A Machine Learning-Based Mobile Health Application for Early Stroke Risk Prediction and Prevention

Stroke remains a leading cause of mortality and disability globally, necessitating effective tools for early risk detection and prevention. This study details the development and evaluation of My-Stroke Guard, a machine learning-based mobile health application designed as an Electronic Personal Health Record (ePHR) for early stroke risk assessment. The objective was to create a user-centered tool that enables individuals to monitor key health indicators and receive personalized risk predictions. Following the Rapid Application Development (RAD) methodology, the system was built for the Android platform, incorporating a machine learning model to analyze user-inputted data including blood pressure, BMI, and lifestyle factors. Usability was evaluated with two distinct user groups: young adults and the elderly. The results demonstrate the system’s efficiency, with the algorithm generating risk assessments in under two seconds. Younger users reported high satisfaction (88.05%), finding the interface intuitive and responsive. Conversely, elderly users reported lower satisfaction (65.15%), facing significant challenges with text readability and navigation. These findings underscore the potential of mobile health applications in promoting preventative health but also highlight the critical need to address digital accessibility barriers for older, high-risk populations to ensure equitable and effective deployment in public health initiatives.