COVID-19 has significantly disrupted global life, leading to high infection rates, considerable mortality, and economic hardships. Mitigation efforts depended on vaccination, preventive behaviors, and lifestyle changes, emphasizing the importance of understanding what influences public compliance. This study investigated COVID-19 preventive behaviors within the Thai Muslim community in Nakhon Si Thammarat Province, southern Thailand, and developed a protection model based on the World Health Organization's framework. Using a multi-stage cluster sampling method, 233 participants from eight villages completed a questionnaire adapted from CDC guidelines. Data were analyzed using Partial Least Squares Structural Equation Modeling (Smart PLS-SEM). Results demonstrated strong reliability (CR > 0.80) despite moderate AVE values (0.314–0.491); HTMT supported discriminant validity. Key predictors of compliance included community nutrition (β = 0.435, p < 0.001) and public trust (β = 0.332, p = 0.001). Additionally, the first service had a significant influence on social support (β = 0.638, p < 0.001), which subsequently affected nutritional status (β = 0.401, p < 0.001). The model explained a moderate amount of variance (R² = 0.232–0.494), showed acceptable fit (SRMR = 0.089), and demonstrated predictive validity (Q² > 0). Overall, findings emphasize the vital roles of nutrition and trust, mediated through services and social support, in encouraging compliance with health protocols.

Geriatric nutritional problems are increasingly characterized by the co-existence of undernutrition and obesity, yet community-level evidence from Southern Thailand remains limited. Grounded in the social determinants of health (SDH) framework, this cross-sectional study examines the prevalence and independent predictors of BMI categories and abdominal obesity among 13,458 community-dwelling older adults across 20 sub-districts in Southern Thailand. Anthropometric measurements and structured interview data were analyzed using chi-square tests and multinomial logistic regression, with normal weight as the reference category. Overall, 50.3% of participants were overweight or obese, 40.5% had abdominal obesity, and 7.7% were underweight. Advancing age, continued employment, and asthma were independently associated with higher odds of underweight, whereas abdominal obesity was the strongest predictor of elevated BMI categories, increasing the odds of overweight, obesity Class I, and obesity Class II. Female sex, literacy, dyslipidemia, hypertension, and diabetes were also associated with excess BMI categories. Moderate, regular physical activity was a protective factor. These findings support risk-stratified community interventions that integrate BMI with WC screening and address regional occupational, behavioral, clinical, and food culture determinants of nutritional health in older adults.