Smart speaker-related perceived companionship and loneliness reduction among oldest-old care facility residents: a 12-week quasi-experimental study

Loneliness is a common mental health concern among oldest-old residents of care facilities. Smart speakers are emerging low-cost tools for supporting older adults, but evidence on their potential role in loneliness reduction remains limited. This quasi-experimental study recruited 140 residents aged ≥ 80 years from six care facilities in a provincial capital city in China. Facilities were cluster-allocated by lottery to a smart speaker intervention group ( n = 70) or a waitlist control group ( n = 70). The intervention lasted 12 weeks, with assessments at baseline, week 6, and week 12. Loneliness was measured using the short-form UCLA Loneliness Scale. Smart speaker-related perceived companionship was assessed using the SSPCS. Intervention-group SSPCS responses reflected actual use, whereas waitlist-control responses reflected anticipated perceptions after a standardized neutral introduction. Exploratory moderated mediation analysis was interpreted cautiously. A total of 132 participants completed all assessments. Loneliness in the intervention group decreased from 19.4 to 14.5 points, with a large between-group effect size at week 12 (Cohen’s d = 1.02). Smart speaker-related perceived companionship increased in the intervention group, whereas anticipated companionship-related perceptions remained relatively stable in the waitlist control group. Exploratory moderated mediation analysis suggested a potential explanatory pathway through smart speaker-related perceived companionship. Participants with better cognitive function and higher baseline loneliness appeared to show greater intervention-related benefits. Sensitivity analyses showed a generally consistent pattern, but did not fully account for within-facility clustering. Smart speakers may help reduce loneliness among oldest-old care facility residents by providing accessible voice-based interaction and companionship-related experiences. These findings provide preliminary support for their use in care facilities, while causal mediation and facility-level clustering should be addressed in future studies. Chinese Clinical Trial Registry, ChiCTR2025030411, registered 09 March 2026. Retrospectively registered.