RESUMEN
BACKGROUND: To test the value of a simulated Family Conference Objective Structured Clinical Exam (OSCE) for resident assessment purposes, we examined the generalizability and construct validity of its scores in a multi-institutional study. METHODS: Thirty-four first-year (PG1) and 27 third-year (PG3) surgery residents (n = 61) from 6 training programs were tested. The OSCE consisted of 2 cases (End-of-Life [EOL] and Disclosure of Complications [DOC]). At each program, 2 clinicians and 2 standardized family members rated residents using case-specific tools. Performance was measured as the percentage of possible score obtained. We examined the generalizability of scores for each case separately. To assess construct validity, we compared PG1 with PG3 performance using repeated measures multivariate analysis of variance (MANOVA). RESULTS: The relative G-coefficient for EOL was .890. For DOC, the relative G-coefficient was .716. There were no significant performance differences between PG1 and PG3 residents. CONCLUSIONS: This OSCE provides reliable assessments suitable for formative evaluation of residents' interpersonal communication skills and professionalism.