AI-Driven Digital Therapeutic for Insomnia: Development and Clinical Validation

Insomnia is a prevalent global health problem, with approximately one in five Taiwanese adults chronically affected. Prolonged use of sedative-hypnotic medications carries risks of adverse effects, dependence, and unnecessary healthcare utilization. Although cognitive behavioral therapy for insomnia (CBT-I) is internationally recommended as first-line treatment, workforce shortages and suboptimal patient adherence limit its reach. This project will develop an artificial intelligence-driven mobile application for insomnia that implements the CBT-I components of stimulus control and sleep restriction. Using retrieval-augmented generation and large language models, we will build a localized sleep knowledge base to deliver personalized, interactive sleep guidance. A pilot randomized controlled trial will assess the intervention's effectiveness in reducing insomnia severity and sedative-hypnotic use. We will subsequently pursue a large-scale clinical trial and iterative model optimization, aiming to deliver the first clinically validated AI-enabled prescription digital therapeutic for insomnia and to advance smart healthcare in sleep care.

Inclusion Criteria: * Aged 18 years or older * An Insomina Severity Index(ISI) score of 8og higher. * Diagnosed with insomnia. * Ownership of a smartphone. * Ability to operate the application and complete online questionnaires independently. * Chinese communication. Exclusion Criteria: * shift worker. * people with combined mental illnesses. * alcohol dependence * unable to use a phone due to cognititive impairment.