BACKGROUND: Safety climate is often measured via surveys to identify appropriate patient safety interventions. The introduction of an insurance premium incentive for simulation-based anesthesia crisis resource management (CRM) training motivated our naturalistic experiment to compare the safety climates of several departments and to assess the impact of the training.

METHODS: We administered a 59-item survey to anesthesia providers in six academic anesthesia programs (Phase 1). Faculty in four of the programs subsequently participated in a CRM program using simulation. The survey was readministered 3 yr later (Phase 2). Factor analysis was used to create scales regarding common safety themes. Positive safety climate (% of respondents with positive safety attitudes) was computed for the scales to indicate the safety climate levels.

RESULTS: The usable response rate was 44% (309/708) and 38% (293/772) in Phases 1 and 2 respectively. There was wide variation in response rates among hospitals and providers. Eight scales were identified. There were significantly different climate scores among hospitals but no difference between the trained and untrained cohorts. The positive safety climate scores varied from 6% to 94% on specific survey questions. Faculty and residents had significantly different perceptions of the degree to which residents are debriefed about their difficult clinical situations.

CONCLUSIONS: Safety climate indicators can vary substantially among anesthesia practice groups. Scale scores and responses to specific questions can suggest practices for improvement. Overall safety climate is probably not a good criterion for assessing the impact of simulation-based CRM training. Training alone was insufficient to alter engrained behaviors in the absence of further reinforcing actions.