Abstract
BACKGROUND: Digital health literacy is a critical digital determinant of health (DDoH) in the Arab Middle East and North Africa (MENA) region, where technological disparities, limited healthcare infrastructure, and diverse socio-cultural contexts significantly impact healthcare access and management. OBJECTIVE: This study evaluates the impact of digital health literacy on health management practices and ensuing health outcomes in Arab Countries, employing predictive modeling as an analysis tool to uncover key determinants. METHODS: A cross-sectional survey of 12,522 respondents from ten Arab MENA countries was analyzed to examine relationships between survey features and health outcomes. We compared multinomial regression to machine learning models, including CatBoost and Random Forest, to predict outcomes and identify significant predictors. RESULTS: CatBoost, a powerful ML model that handles categorical data efficiently, achieved a predictive accuracy of 97.8%, outperforming other models in capturing complex, nonlinear relationships. Five key determinants of digital health literacy on health management outcomes were identified: limited internet access, restricted health service access, confidence in AI health resources, health monitoring tool usage, and social media health information consumption. CONCLUSION: Enhancing digital health literacy is critical for improving healthcare outcomes in the Arab MENA region. This study underscores the need for culturally tailored digital health interventions to address regional technological and healthcare challenges. Policymakers must prioritize these strategies to reduce disparities and empower individuals in managing their health.