Decisions, Decisions: Emergency Physician Evaluation of Low Probability–High Morbidity Conditions
Article Outline
In contrast to many other specialists, emergency physicians spend a substantial amount of their time deciding whether to test for dangerous but unlikely diagnoses. Every time an emergency physician treats a patient who appears well but whose medical history prompts a differential diagnosis that includes acute coronary syndrome, pulmonary embolism, or aortic dissection, a low-probability–high-morbidity decision arises. Although the majority of such patients will not have these diagnoses, the emergency physician must consider consequences for the few who do. So fundamental are these low-probability–high-morbidity decisions to the practice of emergency medicine that one would expect our journals to overflow with investigations about how such decisions are best made. Surprisingly, this has not been the case.
In this issue, Katz et al1 break this pattern and delve into this matter. They report that physicians with similar training, treating similar patients with chest pain, make different decisions. The authors divided 33 emergency physicians into tertiles of low, medium, and high malpractice fear and examined admission and evaluation decisions in 1,134 consecutive patients with symptoms of acute coronary syndrome. They found that physicians in the upper tertile were more likely to admit low-risk patients and were more likely to order cardiac markers and chest radiographs. These findings persisted in analyses that adjusted for physicians' general risk aversion and other patient-specific and department-specific factors. Despite these differences in practice, the rates of missed acute coronary syndrome were similar among physicians in the 3 tertiles. Thus, Katz et al1 found evidence of variation in practice that was not strongly associated with outcome and was partially explained by the physician's fear of malpractice.
Since 1974, when Wennberg and Gittelsohn2 first documented huge variations in practice style, a variety of movements within medicine have attempted to decrease variation. Standardization of residency curricula, evidence-based medicine, clinical guidelines, and case review by various regulatory and payer organizations are but a few examples of forces that would be expected to homogenize practice. Yet, as the Katz et al1 study demonstrates, considerable variation persists. What are the forces that maintain variation despite the many efforts to standardize care? How can we begin to understand why similarly trained emergency physicians, practicing in similar circumstances, make different decisions?
Decision theory, a model for making such decisions, suggests that the rational consideration of low-probability–high-morbidity situations requires an estimation of the likelihood of the diagnosis, the difference in outcome if the condition is evaluated or ignored, the cost of evaluating the condition, and the cost (both economic and medical) of false positive test results.3, 4 If that were not enough, knowledge of societal and patient preferences is needed to reach a final decision. Given the complexity of this model, it is not surprising that emergency physicians seldom invoke this formalism in their routine decisionmaking. We know little about the process that stands in its place, and by taking a first step in trying to understand emergency physician decisionmaking in low-probability–high-risk situations Katz et al1 have done the specialty and the research community an important service.
Knowingly or unknowingly, the emergency physician making such a decision is attempting to maximize utility (the net benefit) for several parties—the patient, the physician, and society—and the utilities are not necessarily aligned. The chest-pain patient at low risk for acute coronary syndrome highlights the potential for misalignment. Imagine that an emergency physician has determined that his or her patient has a 2% chance of having acute coronary syndrome and a 0.5% chance of having an acute myocardial infarction (and ignore for a moment that a different emergency physician could evaluate the same patient and produce very different probability estimates). What disposition decision should be made? For the patient, this decision is not straightforward, because individuals will value the benefits, harms, and costs of hospitalization differently. Some patients may say that the 2% chance of acute coronary syndrome is high enough to warrant hospitalization. Others, concerned about the dangers of hospitalization and financial and opportunity costs, would decide that an outpatient evaluation best meets their needs.
The decision that maximizes the patient's utility may not maximize the physician's. For the physician, a decision to admit does not increase risk. We have yet to hear of a lawsuit filed because the physician admitted a chest-pain patient to a monitored bed when he should have been discharged. For the emergency physician, the potential downsides of an overly generous admission policy are the ire of the admitting staff and the feeling of unease that may arise when the physician realizes that he or she is maximizing personal benefit rather than the patient's or society's.
The decision that maximizes society's utility may differ from that which maximizes the patient's or the physician's. Although the liberal use of admission might maximize the utility for some patients and many physicians, this strategy may drain funds from other important medical or nonmedical uses—education, transportation, the environment—that would produce greater overall good.
It is important to recognize how a seemingly medical decision—should this patient be admitted—deconstructs into a complex set of philosophical issues (the good of the individual versus the good of the many, the right to self determination, moral hazard, etc) under the most superficial scrutiny. Because our society has no mechanism to overtly decide whose utility is to be maximized and because ethics courses in medical school are too vague to be useful in guiding such decisions, it is not surprising that there is considerable variation in the management of low-probability–high-morbidity conditions. Add in the difficulty in determining each patient's actual risk and we have a recipe for medical care that is closer to chaos than homogeneity. Although some will argue that the solution to these problems is technological—better cardiac markers, chest pain centers, etc—these modalities do not eliminate the need to explicitly decide whose risk we are managing and what levels of risk are acceptable. Our specialty has a history of taking on the difficult questions in medicine and providing innovative solutions. It is time to follow the lead of Katz et al1 and initiate the research and dialogue that will lead to a more coherent and consistent treatment of patients with low-probability–high-morbidity conditions.
References
- Emergency physicians' fear of malpractice in evaluating patients with possible acute cardiac ischemia. Ann Emerg Med. 2005;46:525–533
- . Small area variation in health care delivery. Science. 1973;182:1102–1108
- . Theory of Games and Economic Behavior. Princeton: Princeton University Press; 1944;
- . The framing of decisions and psychology of choice. Science. 1981;211:453–458
Supervising editor: J. Stephan Stapczynski, MDFunding and support: The authors report this study did not receive any outside funding or support.Reprints not available from the authors.
PII: S0196-0644(05)01283-7
doi:10.1016/j.annemergmed.2005.06.445
© 2005 American College of Emergency Physicians. Published by Elsevier Inc. All rights reserved.
