Quality improvement strategies for organizational change: a multiphase observational study to increase insight into nonparticipating organizations.

Background: The scope of implementation research is often restricted to the analysis of organizations that participate voluntarily in implementation interventions. The recruitment of participants for a quality improvement collaborative increases awareness of the specific innovation. The objective of this multiphase observational study was to identify differences between organizations that participated in a large-scale implementation project aiming to improve perioperative care, functional recovery, and length of hospital stay after gynecologic surgery and organizations that did not participate. A secondary objective was to explore how perioperative practice changed among nonparticipants.

Methods: Of the seven gynecology departments of nonparticipating Dutch hospitals, five agreed to participate in a retrospective analysis. Baseline data of participating hospitals' (N = 19) characteristics, time to functional recovery, and length of hospital stay were compared. Outcome measures for the subsequent pre-post awareness study in the five nonparticipating hospitals were: (1) overall adherence to predefined evidence-based perioperative elements; and (2) change in functional recovery and length of hospital stay. Multivariable regression models, adjusted for baseline characteristics, were used for analysis.

Results: In retrospect, nonparticipating and participating hospitals did not differ in baseline characteristics, functional recovery, and length of hospital stay. In three of the five nonparticipating hospitals, adherence to the selected evidence-based perioperative elements increased significantly after awareness of the trial (overall mean difference 9.7%, 95% CI 6.9 to 12.5%, p <  0.001). Linear regression models revealed no statistically significant or clinically relevant differences in time to functional recovery (mean difference - 0.2 days, 95% CI -0.7 to 0.2, p = 0.319) or length of hospital stay (mean difference - 0.4 days, 95% CI -1.3 to 0.5, p = 0.419) in the nonparticipating hospitals. None of these hospitals managed to reduce time to functional recovery or length of hospital stay significantly.

Conclusions: No differences in perioperative outcomes between the nonparticipating and participating hospitals were identified at baseline. Despite the statistically significant improvement in overall evidence-based perioperative care, the awareness raised by recruitment activities alone was not enough to reduce time to functional recovery and length of hospital stay in nonparticipating hospitals. Insight into the trends of nonparticipants is valuable to existing implementation effectiveness research.