Association of Behavioral Interviewing With Resident Performance

Abstract

PURPOSE: We sought to evaluate the effectiveness of our institution's incorporation of behavioral interviewing in the resident interview process by determining association of behavioral and traditional interview scores with resident performance evaluations. BACKGROUND: Behavioral interviewing uses past behavior in an attempt to predict future performance, and validation has not yet occurred in healthcare professions. METHODS: Behavioral and traditional interview scores of 44 matched residents from 2006 to 2011 were analyzed with performance evaluation data to determine associations. One interviewer strictly asked behavioral-based questions. The traditional interview score was an average score from four different faculty interviews. A resident panel interview score was also submitted. We measured resident performance by evaluations completed on an ongoing basis throughout the resident's training. Using mixed model analysis, significance of association and slopes of the linear relationship were estimated for each of the three interview scores to resident performance evaluations. RESULTS: Behavioral interviewing had no association with resident performance evaluation scores (faculty evaluations slope100±SE=−0.015±0.1081, P=.8864, nursing evaluations slope100±SE=−0.046±0.1197, P=.6990). Traditional faculty interviews (slope100±SE=0.2443±0.1095, P=.0283) and the resident panel interview (slope100±SE=0.3606±0.0684, P<.0001) both had association to resident performance based on faculty evaluations. The resident panel score (slope100±SE=0.3588±0.1176, P<.0001) had significant association with resident performance in a multivariate model while the traditional faculty interview score (slope100±SE=0.0874±0.1464, P=.3424) did not. DISCUSSION: In our program, behavioral interview scores did not associate with our measure of resident performance, while traditional interviewing methods had significant association.