Naphtal Nyirimanzi

Using mobile health to predict asthma exacerbations in children and adolescents: The Mobile Health for Kids with Asthma (MoKA) study
Asthma exacerbations are a leading cause of emergency visits in children, yet many are preventable with timely intervention. Current predictive tools focus on a single aspect of asthma and forecasting over the medium or long term. The Mobile Health for Kids with Asthma (MoKA) study aims to improve prediction by using real-time, multimodal data through a mobile health (mHealth) solution—the RespiSentinel app. This prospective observational study, conducted across seven pediatric centers in Canada, will recruit 2,000 children aged 1–17 years with history of exacerbation during previous year. Over six months, the study will collect self-reported symptoms, environmental data, and automated nighttime cough recordings via the app.
MoKA’s objectives are to: (1) develop a machine learning–based model predicting severe asthma exacerbations using multimodal data, (2) explore the relationship between nocturnal cough frequency and exacerbations, and (3) assess the app’s acceptability in both current and future forms.
Data from the RespiSentinel app—including questionnaire responses, ambient air quality, and cough audio—will be analyzed. A predictive model will be trained using an 80:20 data split, applying binomial logistic regression and Random Forest methods. Cough frequency will be linked to asthma outcomes via generalized estimating equations. The study also evaluates app engagement and user satisfaction through interaction metrics and questionnaire-based feedback.
Expected outcomes include a validated weekly prediction algorithm and enhanced parental empowerment. The approach will contribute to reduction of asthma exacerbations, emergency visits, and healthcare costs, ultimately improving the lives of children with asthma and their families.