An experimental comparison of rater performance on an SP-based clinical skills exam

Background: Research has provided only limited information regarding how best to structure rater training and the best rating formats for a standardized-patient (SP) based assessment. Purpose: The purpose of this study was to determine the efficacy of rater discussions as an expert rater training method and to compare the expert's rating of recorded video performances with real-time SP ratings on both a global and checklist evaluation instrument. Methods: We implemented an experimental study of a discussion-based expert rater training method to compare generalizability and interrater reliability of trained experts, untrained experts, and SP raters in a 4th-year clinical skills exam. Results: In terms of reliability, SP raters did as well as expert raters on the checklist and better than expert raters on global ratings. For the expert raters, training via a discussion method did not significantly improve the reliability of checklist ratings or global ratings. Within the experts groups, those with checklists performed better than those with global ratings. Conclusions: For SP-based exams designed to assess basic clinical skills, SP ratings appear to perform as well or better than expert ratings. Discussion-based expert training yielded no improvements in expert rater performance, and expert checklist scores generated more reliable results than the expert global rating scores.