Perceived knowledge, attitudes, practices, and compliance to therapeutic regimen among individuals with Diabetes Mellitus in Metro Manila

Diabetes mellitus (DM) management extends beyond clinical care, requiring sustained behavioral engagement shaped by social and contextual factors. Understanding how perceived knowledge, attitudes, and practices influence compliance is critical in urban low- and middle-income settings. This study examined the associations among perceived knowledge, attitudes, practices, and their relationship with self-reported compliance to therapeutic regimens among adults diagnosed with DM in Metro Manila. A cross-sectional online survey of 400 adults with self-reported clinician-diagnosed diabetes was conducted using a culturally adapted KAP instrument rated on a 5-point Likert scale. Domain-level medians and interquartile ranges were computed, with Spearman's correlations and exploratory regression analyses used to assess associations. A sensitivity analysis was performed to evaluate analytic adequacy. Respondents reported high perceived knowledge and frequent self-care practices, but only moderate compliance. Attitudes demonstrated the strongest positive association with compliance (ρ = 0.807, p < .001) and were the sole significant predictor in regression models (β = 0.831, p < .001). Perceived knowledge and practices showed weaker, non-significant effects when attitudes were controlled. Covariate-adjusted regression controlling for age, sex, education, and district confirmed attitudes as the strongest correlate, with demographic factors exerting modest effects. Findings suggest that in this urban LMIC context, attitudinal orientations, including motivation, perceived barriers, and self-efficacy, are more closely linked to treatment compliance than perceived knowledge or general practices. Interventions should therefore incorporate attitude-focused behavioral strategies alongside education and address structural challenges affecting sustained diabetes care. • Attitudes strongly predict compliance, surpassing knowledge and practices. • Covariate-adjusted model shows modest effects of age, sex, education, location (district). • Perceived knowledge alone is insufficient for sustained diabetes adherence. • Findings support equity-focused, culturally tailored interventions in LMICs. • Ordinal-appropriate methods enhance rigor in KAP-based behavioral research.