Objective: Resuscitation of critically ill patients requires medical knowledge, clinical skills, and nonmedical skills, or crisis resource management (CRM) skills. There is currently no gold standard for evaluation of CRM performance. The primary objective was to examine the use of high-fidelity simulation as a medium to evaluate CRM performance. Since no gold standard for measuring performance exists, the secondary objective was the validation of a measuring instrument for CRM performance—the Ottawa Crisis Resource Management Global Rating Scale (or Ottawa GRS).
Design: First- and third-year residents participated in two simulator scenarios, recreating emergencies seen in acute care settings. Three raters then evaluated resident performance using edited video recordings of simulator performance.
Setting: A Canadian university tertiary hospital.
Interventions: The Ottawa GRS was used, which provides a 7-point Likert scale for performance in five categories of CRM and an overall performance score.
Measurements and Main Results: Construct validity was measured on the basis of content validity, response process, internal structure, and response to other variables. One variable measured in this study was the level of training. A t-test analysis of Ottawa GRS scores was conducted to examine response to the variable of level of training. Intraclass correlation coefficient scores were used to measure interrater reliability for both scenarios. Thirty-two first-year and 28 third-year residents participated in the study. Third-year residents produced higher mean scores for overall CRM performance than first-year residents (p < .0001) and in all individual categories within the Ottawa GRS (p = .0019 to p < .0001). This difference was noted for both scenarios and for each individual rater (p = .0061 to p < .0001). No statistically significant difference in resident scores was observed between scenarios. Intraclass correlation coefficient scores of .59 and .61 were obtained for scenarios 1 and 2, respectively.
Conclusions: Data obtained using the Ottawa GRS in measuring CRM performance during high-fidelity simulation scenarios support evidence of construct validity. Data also indicate the presence of acceptable interrater reliability when using the Ottawa GRS.