Factors associated with successful implementation of Clean Cut: a perioperative surgical site infection prevention quality improvement programme-a cohort study of low-resource hospitals.

Abstract

Introduction: "Clean Cut" is an adaptive, multimodal surgical quality improvement (QI) programme that has been associated with significant reductions in surgical site infections. Following implementation in multiple hospitals and countries, we noted variability in impact. We aimed to understand the attributes of hospitals that contribute to the success of a perioperative QI programme in resource-limited settings. We hypothesised that factors related to hospital context before implementation influenced programme success.

Methods: Hospital context assessments were undertaken in 18 hospitals in low-income countries prior to the implementation of Clean Cut, which focuses on improving perioperative infection prevention and control (IPC) standards. We assessed staffing, training, infrastructure and prior QI experience. Hospitals also self-assessed compliance with six standards embedded in the IPC programme and compared reported compliance to the baseline compliance observed by trained data collectors. We defined high-improvement hospitals as those who improved three or more of the six standards by either doubling compliance while also achieving a minimum final compliance>50% or reaching a final compliance>90%. We compared context assessments of high- and low-improvement hospitals.

Results: Infrastructure, trainings and QI experience were not associated with larger improvements. However, high-improvement hospitals had fewer operating room staff (p=0.046) and overestimated their baseline compliance with IPC standards (p=0.032).

Conclusion: Clean Cut implementation was more successful with smaller staff numbers, reflecting challenges with engaging large numbers of stakeholders. High-improvement hospitals overestimated baseline IPC practices, suggesting that this programme is most beneficial when it identifies gaps that hospitals were previously unaware of. Reassuringly, improvements were not dependent on specific resources, indicating that the approach can be implemented in many environments.