Moving behavioral interventions in nursing homes from planning to action: a work system evaluation of a urinary tract infection toolkit implementation.

Background: Implementation evaluations based on a hybrid deductive-inductive approach provide a detailed understanding of organizational choices to introduce and implement complex interventions and may help explain implementation success or failure. However, such evaluations may not be feasible due to resource constraints. Qualitative analyses of artifacts collected for other purposes during implementation may represent a cost-effective method to understand program implementation when robust evaluations are not feasible. This study used a work systems evaluation of how nursing homes (NHs) implemented a urinary tract infection (UTI) recognition and management improvement toolkit.

Methods: Thirty NHs participated in a randomized control trial in which intervention NHs (n = 12) were assigned a clinical coach who employed a standard template to structure coach calls with the NH champion. A hybrid inductive-deductive approach, using the Systems Engineering Initiative for Patient Safety (SEIPS) model, characterized three action domains related to (1) engagement of staff and providers, (2) distribution of toolkit elements, and (3) toolkit use.

Results: A total of 369 coded segments from 148 coach notes generated by three coaches working with 18 NH champions were examined. Planned changes (n = 203) were more frequent compared to actual changes (n = 169). While most NHs quickly engaged staff and providers, which leadership appeared to support, engagement actions were hindered in some NHs due to champion instability or extended champion or medical director absences. Dissemination of materials to family and providers and distribution of tools to staff occurred quickly in 75% of NHs, although delays were encountered in some NHs, usually because of champion instability.

Conclusions: Implementing NH practice change is challenging, and studies examining actions to support planned versus actual change in this setting are limited. The application of the SEIPS model to coach notes collected during the implementation of a structured behavioral intervention to improve the recognition and management of UTI in NHs generated unique insights into the work system and how staff attempted to implement changes. This study identified several factors that interfered with progression from planning to actual change. Future studies are needed to better understand how to best support change interventions in NHs.

Trial Registration: ClinicalTrials.gov, NCT03520010 , Registered May 9, 2018.