Leveraging an AI-Driven Voice Companion to Enhance Mental Health in Older Adults in Post-Acute and Long-Term Care Settings: A Feasibility Study

Loneliness, depression, and anxiety are prevalent among older adults in post-acute and long-term care settings and are associated with adverse health outcomes. Existing psychosocial interventions often face challenges related to scalability and sustained engagement. This preliminary feasibility study assessed an artificial intelligence-based, voice-driven telephone intervention designed to facilitate conversational engagement with older adults. Thirty-three residents were approached, and 28 enrolled, primarily from skilled nursing facilities and assisted living facilities, with additional participation from independent living settings; 23 completed both baseline and follow-up assessments over a 4-week assessment period. Participants demonstrated reductions in depressive and anxiety symptoms, with larger changes observed among those with higher baseline severity. Participants with Patient Health Questionnaire-9 scores ≥10 showed an average reduction of 5.7 points, and overall, Generalized Anxiety Disorder 7-item scale scores decreased by a median of 2.3 points. Given the exploratory nature and methodologic limitations of the study, these initial findings should be interpreted cautiously. Further research is warranted to assess the intervention's effectiveness and generalizability in larger, controlled trials.