Perceptions and effectiveness of a fully-automated brief behavioral insomnia therapy, delivered by a virtual companion, in older and young adults

Abstract

Background and Objectives: One of the most common sleep disturbances in older people is insomnia. Cognitive-behavioral therapy is the first-line treatment for this condition in older adults, but in-person treatment is costly and often unavailable. In this study, in a group of older and young subjects, we aimed to compare: (1) their initial perceptions of a fully-automated mobile health intervention to manage insomnia, (2) how these perceptions related to treatment completion, and (3) the effects of the intervention on insomnia severity and related outcomes. Research Design and Methods: A case series study was conducted with a self-selected sample of older (>65 years) and young (18-35 years) adults (n=5,660) who downloaded a free app, available in France, that delivers a brief behavioral intervention for insomniaaided by a virtual companion. The 17-day intervention included sleep hygiene and stimulus control recommendations. Primary outcome was treatment completion (yes/no). At the beginning of the intervention, treatment acceptability and trust in the virtual companion were assessed with two short questionnaires (completion rate: 1,597 users). Insomnia was evaluated with the Insomnia Severity Index. Results: Logistic regression analyses showed that higher credibility and trust in the app’s virtual companion were associated with higher odds of treatment completion, but only in older adults (Trust scores x Age group: OR=1.12, [95%CI=1.01-1.25], p < .05, and Credibility scores x Age group: OR=1.25, [95%CI=1.06-1.47], p < .01). Within the subset of users who completed the intervention (n=289), insomnia remission (χ2=2.72, NS) and insomnia response rates (χ2=2.34, NS) were comparable across both groups. Discussion and Implications: This brief behavioral intervention appears to be efficacious for the self-managementof insomnia symptoms in older adults. The integration of persuasive interaction elements, such as avatars and virtual coaches, in fully-automated interventions could be particularly useful to stimulate older adults’ engagement.