Bayesian Modelling of Family Planning Discontinuation under Socioeconomic and Environmental Uncertainty in Rural Communities in South East Nigeria

Abstract

Despite increasing contraceptive awareness, discontinuation rates continue to contribute to unintended pregnancies, maternal health risks, and population pressure. This study investigated the determinants of family planning discontinuation among women of reproductive age in South-East Nigeria using a hierarchical Bayesian Stochastic modelling framework. The study utilised the IRPGS dataset collected from the five states of South-East Nigeria, comprising household-level reproductive-health, socioeconomic, migration, healthcare accessibility, and environmental variables. A hierarchical Bayesian logistic regression model was employed to estimate the probability of contraceptive discontinuation while accounting for latent state-level heterogeneity. Fixed effects included age, healthcare accessibility, food price instability, flood exposure, migration experience, drought exposure, educational attainment, and family planning knowledge, while state-level random effects captured contextual variability. Posterior estimation, convergence diagnostics, sensitivity analysis and predictive information criteria,  were used for model validation. The findings revealed that environmental shocks significantly influenced contraceptive discontinuation behaviour. Drought exposure substantially increased discontinuation risk, while flood exposure also exhibited a positive association with discontinuation. Migration experience emerged as a major protective factor, significantly reducing discontinuation probability. Increased travel time to healthcare facilities slightly elevated discontinuation risk. State-level random effects indicated substantial geographic heterogeneity, with Ebonyi, Enugu and Imo States exhibiting higher latent discontinuation vulnerability compared with Abia and Anambra States. Convergence diagnostics showed satisfactory posterior convergence with Gelman-Rubin statistics below 1.1, while sensitivity analysis confirmed robustness of posterior estimates. The study concludes that family planning discontinuation in South-East Nigeria is shaped by complex interactions among environmental uncertainty, healthcare accessibility, migration dynamics, and latent geographic factors.