Objective: To determine the reliability and validity of "standardized patients" to assess clinical performance of internal medicine residents.
Design: Each resident spent 2 half-days interacting with 19 standardized patients (nonphysicians taught to portray patients in a reproducible fashion). Each resident was asked to obtain focused histories, perform relevant physical examinations, and provide patient education or counseling. At the end of each 10- to 15 minute encounter, the resident was asked to identify positive findings on physical examination and to prioritize a differential diagnosis. Other, more traditional indicators of clinical skills were also obtained.
Participants: Three hundred ten residents from nineteen internal medicine training programs in the New England region of the United States.
Results: Reproducible estimates of data gathering and interviewing skills were achieved with 1 day of testing. Validity analyses were difficult to interpret because of the lack of an external "gold standard." However, faculty judgments of performance from reviewing videotapes corresponded with standardized-patient-based scores. Differences in group performance were also demonstrated across years of training and between U.S. or Canadian and foreign medical graduates.
Conclusions: Systematic and effective procedures for developing standardized-patient-based test materials and for training standardized patients have been developed. This technique is best used for measuring data gathering and interviewing skills. Correlations with commonly used evaluation methods were generally low; this may be because standardized patients measure different skills. Local use of this technique by residency programs and the development of regional consortia sharing resources, costs, and expertise are advocated.