Personalized Machine Learning Intervention to Improve Sleep Quality Using Wearable Technology in Healthy Middle-Aged...

Area of Science:

• Sleep Science
• Artificial Intelligence in Healthcare
• Digital Health

Background:

• High prevalence of sleep dissatisfaction globally, with significant percentages of adults reporting poor sleep quality and lack of energy.
• Consistent reports of sleep problems and dissatisfaction across diverse adult populations, including in Mexico.
• Growing need for effective interventions to improve sleep quality in the general adult population.

Purpose of the Study:

• To test the efficacy of a personalized sleep intervention driven by machine learning (ML) using consumer wearable data.
• To compare the effectiveness of ML-personalized sleep recommendations against generic sleep hygiene education.
• To evaluate sleep score improvements in healthy middle-aged adults from Mexico City.

Main Methods:

• Pilot randomized controlled trial (RCT) with 32 participants, stratified by sex.
• 60-day data collection using Samsung Galaxy Watch 4 devices for objective sleep metrics.
• Machine learning models (Shapley Additive Explanations, recursive feature elimination) to identify key sleep parameters for personalized recommendations.

Main Results:

• The study is ongoing, with data collection expected by December 2025 and results anticipated by March 2026.
• The primary outcome is the change in sleep score (1-100) from baseline to intervention phase.
• Secondary outcome includes changes in the Pittsburgh Sleep Quality Index (PSQI) global score for subjective validation.

Conclusions:

• Establishes feasibility and preliminary effect size of ML-personalized sleep interventions via consumer wearables.
• Suggests personalized sleep interventions may outperform generic advice for improving sleep quality.
• Provides a framework for scalable, individualized sleep health interventions.