Taylor, MD, Reed Tuckson, MD, Silas Smith, MD, and Charlotte Yeh, MD. Supervising editor: Amy H. Kaji, MD, PhD Funding and support: By Annals policy, all authors are required to disclose any and all commercial, financial, and other relationships in any way related to the subject of this article that might create any potential conflict of interest. See the Manuscript Submission Agreement in this issue for examples of specific conflicts covered by this statement. This work was supported in part by a contract from the Georgia Division of Public Health and grants from the Robert W. Woodruff Foundation and the de Beaumont Foundation. Publication date: Available online June 3, 2010.