Revisiting the health belief model: The role of fertility preferences and health system interaction in contraceptive use in Pakistan

BACKGROUND: Unintended pregnancies remain a major public health challenge in Pakistan, driven in part by low and inconsistent contraceptive use. While prior research has largely focused on demographic and access-related determinants, less attention has been given to the role of health beliefs within a broader reproductive context. This study applied a revised Health Belief Model (HBM) to examine how perceived threat and perceived benefits are associated with contraceptive use among married women in Pakistan, alongside fertility preferences and health system interaction. METHODS: This study used data from the Pakistan Demographic and Health Survey (PDHS 2017-18), including 14,494 currently married women aged 15-49 years with complete data. Partial Least Squares Structural Equation Modelling (PLS-SEM) was used to estimate associations between HBM constructs and current contraceptive use. Psychological constructs were operationalized using proxy indicators, while health system interaction was captured through contact with family planning providers and facilities. Three models were estimated: a baseline HBM model, a model adjusted for demographic and socioeconomic characteristics, and an extended model incorporating fertility preferences and partner-related factors. Model performance was assessed using R², Cohen's f² effect sizes, and bootstrapped confidence intervals based on 5,000 resamples. RESULTS: Perceived threat (β = 0.045; 95% CI: 0.029, 0.061) and perceived benefits (β = 0.018; 95% CI: 0.002, 0.038) were positively associated with contraceptive use, but with small effect sizes. Health system interaction remained consistently associated across models (β = 0.079; 95% CI: 0.063, 0.096). In the extended model, fertility-related variables showed the largest magnitude associations, particularly preferred birth spacing (β = 0.501; 95% CI: 0.451, 0.550) and pregnancy intention (β = -0.298; 95% CI: -0.349, -0.245), followed by births in the last five years (β = 0.096; 95% CI: 0.079, 0.113) and husband's fertility preference (β = -0.090; 95% CI: -0.103, -0.075). Model explanatory power increased across specifications, with R² rising from 0.032 in the baseline model to 0.138 in the extended model. CONCLUSIONS: Contraceptive use in Pakistan is more strongly associated with fertility preferences and health system interaction than with health beliefs alone. The limited contribution of perceived threat and perceived benefits highlights the need to move beyond awareness-based approaches toward interventions that strengthen service delivery and align with women's reproductive intentions. Programs that improve access to providers, support birth spacing goals, and address partner dynamics may be more effective in increasing contraceptive uptake. Given the cross-sectional design, findings should be interpreted as associations rather than causal relationships.