Predictive Power of Resident Selection Criteria for Clinical Performance in General Surgery Residency

Abstract

To evaluate the ability of the resident selection criteria to predict clinical performance of general surgery residents, the application files and resident evaluations of 35 general surgery residents who were in the residency program of the Department of Surgery, Faculty of Medicine Siriraj Hospital during the 2000 - 2001 and 2001 - 2002 academic years were reviewed. A correlation study was done using scores from three selection criteria (medical school grades, letters of recommendation, and interview) predictors and clinical performance ratings as outcomes. The interview scores were the best predictor for overall performance of residents in the first and second years. The GPA scores were the best predictor for overall third-year performance. Each selection criterion contributed unique predictive ability for resident performance. The combination of interview scores, scores from letters of recommendation, GPA scores, and ages at admission could predict 60.5% of the total variance in the overall first-year performance scores (R=0.778, p=0.012). The combination of interview scores and score from letters of recommendation could predict 31.4% of the total variance in the overall performance score in the second year (R=0.56, p=0.049). None of the multiple regression models demonstrated statistically significant prediction for the third-year overall performance.