Intervention Without Borders - an Automated Self-Guided AI-Enhanced Psychoeducation Intervention for Dementia...

To examine whether a fully automated, self-guided intervention (PDC30) could improve caregiver well-being over a 1-month waitlist control in an international sample.

Randomized waitlist-controlled trial.

Web-based platform accessible globally.

441 individuals responded to study promotion on the internet, of whom 274 from 43 countries met the study criteria and were randomized. Eligible participants were adults providing ≥10 care hours weekly to community-dwelling relatives with dementia, scoring ≥5 on Patient Health Questionnaire-9 (PHQ-9), and without recent caregiver intervention.

Available 24/7, PDC30 is a self-guided, automated intervention consisting of a Guidebook, an AI-powered counseling chatbot, and interactive applications for cognitive-behavioral techniques, relaxation, and caregiver-recipient bonding.

At baseline and follow-ups at 1, 2, and 3 months, depression was assessed by PHQ-9. Secondary outcomes were measured with validated brief versions of anxiety, burden, and positive gains.

Intent-to-treat analysis using mixed-effects regression showed treatment x time2 effects on all outcomes except anxiety. At 1-month follow-up, coinciding with exclusive access to PDC30, intervention caregivers showed significant improvements in depression (d = -0.37), burden (d = -0.34), and positive gains (d = 0.42). The differences mostly disappeared after control participants received the intervention, while improvements in both groups were sustained thereafter. Participants reported using the website several times weekly, were generally satisfied with it, and found the chatbot most helpful.

The effects on depression and other outcomes were consistent with those observed for in-person programs, suggesting the viability of well-designed automated intervention. The study demonstrates the feasibility, acceptability, and potential global health impact of PDC30.