Testing a Machine Learning-Based Adaptive Motivational System for Socioeconomically Disadvantaged Smokers...

Area of Science:

• Public Health
• Digital Health
• Behavioral Science

Background:

• Socioeconomically disadvantaged individuals exhibit high smoking rates and face barriers to cessation interventions.
• Computer-tailored health communication shows potential for promoting behavior change.
• The Adapt2Quit recommender system, a machine learning approach, was developed to enhance message relevance and smoking cessation effectiveness.

Purpose of the Study:

• To evaluate the efficacy of the Adapt2Quit recommender system in increasing smoking cessation among socioeconomically disadvantaged smokers through a randomized controlled trial.

Main Methods:

• Recruitment of socioeconomically disadvantaged smokers from low-income insurance groups and community health centers.
• Intervention group received tailored text messages via the Adapt2Quit system for 6 months, including facilitation and status assessment messages.
• Control group received only facilitation and status assessment messages; primary outcome measured by 7-day point-prevalence smoking cessation at 6 months, verified by carbon monoxide testing.

Main Results:

• Recruitment of 757 participants, with a diverse sample including 64% female, 35% Black or African American, and 51.1% White individuals.
• Participant characteristics include 52.6% with a high school education, 70% smoking within 30 minutes of waking, and 50% having abstained from smoking for at least one day in the past year.
• The study is ongoing, with recruitment and 6-month follow-up completed; main analysis is pending.

Conclusions:

• A diverse sample of socioeconomically disadvantaged smokers was recruited for the Adapt2Quit study.
• A rigorous protocol was established to evaluate the Adapt2Quit recommender system's effectiveness.
• Future publications will present the primary analysis of the randomized controlled trial data.