Investigating Rare Events and a Discussion on Waiting Room Chest Pain Evaluations: Answers to the November 2010 Journal Club Questions

Abstract

Discussion Points1This abstract's conclusion states, “An organized approach to triage and waiting room evaluation for stable chest pain patients is safe and efficient. Although waiting room evaluation is not ideal, it may be a viable contingency strategy for periods when ED crowding or volume surges lead to compromised access and delays to stretcher placement.”1Scheuermeyer F. Christenson J. Innes G. et al.Safety of assessment of patients with potential ischemic chest pain in an emergency department waiting room: a prospective comparative cohort study.Ann Emerg Med. 2010; 56: 455-462Abstract Full Text Full Text PDF PubMed Scopus (16) Google Scholar ANearly 8% of patients managed in the waiting room were diagnosed with acute coronary syndrome. List the pros and cons of treating patients with “chest pain of potential cardiac origin” in the emergency department (ED) waiting room, making separate lists for each of the following perspectives: patient, emergency physician, hospital, regional health care system.BThe Massachusetts experience suggests that ED visits will increase after the implementation of federal health care reform. The higher patient volume may result in additional ED crowding and longer waiting room times. Should treating patients in ED hallways and waiting room be an accepted practice norm? What other innovative approaches have EDs tried to deal with crowding?CHow might these ED strategies to mitigate crowding affect trainees' education and ED nurse satisfaction and retention?2A. Fifty-year-old twins, identical in all biological and lifestyle respects, present to your ED at different times with identical chest pain symptoms. One arrives at a quiet time and is placed in a standard ED examination room. The other arrives at a busy time and is triaged to waiting room care. List what is likely done for each patient. BIf you managed the waiting room patient in the manner reported in this article, do you think you would pass the American Board of Emergency Medicine (ABEM) oral examination? Do you think the Board would have to adjust the required critical actions for the waiting room setting? What important elements of the history and physical examination might be difficult to ascertain and perform in the waiting room? Would you deviate from your center's standard ED chest pain protocol?3Imagine a condition for which patients generally do well but, according to national averages, in which 1 in 1,000 will die or have a bad complication. Now imagine 2 similar neighboring EDs with patient populations of identical demographic makeup. Each ED treats 2,000 patients with this condition each year. The hospitals use different strategies to manage this condition. ED A's strategy reduces the risk of a bad outcome to 1 in 2,000, half of the national average. ED B's approach produces 1 bad event per 500 cases. AOn average, how many bad events can we expect at each hospital each year?BWhat is the probability that ED B will have the same number of events or fewer events than ED A in any given year? Same question if quality levels were such that ED A expected 1 event per year and ED B 2? What about ED A having 2 events and ED B having 4?CHow many years of observation would be required (guess if you must) to prove with 95% confidence (assuming that the patient populations and hospitals were similar in all other respects and did not change during the length of the observation period) that ED A was indeed safer than ED B for this condition, assuming that rates are as stated in the original question?4A. In his editorial, Dr. Wears cautions that in circumstances involving rare events, it is easy to be seduced into thinking a system is safe when it is not. Discuss his contention in light of your answers to question 3. BDr. Schriger points out, however, that the converse could also be true. He opines that if risk-averse physicians develop an overly conservative practice style, even a prolonged period of event-free activity may be insufficient to change practice to a much less expensive but still adequately safe style. Again, in light of your answers to question 3, discuss whether we can expect to gain scientific proof that one strategy is safer than another.5Dr. Kellermann laments that regardless of what chest pain rule-out system is correct, it is a shame that change occurs because of crowding and other political forces instead of a rational process based on science. He fears that emergency physicians, by cleverly developing workarounds to untenable situations, allow the fundamental problems to perpetuate. AConsidering all of these perspectives, have a discussion about how a community might best determine what method for handling soft “rule-out acute coronary syndrome” patients is best.Answer 1Q1. This abstract's conclusion states, “An organized approach to triage and waiting room evaluation for stable chest pain patients is safe and efficient. Although waiting room evaluation is not ideal, it may be a viable contingency strategy for periods when ED crowding or volume surges lead to compromised access and delays to stretcher placement.”1Scheuermeyer F. Christenson J. Innes G. et al.Safety of assessment of patients with potential ischemic chest pain in an emergency department waiting room: a prospective comparative cohort study.Ann Emerg Med. 2010; 56: 455-462Abstract Full Text Full Text PDF PubMed Scopus (16) Google ScholarQ1.a Nearly 8% of patients managed in the waiting room were diagnosed with acute coronary syndrome. List the pros and cons of treating patients with “chest pain of potential cardiac origin” in the emergency department (ED) waiting room, making separate lists for each of the following perspectives: patient, emergency physician, hospital, regional health care system.We created Table 1 which is neither comprehensive nor authoritative. Was yours similar or were there items that we called “pro” that you called “con” or vice versa?Table 1Pros and cons of waiting room care.PerspectiveProsConsPatientShorter time to physician evaluationEarlier receipt of aspirin and/or analgesiaPotential for improved functional outcome by earlier identification and treatment of evolving STEMI or NSTEMIFaster disposition if chest pain determined to be of benign originLack of privacyNoise and chaos of a crowded waiting roomConcern for infection transmission in crowded WRMay have to sit in a chair while tethered to a portable monitor or intravenous fluid poleMay not be able to receive adequate analgesia (eg, intravenous narcotic or nitroglycerine) in WRLess intensive monitoring and reassessmentsPotential for physicians to erroneously minimize importance of patient's complaints when patients is sitting in WRTotal ED length of stay may be prolonged compared with that of patients treated primarily in ED2Russ S. Jones I. Aronsky D. et al.Placing physician orders at triage: the effect of length of stay.Ann Emerg Med. 2010; 56: 27-33Abstract Full Text Full Text PDF PubMed Scopus (38) Google ScholarRisk of patients with peripheral IV leaving with IV in place, resulting in morbidityPotential for unrecognized sudden deathEmergency physicianPeace of mind that patients with potentially serious causes of chest pain have been screened and are not “ticking time bombs” in a WR chairEarlier identification of patients with chest pain who have a STEMI or have positive markersAdditional space to treat patients when department is fullFaster evaluation and ED discharges for patients with benign causes of chest painMore patients treated per hour and hence more potential revenueMay not obtain complete or true history because of lack of privacy and patient reluctance to disclose pertinent informationMay not be able to perform complete physical examination (eg, rectal examination for occult bleeding)Limited treatment options for patient at increased risk for eloping or less intensive nurse monitoringPatients angry about staying in WR may downgrade physician on satisfaction surveysCareer dissatisfaction caused by work conditionsIncreased malpractice liability if patients in the waiting room have untoward outcomesHospitalMore rapid throughputIncrease rate of meeting federal metrics for chest painIncrease revenues without hospital expenditure on expanding additional ED, observation unit, or inpatient spaceReserve limited ED rooms for the most critically ill patientsAnother short-term fix so that hospitals can continue to board admitted patients in ED and use hospital beds for more profitable insured elective surgery patientsFewer left without being seen (LWBS)Theoretically decrease likelihood of patient decompensation in WR and potential malpractice eventsPotential for lower patient satisfaction scores and increased patient complaintsHospital internists and cardiologists are upset when their private patients are treated in the WRLower patient satisfaction, leading to loss of future business for the hospitalPotential for increased malpractice payouts if a missed myocardial infarction or pulmonary embolism occurs because of limited testing in WR patientsIncreased physician and nurse attrition because of unpopular work conditionsPotential HIPAA violationsRegional health care systemEncourage hospitals with less crowding to publicize their ED's availability, lack of use of the WR, and short WR times to improve the distribution of patients to EDs across the regionWR evaluations may reduce interventions for social issues such as substance abuse, smoking, and domestic violence counselingPatients disappointed with WR dispositions may avoid seeking medical care for chronic medical conditionsSTEMI, ST segment elevation MI; NSTEMI, non ST segment elevation MI; WR, waiting room; IV, intravenous catheter; LWBS, left without being seen; HIPAA, Health Insurance Portability and Accountability Act; RVU, relative value unit. Open table in a new tab Q1.b The Massachusetts experience suggests that ED visits will increase after the implementation of federal health care reform. The higher patient volume may result in additional ED crowding and longer waiting room times. Should treating patients in ED hallways and waiting room be an accepted practice norm? What other innovative approaches have EDs tried to deal with crowding?The short answer to this question is no. Hallway and waiting room medicine is our specialty's Band-Aid for ED crowding and a dysfunctional health care system. Treatment anywhere outside of a standard ED treatment room or bay is not an optimal patient experience for the reasons stated in answer 1A.As Dr. Kellerman points out in his editorial, our celebration of the dedication and ingenuity of the thousands of emergency physicians who develop ways to compensate for crowding should be tempered by the recognition that such efforts, by allowing the status quo to persist, may impede real reform.3Kellermann A.L. Waiting room medicine: has it really come to this?.Ann Emerg Med. 2010; 56: 468-471Abstract Full Text Full Text PDF PubMed Scopus (5) Google Scholar Do you think the Scheuermeyer group should be praised or castigated?Hospitals' reactions to crowding have ranged from laissez faire neglect to systems reengineering. Many hospitals have incorporated “team triage” strategies that aim to jump-start ED evaluations in triage and discharge low-acuity patients directly from triage.2Russ S. Jones I. Aronsky D. et al.Placing physician orders at triage: the effect of length of stay.Ann Emerg Med. 2010; 56: 27-33Abstract Full Text Full Text PDF PubMed Scopus (38) Google Scholar Emergency physicians and midlevel providers now treat patients in hallways, waiting rooms, triage rooms, converted conference rooms, radiology overflow areas, and even drive-through flu stations in hospital parking lots.2Russ S. Jones I. Aronsky D. et al.Placing physician orders at triage: the effect of length of stay.Ann Emerg Med. 2010; 56: 27-33Abstract Full Text Full Text PDF PubMed Scopus (38) Google Scholar, 4Cruz A.T. Patel B. DiStefano M.C. et al.Outside the box and into thick air: implementation of an exterior mobile pediatric emergency response team for North American H1N1 (swine) influenza virus in Houston, Texas.Ann Emerg Med. 2010; 55: 23-31Abstract Full Text Full Text PDF PubMed Scopus (29) Google Scholar Patients are shuffled from core ED rooms to waiting rooms, hallways, or observation units in an attempt to keep key ED beds available.Solutions to ED crowding will not come from the ED alone but require institutional cooperation in controlling both the inflow (how are elective admissions and ED admissions prioritized) and outflow (are discharges performed continuously or only after morning rounds) of admitted patients.Until the federal government and health care payers penalize hospital administrators for ED crowding, ED physicians and staff will be forced to keep reaching into our “bag of tricks.”Q1.c How might these ED strategies to mitigate crowding affect trainees' education and ED nurse satisfaction and retention?ED crowding has been shown to harm patients,5Government Accountability OfficeHospital Emergency Departments: Crowding Continues to Occur, and Some Patients Wait Longer Than Recommended Time Frames. Government Accountability Office, Washington, DC2009Google Scholar, 6Pines J.M. Localio A.R. Hollander J.E. et al.The impact of emergency department crowding measures on time to antibiotics for patients with community-acquired pneumonia.Ann Emerg Med. 2007; 50: 510-516Abstract Full Text Full Text PDF PubMed Scopus (223) Google Scholar, 7Fee C. Weber E.J. Maak C.A. et al.Effect of emergency department crowding on time to antibiotics in patients admitted with community acquired pneumonia.Ann Emerg Med. 2007; 50: 501-509Abstract Full Text Full Text PDF PubMed Scopus (202) Google Scholar, 8Hwang U. Richardson L.D. Sonuyi T.O. et al.The effect of emergency department crowding on the management of pain in older adults with hip fracture.J Am Geriatr Soc. 2006; 54: 270-275Crossref PubMed Scopus (198) Google Scholar but let's consider its effect on resident education. Residents train predominantly at large, urban, tertiary care medical centers that are inundated with patients. Crowding reduces the number of available ED rooms, thereby decreasing turnover and the number of new patients available to house staff for evaluation and treatment. Boarding of admitted patients often causes ED diversion and the redirection of ill patients to other facilities.9McConnell K.J. Richards C.F. Daya M. et al.Ambulance diversion and lost revenues.Ann Emerg Med. 2006; 48: 702-710Abstract Full Text Full Text PDF PubMed Scopus (51) Google Scholar Residents treat fewer new patients per shift and miss out on the high-acuity patients diverted to other EDs.Team triage strategies, although beneficial in expediting patient care, might also harm resident education. Orders placed by attending physicians in triage compromise resident education by focusing care before the resident has had a chance to conceptualize the case. For example, a 50-year-old woman who presents with general weakness, weight gain, and a pulse rate of 50 beats/min might be far more valuable to residents' education than the same patient with a thyroid stimulating hormone result of 30μ IU/mL, which was ordered by the attending physician at triage some 3 hours ago, clipped to the front of the chart. Furthermore, triage physicians often manage and discharge low-acuity patients without ED resident participation, depriving them of simple but important procedures such as splinter removal and paronychia drainage.Treating patients in nonstandard ED locations may also foster bad habits such as failing to ask sensitive questions or failing to undress and fully examine patients because of privacy concerns. At the other extreme, trainees may learn that it is all right to violate a patient's sense of privacy in the name of expedience.Emergency nurses, like physicians, work in the ED because they enjoy the fast-paced, semichaotic environment in which they have the opportunity to resuscitate and manage sick patients. Boarding turns ED nurses into ICU and ward nurses. The medical and trauma resuscitations are replaced with pages of inpatient orders awaiting completion. Patients and their families, understandably upset about their prolonged ED stays, often vent their discontent on the ED nurses. This crowding and redefinition of the emergency nurse duties has resulted in lower job satisfaction and increased attrition rates.10Gooch P. ED overcrowding.Implications for nurse leaders. Nurs Manage. 2009; 40: 50-54Crossref PubMed Scopus (2) Google ScholarAnswer 2Q2.a Fifty-year-old twins, identical in all biological and lifestyle respects, present to your ED at different times with identical chest pain symptoms. One arrives at a quiet time and is placed in a standard ED examination room. The other arrives at a busy time and is triaged to waiting room care. List what is likely done for each patient.We were looking for a chart like this one (Table 2). The individual answers are based on some combination of the information in the Scheuermeyer article, our experience, and common sense.1Scheuermeyer F. Christenson J. Innes G. et al.Safety of assessment of patients with potential ischemic chest pain in an emergency department waiting room: a prospective comparative cohort study.Ann Emerg Med. 2010; 56: 455-462Abstract Full Text Full Text PDF PubMed Scopus (16) Google ScholarTable 2A comparison of care in ED and ED waiting room settings.ActionED Room EvaluationWaiting Room EvaluationECG on arrivalYesYesSerial ECGs at frequent intervalsMore likelyLess likelyAspirin administrationYesYesNitroglycerin administrationMore likely if indicatedLess likely without cardiac and blood pressure monitoringCardiac monitoringYesLess likelySupplemental oxygenYesUnlikelyClose nursing supervisionMore likelyLess likelyInitial cardiac biomarker level results sentYesYesRepeated set of biomarker level results sentMore likelyLess likelyComplete history, including questions about illicit drug useYesLess likely to be asked or answered honestly in nonprivate settingComplete physical examination performedMore likely to be conducted and include a rectal examination for occult blood when indicatedLess likely that a complete examination is performed in the waiting room. Potential for critical elements to be skippedProvocative cardiac testingMore likelyLess likelyCardiology consultationMore likelyLess likelyDetection of brief arrhythmiasMore likelyLess likely, especially if patient does not receive a portable cardiac monitor Open table in a new tab Q2.b If you managed the waiting room patient in the manner reported in this article, do you think you would pass the American Board of Emergency Medicine (ABEM) oral examination? Do you think the Board would have to adjust the required critical actions for the waiting room setting? ⁎This Journal Club's authors do not serve as ABEM board examiners and have no insider knowledge about the scoring system for that examination. This answer reflects our opinions and not ABEM's position on this topic. What important elements of the history and physical examination might be difficult to ascertain and perform in the waiting room? Would you deviate from your center's standard ED chest pain protocol?We do not anticipate that ABEM would present an oral board examination case that required waiting room management. However, this hypothetical scenario provides an opportunity to examine whether the actions that might be considered critical on an ABEM examination can be accomplished in the waiting room. According to the analysis in the answer to question 2A, it appears that many of the medical critical actions, including rapid ECG testing, review for ST-segment elevation, and aspirin administration, can be accomplished in the waiting room. Other potential critical actions, however, such as maintaining patient privacy and dignity while performing a complete history and physical examination, might be impossible to fulfill in the waiting room. Whether an examinee passes or fails the examination would ultimately depend on the extent to which ABEM values the completeness of the history and physical examination or ensuring a patient's privacy. Consider a case testing the management of chest pain caused by acute cocaine use. Would it be responsible to ask a patient about cocaine use in the waiting room, taking into account HIPAA and privacy issues? What about asking whether a man has recently used sildenafil (Viagra™) before administering nitroglycerin? Patients might feel very uncomfortable revealing medical history such as HIV status, previous surgeries, and psychiatric illness in a nonprivate setting. Performing a complete physical examination would also be difficult in the waiting room. Examining a patient's chest wall for tenderness, bruising, or herpes zoster would be challenging while maintaining privacy. Genitourinary and rectal examinations would most likely have to be deferred. If physicians fail to ask about cocaine or Viagra™ use and delay rectal examinations during the initial waiting room evaluation, there is no guarantee that this information will be acquired later, yet subsequent providers may erroneously assume that these things had been done.Chest pain protocols are designed for ED treatment settings in which patients are in proper examination rooms with appropriate physician and nursing monitoring. Common chest pain treatments might cause more harm than good to waiting room patients. Would it be wise to administer sublingual nitroglycerin or morphine in a setting in which blood pressure is not easily monitored? What if the patient's blood pressure decreases and he or she falls out of their chair, sustaining injury? Drugs that are unequivocally considered beneficial in a normal ED setting may not have the same benefit/harm results when used in the waiting room. The austere conditions of the waiting room would likely make chest pain protocol deviations the norm, rather than the exception.Answer 3Q3. Imagine a condition for which patients generally do well but, according to national averages, in which 1 in 1,000 will die or have a bad complication. Now imagine 2 similar neighboring EDs with patient populations of identical demographic makeup. Each ED treats 2,000 patients with this condition each year. The hospitals use different strategies to manage this condition. ED A's strategy reduces the risk of a bad outcome to 1 in 2,000, half of the national average. ED B's approach produces 1 bad event per 500 cases.Q3.a On average, how many bad events can we expect at each hospital each year?At ED A, we would expect 1 bad outcome. At ED B, we would expect 4 bad outcomes.Q3.b What is the probability that ED B will have the same number of events or fewer events than ED A in any given year? Same question if quality levels were such that ED A expected 1 event per year and ED B 2? What about ED A having 2 events and ED B having 4?To answer this question, we need to understand a little bit about the mathematics of rare events. Such events are quantized: There can be 0, 1, 2, etc events per year; there can't be 1.23 events in a year. The Poisson distribution is typically used to model such phenomena. Simulated Poisson distributions for phenomena that, like the EDs in these examples, have 1, 2, and 4 expected events year are shown in the Figure. From these distributions, we can calculate the probability that there will be a specific number of events in the next year according to the expected number of events. The spreadsheet that does these calculations is available for download (Table E1, available at http://www.annemergmed.com). The general approach is to sum the probabilities of all situations that meet the required conditions. In the first example, pairs of number of events such as (ED A 0, ED B 0) or (1, 1) or (1, 0) or (2, 0) all meet the requirement that B≤A. We calculate the probability for each pair by multiplying the probability of ED A having the specified number of events (which can be gleaned from the top panel of the Figure) by the probability of ED B having the specified number of events (the bottom panel of the Figure) because we assume that the 2 probabilities are independent.Our calculations reveal that the probability that ED A has the same number of events or more events than ED B in a year, given that each ED treats 2,000 patients per year and true rates are 1 per 2,000 (ED A) and 4 per 2,000 (ED), is about 12.3%. If the true rates are ED A 1 per 2,000 and ED B 2 per 2,000, then it is 39.4%, and if they are 2 per 2,000 and 4 per 2,000, it is 25.5%.These examples illustrate that when events are rare, it is not uncommon to see EDs that are inherently worse performing better than EDs that are inherently better. As will be seen from the next question, when events are rare, long periods of observation are required to say with confidence that event rates at one hospital are lower than at another.Q3.c How many years of observation would be required (guess if you must) to prove with 95% confidence (assuming that the patient populations and hospitals were similar in all other respects and did not change during the length of the observation period) that ED A was indeed safer than ED B for this condition, assuming that rates are as stated in the original question?There are a number of ways to perform a sample size calculation to answer this question. Here is one approach.The rate at ED A is 1/2,000, or .0005, and at ED B it is 4/2,000, or .002. We use the sampsi (sample size) command in Stata 11 (StataCorp, College Station, TX) to determine the number of subjects needed to have an 80% chance of achieving a statistically significant result if these are the true rates (p[ower]=.8).. sampsi .0005 .002, p(.8)Estimated sample size for 2-sample comparison of proportionsTest Ho: p1=p2, where p1 is the proportion in population 1 and p2 is the proportion in population 2Assumptions:α=.0500 (2-sided)power=0.8000p1=0.0005p2=0.0020n2/n1=1.00Estimated required sample sizes:n1=9,998n2=9,998This analysis shows that it would take roughly 10,000 patients, or 5 years at each ED, to show that ED A is indeed safer than ED B, and that is assuming that nothing else about these EDs changes during that time. This example has profound implications for quality improvement efforts in emergency medicine. It suggests that a quantitative approach to quality–that we will prove with numbers that one institution is safer than another with respect to rare but important outcomes–is untenable.Answer 4Q4.a In his editorial, Dr. Wears cautions that in circumstances involving rare events, it is easy to be seduced into thinking a system is safe when it is not. Discuss his contention in light of your answers to question 3.Question 3's example provides strong support for Dr. Wears' contention that it is easy to be seduced into thinking something is safer than it is.11Wears R.L. Cook R.I. Getting better at being worse.Ann Emerg Med. 2010; 56: 465-467Abstract Full Text Full Text PDF PubMed Scopus (9) Google Scholar ED B has 4 times the bad outcomes of ED A, but it would take heroic efforts to know this. Most of us have little quantitative basis to know whether our practice is safe, too safe, or not safe enough. Scheuermeyer et al1Scheuermeyer F. Christenson J. Innes G. et al.Safety of assessment of patients with potential ischemic chest pain in an emergency department waiting room: a prospective comparative cohort study.Ann Emerg Med. 2010; 56: 455-462Abstract Full Text Full Text PDF PubMed Scopus (16) Google Scholar may believe that waiting room evaluation of low-risk chest pain patients is safe and they may well be right. What is important to understand is that we are unlikely to ever get data of sufficient quality and quantity to prove or disprove their thesis unless such trials are conducted on a national level with huge numbers of patients. This may be possible in countries with national health plans and centralized databases, but is unlikely to happen in the United States.Q4.b Dr. Schriger points out, however, that the converse could also be true. He opines that if risk-averse physicians develop an overly conservative practice style, even a prolonged period of event-free activity may be insufficient to change practice to a much less expensive but still adequately safe. Again, in light of your answers to question 3, discuss whether we can expect to gain scientific proof that one strategy is safer than another.The quantitative difficulties related to differentiating event rates when events are rare cuts both ways. Just as we cannot easily determine whether the waiting room strategy is safe enough, we also cannot tell when a strategy is too safe. For example, how would you know whether you or your department is ordering too many computed tomography angiograms to rule out pulmonary embolism? The understanding that we have little ability to quantitatively determine whether we are doing the right thing leads to some scary questions such as, If we have no mechanism to determine whether our practice style is properly calibrated, what determines our practice style? The answer to this question is a good topic for another journal club.Answer 5Q5. Dr. Kellermann laments that regardless of what chest pain rule-out system is correct, it is a shame that change occurs because of crowding and other political forces instead of a rational process based on science. He fears that emergency physicians, by cleverly developing workarounds to untenable situations, allow the fundamental problems to perpetuate.Q5.a Considering all of these perspectives, have a discussion about how a community might best determine what method for handling soft “rule-out acute coronary syndrome” patients is best.This is a bit of a trick question because, w