Relationship between lifestyle factors and cardiovascular disease prevalence in Somaliland: A supervised machine...

Cardiovascular diseases (CVDs) are leading contributors to global morbidity and mortality, with low- and middle-income countries experiencing disproportionately high burdens. In Somaliland, urbanization and lifestyle transitions have increased the prevalence of CVDs, necessitating an in-depth exploration of associated risk factors.

This study investigated the relationship between lifestyle factors and CVD prevalence among adult patients in Somaliland using data from the Hargeisa Group Hospital in 2024.

A cross-sectional design was employed, enrolling 411 adults aged ≥ 18 years. Data were collected through structured questionnaires and analyzed using traditional statistical methods and seven supervised machine learning models: Logistic Regression, Random Forest, Support Vector Machine (SVM), Probit Regression, KNN and Decision Tree. The study assessed associations between sociodemographic variables, lifestyle factors, and CVD prevalence while evaluating the predictive model performance.

Age and smoking were the most significant predictors of CVD prevalence across all models, with individuals aged ≥ 60 years exhibiting the highest risk. Urban residence was associated with lower CVD prevalence, while behaviors such as khat chewing and physical inactivity increased the risk. Machine learning models, notably SVM, demonstrated robust predictive performance, achieving an accuracy of 63.4 % and an AUC of 67.1 %.

Lifestyle factors, particularly smoking, khat chewing, and dietary habits, are critical drivers of the CVD prevalence in Somaliland. These findings underscore the need for targeted public health interventions focusing on smoking cessation, dietary improvements, and culturally sensitive awareness campaigns. Machine learning techniques offer valuable tools for enhancing the predictive accuracy and guiding tailored health strategies.