Cognitive differences in clinical reasoning related to postgraduate training

Abstract In this study, first‐year residents in family medicine (FM), first‐ and second‐year residents in internal medicine, and experienced academic nephrologists were confronted with a series of eight complex nephrology problems, represented by a brief written history and a standard array of laboratory data. Subjects were asked to solve the problem while "thinking aloud."; Results showed a large gradient in diagnostic ability over the three cohorts, ranging from a mean accuracy of 25% for FM residents to 91% for nephrologists. Increased experience was associated with increased clustering of individual data into meaningful relations and more extensive use of causal explanations but no difference in use of forward or backward reasoning. These results are at variance with an extensive literature showing (a) no or little difference in outcome measures of clinical competence related to postgraduate education but (b) greater use of forward reasoning and less use of causal biomedical explanations, associated with increased expertise. An explanation in terms of Schmidt and Norman's stage theory is proposed.