Zhenti Cui, Rongrong Guan, Tianrui Xue, Yongchang Sun, Maw Pin Tan, Mas Ayu Said, Yingdong Cao
The care package was feasible and produced a preliminary behavioral signal, but medication effects were imprecise, and functional improvement was not sustained. A larger cluster trial with an attention-control condition is required.
BACKGROUND: Doll therapy is a person-centered intervention for behavioral and psychological symptoms of dementia, but evidence from Chinese nursing homes and information about implementation, medication use, and ethical acceptability remain limited.
METHODS: We conducted an exploratory pilot and feasibility mixed-methods cluster-randomized clinical trial in six nursing homes in Zhengzhou, China. Homes were randomized 1:1 to a staff-mediated doll-therapy care package plus routine care or routine care alone. Residents with dementia were assessed at baseline, 6 weeks, and 12 weeks. Registered primary outcomes were class-specific standardized daily psychotropic doses; NPI-12 was the main proximal clinical outcome. Feasibility, fidelity, safety, and interviews with intervention-home staff and family caregivers were also evaluated. Marginal models used nursing home as the top-level cluster and the Mancl-DeRouen small-sample variance correction.
RESULTS: Among 185 residents screened, 142 were enrolled, and 130 (91.5%) completed 12-week follow-up. Sixty-eight of 71 intervention participants (95.8%) attended at least 6 of 8 sessions; mean fidelity was 94.1%. The intervention-control difference in NPI-12 change was -3.61 points at 6 weeks (95% confidence interval [CI], -5.86 to -1.36) and -4.40 at 12 weeks (95% CI, -6.50 to -2.30). The Barthel difference was 3.96 points at 6 weeks (95% CI, 1.06 to 6.86) and 3.07 at 12 weeks (95% CI, -2.33 to 8.47). All medication-dose intervals showed no difference. Six possible or probable related events were mild or moderate; no serious event was related. Interviews with 15 staff and 9 family caregivers indicated engagement for some residents, absent or adverse responses for others, and the importance of training, non-deceptive introductions, prompt respect for refusal, and operational ownership.
CONCLUSION: The care package was feasible and produced a preliminary behavioral signal, but medication effects were imprecise, and functional improvement was not sustained. A larger cluster trial with an attention-control condition is required.
TRIAL REGISTRATION: ClinicalTrials.gov, NCT06506487; https://clinicaltrials.gov/study/NCT06506487.