A few years ago, I was talking with a medical student who had just finished a shift in the ED. He was an excellent student, but today he had a string of patients who left him frustrated. The 40 year-old woman who presented with abdominal pain left the ED with a diagnosis of “abdominal pain.” We discharged the 35 year-old man who presented with a severe headache with a diagnosis of “possible tension headache.” A five year-old girl with a fever left with “fever of probable viral etiology.” We had gotten blood tests and a CT scan on the woman with abdominal pain, the results of which were normal. We had treated the man’s headache with analgesics, and he got better. We had given the five year-old some Jell-O when she stopped vomiting, and she was playing happily with some toys. The student said, “These people came in bent over in pain or crying. We put them in a bed. We talked to them and examined them. We ordered some tests. A few hours later, they felt better, and we sent them home. They walked out standing up straight and smiling. But we didn’t really do anything. We never figured out what was wrong with them.”
I had to smile because, in many ways, the student was right. He had been learning about the dangerous causes of headache–subarachnoid hemorrhage, meningitis, brain tumors. But he had not learned much about tension headache, or migraine–the causes of most headaches that often improve on their own. He had learned about pancreatitis, appendicitis, ulcers, and colon cancer as causes for abdominal pain, but not abdominal pain of unknown etiology, a common problem that typically gets better without treatment or knowing what caused it. He had learned about fever in a five year-old caused by strep throat or meningitis but not about the various self-limited viruses that caused a fever and vomiting in that same five year-old.
When he had interviewed each patient, his differential diagnosis had included a list of the most serious causes of the problem, as it should. But he had no idea how frequently they occurred or what clues might tip him off as to how to recognize the more common and less serious causes of the same symptoms. His frustration was a function of the emphasis in his education of various specific causes of pathology, which populated his problem based learning cases but did not align with the real world of problems that often are not connected to any pathological condition or given a diagnostic name beyond the symptom the patient described–headache, abdominal pain, fever.
When I read the study of ambiguity tolerance in medical students by Caulfield et al I flashed back to this medical student. Was he frustrated by the lack of certainty in our diagnoses and the ambiguity surrounding our treatment decisions? He chose radiology as his residency, a specialty in the middle range for students’ scores on the Tolerance for Ambiguity (TFA) scale. If I had known about this test, I might have discussed it with him. However, as Caulfield et al noted, many questions still exist about what the results on the TFA test mean and how changeable they are over time. The two blog posts by Luther and Caulfield provide different perspectives on the TFA test.
Our tolerance for ambiguity in medicine also changes with experience. In the cases I described earlier, I expected that the woman with the abdominal pain would have a negative CT because her presentation suggested that her problem was not serious and would likely improve on its own. I ordered the CT scan to lower the probability of a serious condition below my threshold of concern. I could tolerate the remaining uncertainty about the diagnosis. But the student had no such threshold in his mind. The entire spectrum of possibilities swirled around him, and he could not understand why we stopped our work up with the negative CT scan. Even more perplexing for him were the other cases for which we did virtually no tests, and I decided to send the patients home. While I was relatively confident because I felt the risk was low, he was anxious because he had no base of experience to estimate that risk.
Whether the TFA test will be helpful for advising medical students about career choice or even evaluating their suitability for a career in medicine is not clear. I applaud Caulfield et al for their efforts to bring the TFA test to our attention as medical educators, but I urge caution in applying it. I look forward to further work in this area. In the meantime, I suspect the list of specialties ranked from lowest to highest tolerance for ambiguity will likely be the topic of discussion at meetings of medical school department chairs. Perhaps they will decide that when the chairs of dermatology and otolaryngology (low tolerance for ambiguity) and the chairs of emergency medicine and neurosurgery (high tolerance for ambiguity) can agree on a decision, it is likely that everyone else will agree, and they can adjourn the meeting.