Predictive Validity of Family Medicine Resident's Performance at Induction with Their Future Performance During Residency.

OBJECTIVE: To determine the predictive validity of induction assessment scores of family medicine residency program for performance in final year of residency.

STUDY DESIGN: Psychometric (predictive-validity) study.

PLACE AND DURATION OF STUDY: The Aga Khan University Hospital, Karachi, from 2008 to 2014.

METHODOLOGY: All family medicine residents were evaluated. Family Medicine Residency induction assessment scores were used as independent predictor variables, whereas resident assessment scores during the final year residency program were used as dependent or outcome variables. Data was analyzed using SPSS version 19. Linear regression was used to determine predictive validity of induction scores as independent variables with outcome variables at 95% confidence level.

RESULTS: There were 33 residents. MBBS scores accounted for 30.1% of variance in final year Objective Structured Clinical Exam (OSCE) scores; whereas, induction written test accounted for 37.1% of variance in final year written test. Induction communication skill scores did not correlate with in-training communication skill scores or with the final year OSCE scores. Induction professionalism scores accounted for 13.7% of the variation in final year OSCE scores, but not with in-training continuous professionalism scores.

CONCLUSION: Induction knowledge scores have acceptable predictive value for future knowledge and its application. Other valid and reliable assessment methods, such as multiple mini-interviews, should be explored for assessment of noncognitive domains at induction.