Related Experiment Video
Updated: Mar 29, 2026
07:24
Home-Based Monitor for Gait and Activity Analysis
Published on: August 8, 2019
7.4K
Prediction of Adherence to an Online Wellness Program for People with Mobility Limitations: A Machine Learning
Salma Aly1, Hui-Ju Young2,3,4, James H Rimmer2,3,4,5
1Department of Family and Community Medicine, Heersink School of Medicine, University of Alabama at Birmingham, Birmingham, AL 35205, USA.
Healthcare (Basel, Switzerland)
|March 28, 2026
Summary
Digital wellness program adherence for individuals with mobility limitations is modestly predicted by baseline factors. Psychosocial and socioeconomic elements are key, suggesting personalized support can reduce dropout.
Area of Science:
- Digital health interventions
- Machine learning in healthcare
- Behavioral science
Background:
- Individuals with mobility limitations experience higher chronic disease rates and lower adherence to wellness programs.
- Digital telewellness programs like MENTOR offer accessibility but struggle with participant attrition.
- Predicting and understanding adherence is crucial for improving program effectiveness.
Purpose of the Study:
- To apply machine learning (ML) to predict adherence to the MENTOR telewellness program.
- To identify key baseline predictors of participant attendance and engagement.
- To explore interactions between predictors for a deeper understanding of adherence patterns.
Main Methods:
- Utilized data from 1218 adults enrolled in the MENTOR program (2023-2024).
- Trained 13 ML regression models on demographic, socioeconomic, psychosocial, and health variables to predict session adherence.
- Employed SHAP and synergy analyses to interpret model predictions and identify significant predictors and their interactions.
Main Results:
- Bayesian ridge regression demonstrated the best predictive performance (MAE 20.98).
- Key predictors included education, race, emotional support, Area Deprivation Index, household size, mindfulness, life satisfaction, and disability onset.
- Higher emotional support, mindfulness, and life satisfaction correlated with increased adherence, while socioeconomic disadvantage predicted lower adherence.
Conclusions:
- Baseline characteristics offer modest prediction of digital wellness program adherence.
- Psychosocial factors (emotional support, life satisfaction) and socioeconomic status significantly influence adherence.
- Personalized support strategies targeting these factors are essential to mitigate dropout rates in telewellness programs for individuals with mobility limitations.
