Smartphone Adaptive Intervention Using Reinforcement Learning to Help Adults Aged 18 to 40 Reduce Smoking and Cannabis Use

What this Trial Is About

This research aims to develop and test a smartphone-based adaptive intervention to help adults aged 18 to 40 reduce smoking and cannabis use. The study focuses on tobacco use, a leading cause of preventable death in the United States, and evaluates a reinforcement learning RL algorithm designed to personalize treatment messages in real-time based on user context such as location and stress levels. This approach builds on previous work that showed feasibility in reducing smoking urges using cognitive-behavioral and mindfulness-based therapies delivered via smartphone. Participants first complete a 14-day training phase using a smartphone app that prompts them to answer surveys up to three times daily and passively collects GPS data. This data identifies high-risk locations and times, which inform the RL algorithm during a subsequent 30-day intervention phase. During this phase, the app delivers personalized messages based on cognitive-behavioral therapy, acceptance and commitment therapy, or attention control, triggered when participants enter designated high-risk areas. Throughout the 44-day study, participants smoking urges and cigarette use are assessed through surveys shortly after message delivery and at baseline and follow-up. Usability of the app is measured using the System Usability Scale at the end of the intervention. Additional outcomes include changes in stress, negative feelings, message helpfulness, activity completion, smoking abstinence, psychological flexibility, and tobacco product switching. The study monitors participants remotely using smartphone data and questionnaires to evaluate the interventions impact and user experience.

Adaptive Mobile Interventions to Reduce Cancer Risk Behaviors