Predicting nursing home adherence to a clinical trial intervention: lessons for the conduct of cluster randomized trials

Abstract

Objectives: To describe factors predictive of nursing home (NH) adherence to a clinical trial intervention.

Design: Post hoc analysis of a cluster randomized trial (CRT) evaluating a structured communication intervention to improve nurse-physician telephone communication in NHs.

Setting: NH.

Participants: All eligible licensed nursing staff in all participating NHs.

Measurements: Adherence was defined as active participation for at least 3 months of the 12-month trial. NH characteristics hypothesized to affect trial outcomes (profit status, bed size, nursing staff time, NH quality, and leadership turnover) were measured a priori. The association between intervention adherence, NH characteristics and preintervention questionnaire response rate was examined.

Results: Of 13 intervention NHs, seven adhered to the intervention. Three factors differentiated adherent from nonadherent NHs: director of nursing turnover (nonadherent NHs 50% vs adherent NHs 0%, P = .03); Centers for Medicare and Medicaid Services (CMS) nurse staffing rating (range: 1-5) (nonadherent NHs mean 3.7 ± 0.5 vs adherent NHs mean 4.3 ± 0.5), P = .048); and questionnaire response rate (nonadherent NHs 15.6 ± 10.0% vs adherent NHs 34.2 ± 12.1%, P = .02). Profit status, bed size, and number of NH deficiencies on state surveys were not significantly associated with intervention adherence.

Conclusion: CMS nurse staffing rating, leadership turnover, and questionnaire response rate are associated with adherence to a CRT intervention. Pretrial evaluation of NH staffing rating by CMS and of response to a questionnaire can help investigators improve trial efficiency by screening for NHs likely to adhere to a CRT intervention.