As an internist, working in the emergency room feels at times like the dark underbelly of medicine. The frenetic pace, the need to make decisions within highly uncertain conditions, and reliance on technology all cut against the grain of the internists credo of “being a doctor’s doctor.” If internists are biased in how they arrive at diagnoses, emergency medicine doctors face such bias on an exponential scale. Clinical decision-making is a tricky process, and so far no checklist or educational strategy has made it less tricky. Below are three strategies to further understand and demystify clinical bias.
1. Create a diagnostic batting average for physicians. Despite remarkable gains in information technology, we still have little data on comparative effectiveness of physicians. On a macro-level, quality metrics such as percentage of patients with diabetes or controlled hypertension capture the quality of healthcare systems, rather than individual clinicians. How then might we capture the difference in diagnostic accuracy amongst clinicians? I believe the emergency room is a good place to start. In a recent twelve-hour shift in the ED, I ordered three abdominal CT scans. Despite being assured in my diagnostic reasoning of gallbladder disease, pancreatitis, and diverticulitis, I wrong in 2/3 cases. Collecting this information is simple, and comparing it incredibly useful for understanding differences in diagnostic reasoning.
2. Foster and environment that combats testing momentum. Diagnostic momentum is tendency to continue with a diagnosis despite further information. A less-discussed bias is what I call “testing momentum,” the tendency to “test oneself to certainty.” I believe that this type of bias is more subtle and institutionalized. It can be heard in such statements as “he just bought himself a head CT” or “we need an abdominal CT to put his complaints to rest.” Part of this tendency is both practical and prudent; indeed, a negative abdominal CT scan will make everyone feel better sending the patient home (increased radiation exposure aside). But it also reflects a tunnel vision of care, in that had the patient presented for care in a different context, they would not have undergone as aggressive testing. As the problem is institutional, the solution may also be to continue to ask the question, “If the patient presented to the primary care clinic, what would you order?”
3. Push residents and students to defend testing choices. The recent push of the Choosing Wisely campaign will result in an extraordinary savings of inappropriate tests. The fact is, however, that the vast majority of tests ordered live in a gray zone of appropriateness. Does a person with a clear COPD exacerbation need a troponin test to measure for heart attack? Does a person with severe hypertension and vision changes need a CT scan? These gray zones represent the silent majority of clinical decision-making. In approaching them, we may do well to think of the concept of framing. Instead of asking, “What tests will you order?” we can ask, “Do we need to do any testing at all?” and finally, “What is the evidence behind your choice?”
The problem of cognitive bias in decision-making is both technical and cultural. To further understand it, we need to commit to collecting data differently, to actively question our decisions, and systematically explore the gray zones of medicine.
Tom Peteet is an internal medicine resident.