By: Marie Caulfield, PhD, manager of data operations and services, Association of American Medical Colleges, Washington, DC

Last year my daughter was admitted to a teaching hospital with a kidney infection. The medical student and residents on her treatment team recommended an MRI to rule out a possible anatomical cause. Dr. R., the attending physician, guided them through the case: the medical history did not suggest an anatomical problem, treatment would be the same regardless of the findings, and scans are uncomfortable and expensive. Dr. R. firmly said, “No MRI.” The medical student was clearly uncomfortable with leaving this stone unturned.

Two days later, when my daughter had responded to treatment but still had a significant fever, the medical student and residents recommended keeping her inpatient on IV antibiotics until her fever resolved. One of them brought up the MRI again. Dr. R. masterfully and respectfully talked them through the situation again: treatment was going well, the same drug is available in oral form that could be taken at home, an otherwise healthy 14-year-old is better off at home, and there would be a follow-up by her pediatrician. The team deferred to his judgment, but I could feel their discomfort. They clearly wanted to be sure.

This case was not rocket science. I could see that Dr. R. had seen it a thousand times, so he had the advantage of a perspective that the team did not. Did the team suggest anything off-the-wall or extraordinary? No. Could Dr. R. have let them do more tests? Yes. Would our insurance have paid for another night in the hospital? Absolutely. But Dr. R. was teaching something important. He modeled for them the need to balance a little uncertainty with other values, including health care costs and the comfort of the patient. He weighed the low probability that sending my daughter home with a fever would have a bad outcome, against the high probability that the treatment would keep working and she would be happier. He modeled being satisfied with successful treatment even when questions lingered.

There has been recent attention to the need for physicians, and by extension, medical students, to have a high tolerance for ambiguity. This quality is sometimes described as having less need for certainty. Research has found that physicians who are uncomfortable or stressed when they do not know “the right answer” may suffer from burnout. According to our study recently published in Academic Medicine, physicians with a high tolerance for ambiguity may be attracted to some medical specialties, such as emergency medicine and psychiatry. Physicians with low tolerance for ambiguity working in those fields may experience high stress.

Luther and Crandall observed in their 2011 commentary that, despite the call for increased tolerance for ambiguity among physicians, the culture of academic medicine “has little tolerance for ambiguity and uncertainty.” It is true that the traditional route to medical school is not about uncertainty. Individuals who get into medical school likely have done very well in premedical curricula that heavily weight fact-based science exams. Similarly, a large amount of energy in the first two years of medical school is devoted to preparing for the Step 1 exam, which assesses understanding of basic science principles and processes. It is a difficult multiple choice exam for which there are single correct answers. Successful medical students are expert learners of facts, and they have been well-rewarded for that skill. But the care of patients is full of ambiguity and uncertainty. It is messy and complicated, and real cases may not look like the ones in their text books. Judgments need to be made, sometimes in the absence of complete information. There are daily decisions about which tests are necessary and sufficient. Working with interdisciplinary teams is complicated, hospital systems are confusing and always changing. Should medical schools therefore screen for such qualities and seek to admit students who appear to have high tolerance for ambiguity? Or, assuming that most people fall in a moderate range on this trait, can medical schools foster tolerance for ambiguity and decrease discomfort with uncertainty? What would that look like? Perhaps it requires that mentors talk openly about their decision-making processes when the way forward is not comfortably clear. Maybe it begins with attending physicians like Dr. R., who take the time to explain when a little discomfort is okay.