Missing celiac disease in family medicine: the importance of hypothesis generation

Delays in diagnosing celiac disease average 13 years. We aimed to identify reasons for misdiagnosis in family medicine.

During a larger study on diagnosis, a scenario describing a 30-year-old female with 3-month abdominal pain, diarrhea, and microcytic anemia consistent with celiac disease was presented on a computer to 84 family physicians. Their information gathering and diagnoses were recorded. Fifty physicians misdiagnosed, and 38 of these took part in "stimulated recall'': they were asked to recall their hypotheses and inferences step by step, aided by a record of their information gathering. They were unaware of the misdiagnosis.

Transcripts were analyzed to identify whether celiac disease was mentioned and how information was interpreted. Two raters independently assessed information interpretation against the published evidence (kappa = 0.85).

Physicians did not change their diagnoses during stimulated recall. Only 10 physicians mentioned celiac disease as a hypothesis (26%). "Diarrhea'' and "pain relief by defecation,'' consistent with both celiac disease and irritable bowel syndrome (IBS), were only linked to IBS. "Absence of weight loss'' led to rejecting celiac disease, although weight loss is characteristic of advanced disease. A complete blood count was requested as a routine test and not specifically for celiac disease. Thus, the unexpected result of "microcytic anemia,'' inconsistent with IBS, did not trigger the correct diagnosis.

Most physicians never considered celiac disease. Information inconsistent with the favorite IBS diagnosis was overlooked. Reviewing the case did not prompt physicians to consider celiac disease, re-evaluate the evidence, or rethink the IBS diagnosis.