Novice diagnostic reasoning in medicine: Accounting for evidence

This study investigated the strategies used by beginning, intermediate, and advanced medical students when solving clinical cases containing inconsistent data. Written clinical cases were used in which the initial patient data suggested a very common and prototypical disease. Subsequent data in the cases were inconsistent with this first suggestion. The cases were presented on a computer so that the subjects had the opportunity to reason aloud about the evidence. Verbal protocols were collected and analyzed. Our results showed that the beginning students generated the suggested hypothesis and either ignored or reinterpreted the inconsistent evidence to fit their initial hypotheses. Advanced novices generated several hypotheses with a similar underlying problem; this allowed them to narrow their initial hypothesis set in the face of inconsistent evidence and make fewer data reinterpretations. Intermediate novices showed a similar pattern but failed to change their hypotheses with inconsistent data. They maintained several hypotheses of a diverse nature concurrently, without evaluating them efficiently. They showed a deterioration in performance-commonly known as the intermediate effect-which we interpreted as the result of a process of knowledge reorganization that interferes with problem solving. Results are discussed in terms of change in reasoning strategies as a function of the type of learning in medical school.