Evaluating problem solving based on the use of history findings in a standardized-patient examination

To evaluate a novel item format for assessing clinical problem solving in a standardized-patient examination (SPE).

In 1992-93 a key-findings item format was included in two versions of three stations in an SPE (given in the style of an objective structured clinical examination) that was taken by 198 third-year students at the end of their three-month internal medicine clerkship at the University of Texas Medical Branch at Galveston. Each of the stations involved an extended matching question that listed ten to 12 findings. The students were told to select as many or as few findings as they wished that were key in leading them to their diagnosis of the standardized patient's (SP's) problem. The findings fell into three categories: (1) key to the diagnosis, (2) provided by the SP but not key to the diagnosis, and (3) not provided by the SP.

A total of 169 students (85%) identified at least one of the findings determined to be key in each of the stations. Correctly identifying key findings was related to correctly diagnosing the SP's problem. A total of 145 students (73%) indicated at least one finding across the three cases as key to their diagnosis but not given by the SP. Selecting nonexistent findings as key to diagnosis was not found to be related to performance on the SPE. It was hypothesized that once a diagnosis was made, the students had difficulty differentiating key findings that had led them to the diagnosis from other key features of their cognitive model of illness.

The students were generally able to obtain and recognize at least some of the key information they needed to formulate appropriate differential diagnoses, and the ability to identify key findings was shown to be related to identifying the most appropriate diagnosis. The key-findings item format has potential both in assessment and for gaining a better understanding of the clinical problem-solving process.