0914 Catching Z’s: Development of a Chatbot-Based Sleep Intervention for Adolescents with Autism Spectrum Disorder

Abstract Introduction Sleep disturbances are highly prevalent among autistic adolescents and adversely affect neurobehavioral, emotional, and social functioning. Artificial intelligence chatbots show promise for delivering scalable sleep coaching, but have not been adapted for this population. This study aimed to design and pilot-test Catching Z’s, an app-based, chatbot-integrated sleep intervention for autistic adolescents with poor sleep health. Methods Building on a CBT-I–based chatbot our team previously developed for young adults, we developed Catching Z’s using a design-thinking approach involving: (1) two focus groups with 16 autistic adolescents and caregivers to assess needs and preferences for an AI-chatbot; (2) expert consultation with sleep and autism specialists to develop the prototype; and (3) think-aloud usability testing with five autistic adolescents to refine the intervention. We then conducted a 4-week, single-arm beta test. Feasibility, usability, and acceptability were evaluated using completion and retention rates, the Adapted System Usability Scale (SUS), acceptability ratings, and post-intervention interviews. Results Insights from focus groups and prototype testing informed intervention components: (1) a rule-based chatbot delivering four structured sleep-coaching modules with visual supports and sensory-friendly design; (2) a natural-language–processing chatbot for real-time Q&A; (3) daily sleep tracking with graphic feedback; and (4) weekly to-do lists and milestone badges. For beta-testing, five autistic adolescents (ages 14–17; three female) have been enrolled to date, with a target sample of 10-15. Personalized coaching schedules and weekly summaries were provided to parents. All adolescents completed the 4-week intervention; three followed the schedule and two required additional reminders. The average SUS score (67) approached the favorable usability threshold (68), and 80% reported confidence using the chatbot to support sleep improvements. All adolescents rated the intervention acceptable or highly acceptable. In interviews, all adolescents reported improvements in sleep knowledge, engagement in sleep hygiene practices, and overall sleep quality. They described the app as user-friendly and flexible while recommending enhancements, including more concise dialogue, richer visuals, and adaptive reminders. Conclusion The chatbot-based sleep intervention appears feasible and acceptable, with autistic adolescents providing valuable suggestions for refinement. Catching Z’s also shows promising potential for improving sleep health, warranting further efficacy trials. Support (if any) Maggie E. Neumann Health Sciences Research Fund pilot research grant.