Background: Many hospitals, including those affiliated with New York University Medical Center, have adopted the use of rapid response teams (RRTs) to improve patient care. There is currently no standardized training for residents leading RRTs comparable to Advanced Cardiac Life Support (ACLS) training. We developed and validated an assessment tool of leadership skills for use in a new RRT leadership training program for second year internal medicine residents. Methods: Using a high-fidelity patient simulator, we created a patient scenario to assess resident leadership skills. The scenario involved an elderly post-surgical patient with multiple comorbidities who develops hypotension. Residents were expected to gather information, generate a differential diagnosis, plan and execute interventions, and manage the care team. After performing a literature search to identify twenty-nine components of excellent leadership in acute medical situations, we created a behavior-anchored scoring system using those components to assess resident performance during the scenario. These behaviors can be grouped under the domains of communication, decision making, implementation, resource management, situational awareness, teamwork, and workload management. Each was scored as "not done," "partly done," or "well done." Finally, each resident's leadership performance was assessed overall as being "poor," "average," or "good." Results: We assessed the internal consistency and reliability and found the Cronbach's alpha to be 0.907 for the overall score. Furthermore, the overall percentage of items well done varied with the overall global assessment. Residents with a "poor" overall performance only received a score of "well done" on 16.7% of behaviors (SD 13.6); those with an "average" overall performance received a score of well done on 29.5% of behaviors (SD 12.1); and those with a "good" overall performance received a score of "well done" on 55.9% of behaviors (SD 6.4), p < 0.001). The residents scores of percent "well done" in the six domains are described in table one. The overall behavior percent "well done" was 35% (SD 18.7). Conclusions: The tool we developed to evaluate resident leadership skills was internally consistent and the performance of the individual components predicted global performance. Overall, residents did not perform well either in percent of behaviors well done or in the six global domains. We plan to use this tool to help measure the effect of a training program to improve RRT management. Table One -Behavioral Domains (N=22) Behavioral Domain Mean Percentage Well Done Standard Deviation Communication 43.9% 24.4% Decision-Making 34.1% 28.4% Implementation 68.2% 36.3% Resource Management 51.5% 39.5% Situational Awareness 40.9% 36.6% Teamwork 26.4% 28% Workload Management 9.1% 15.2%

No abstract