Implementation of a multicomponent intervention to prevent physical restraints in nursing homes (IMPRINT): A pragmatic cluster randomized controlled trial

Background: Despite clear evidence for the lack of effectiveness and safety, physical restraints are frequently applied in nursing homes. Multicomponent interventions addressing nurses' attitudes and organizational culture have been effective in reducing physical restraints. Objective: To evaluate the effectiveness of two versions of a guideline and theory-based multicomponent intervention to reduce physical restraints in nursing homes. Design: Pragmatic cluster randomized controlled trial. Setting: The study was conducted in 120 nursing homes in four regions in Germany. Participants: All residents living in the participating nursing home during follow-up, newly admitted residents were also included. A total of 12,245 residents included in the primary analysis (4126 and 3547 residents in intervention group 1 and 2 and 4572 residents in the control group). Methods: Intervention group 1 received an updated version of a successfully tested guideline-based multicomponent intervention (comprising brief education for the nursing staff, intensive training of nominated key nurses in each cluster, introduction of a least-restraint policy and supportive material), intervention group 2 received a concise version of the original program and the control group received optimized usual care (i.e. supportive materials only). Primary outcome was physical restraint prevalence at twelve months, assessed through direct observation by blinded investigators. Intervention and control groups were compared using baseline-adjusted linear regression on cluster level, Bonferroni-adjusted for double testing. Secondary outcomes included falls, fall-related fractures, and quality of life. We also described intervention costs and performed a comprehensive process analysis. Results: At baseline, mean physical restraint prevalence was 17.4% and 19.6% in intervention groups 1 and 2, and 18.8% in the control group. After twelve months, mean prevalence was 14.6%, 15.7%, and 17.6%. Baseline-adjusted differences between mean prevalences were 2.0% (97.5% CI, -5.8 to 1.9) lower in intervention group 1 and 2.5% (97.5% CI, -6.4 to 1.4) lower in intervention group 2 compared to controls. Physical restraint prevalence showed a pronounced variation between the different clusters in all study groups. We found no significant differences in the secondary outcomes. According to the process evaluation, the intervention was mainly implemented as planned, but the expected change towards a least restraint culture of care was not achieved in all clusters. Conclusions: Neither intervention showed a clear advantage compared to control. The pronounced center variation in physical restraint prevalence indicates that other approaches like governmental policies are needed to sustainably change physical restraint practice and reduce center variations in nursing homes. (C) 2019 The Author(s). Published by Elsevier Ltd.