ASSESSMENT OF CLINICAL-PERFORMANCE DURING SIMULATED CRISES USING BOTHTECHNICAL AND BEHAVIORAL RATINGS

Background: Techniques are needed to assess anesthesiologists' perform ance when responding to critical events. Patient simulators allow pres entation of similar crisis situations to different clinicians. This st udy evaluated ratings of performance, and the interrater variability o f the ratings, made by multiple independent observers viewing videotap es of simulated crises. Methods: Raters scored the videotapes of 14 di fferent teams that were managing two scenarios: malignant hyperthermia (MH) and cardiac arrest. Technical performance and crisis management behaviors were rated. Technical ratings could range from 0.0 to 1.0 ba sed on scenario-specific checklists of appropriate actions. Ratings of 12 crisis management behaviors mere made using a five-point ordinal s cale. Several statistical assessments of interrater variability were a pplied. Results: Technical ratings were high for most teams in both sc enarios (0.78 +/- 0.08 for Mil, 0.83 +/- 0.06 for cardiac arrest). Rat ings of crisis management behavior varied, with some teams rated as mi nimally acceptable or poor (28% for MH, 14% for cardiac arrest). The a greement between raters was fair to excellent, depending on the item r ated and the statistical test used. Conclusions: Both technical and be havioral performance can be assessed from videotapes of simulations. T he behavioral rating system can be improved; one particular difficulty was aggregating a single rating for a behavior that fluctuated over t ime. These performance assessment tools might be useful for educationa l research or for tracking a resident's progress. The rating system ne eds more refinement before it can be used to assess clinical competenc e for residency graduation or board certification.