Oral Health Literacy and Preventive Behaviors Among Caregivers of Dependent Older Adults in Sisaket Province, Thailand: A Structural Equation Modeling Study

Background: Dependent older adults experience a high burden of preventable oral conditions, particularly gingivitis, while caregivers’ preventive practices remain insufficiently explained by conventional single-factor models. This study aimed to develop and validate a theory-informed structural equation model (SEM) to explain how oral health literacy and psychosocial determinants shape preventive behaviors for gingivitis among caregivers of dependent older adults in Sisaket Province, Thailand. Methods: A cross-sectional analytical study was conducted among 420 caregivers selected using multistage random sampling. Data were collected using a validated questionnaire based on the Health Belief Model and Social Cognitive Theory, covering knowledge, attitude, perceived severity, perceived benefits, self-efficacy, preventive behaviors, and social support. Reliability was excellent. Descriptive statistics, Pearson’s correlation, and SEM with maximum likelihood estimation were applied. Results: The SEM demonstrated excellent fit (χ2/df = 0.80, p = 1.000; CFI = 1.000; TLI = 1.008; RMSEA = 0.000; SRMR = 0.092). Knowledge directly influenced attitude (β = 0.355, p = 0.013) and self-efficacy (β = 0.381, p = 0.003). Attitude (β = 0.406, p < 0.001), self-efficacy (β = 0.384, p < 0.001), and social support (β = 0.260, p < 0.001) were significant predictors of preventive behaviors. The model explained 68.3% of variance in behaviors, with knowledge exerting significant total effects through mediated pathways (β_total = 0.290, p < 0.001). Conclusions: A theory-based SEM effectively explains caregiver preventive oral health behaviors in a rural, resource-limited setting. Interventions should integrate literacy-focused education with strategies that strengthen attitudes, self-efficacy, and social support to achieve sustainable improvements in dependent older adults’ oral health.