Surgical safety checklist audits may be misleading! Improving the implementation and adherence of the surgical safety checklist: a quality improvement project.

Despite good quality evidence for benefits with its use, challenges have been encountered in the correct and consistent implementation of the surgical safety checklist (SSC). Previous studies of the SSC have reported a discrepancy between what is documented and what is observed in real time. A baseline observational audit at our institution demonstrated compliance of only 3.5% despite a documented compliance of 100%. This project used quality improvement principles of identifying the problem and designing strategies to improve staff compliance with the SSC. These included changing the SSC from paper-based to a reusable laminated form, a broad multidisciplinary education and marketing campaign, targeted coaching and modifying the implementation in response to ongoing staff feedback. Five direct observational audits were undertaken over four Plan-Do-Study-Act cycles to capture real-time information on staff compliance. Two staff surveys were also undertaken. Compliance with the SSC improved from 3.5% to 63% during this study. Staff reported they felt the new process improved patient safety and that the new SSC was easily incorporated into their workflow. Improving compliance with the SSC requires deep engagement with and cooperation of surgical, anaesthesia and nursing teams and understanding of their work practices and culture. The prospective observational audit highlighted an initial 3.5% compliance rate compared with 100% based on an audit of the patient notes. Relying solely on a retrospective paper-based model can lead to hospitals being unaware of significant safety and quality issues. While in-person prospective observations are more time-consuming and resource-consuming than retrospective audits, this study highlights their potential utility to gain a clear picture of actual events. The significant variation between documented and observed data may have considerable implications for other retrospective studies which rely on human-entered data for their results.