Short-term memory constraints upon the generation of diagnosis hypotheses by novice clinicians

The structure and function of memory affect medical problem solving. The purpose of this 1982 study was to test the assertion that the number of diagnoses actively entertained by novice clinicians during a simulated case would not be greater than seven, the average maximum storage capacity of short-term memory. A total of 24 junior physician's assistant students took mock medical histories and gave mock physical examinations using simulated patients. Each of the eight cases investigated by each group included a primary problem with various numbers of secondary problems. Three successive time periods, called interludes, were provided in which each student had three minutes to list all the hypotheses he or she was actively considering concerning the diagnosis of the patient's problem. Each interlude was scheduled at a progressively more complete point in the student's knowledge of the case. The results supported the assertion that was tested.