Analyzing factors affecting age at first birth among married women in Somalia: a Bayesian shared frailty modelling approach using SDHS 2020.

Abstract

This study examines factors influencing the age at first birth among married women in Somalia, focusing on sociodemographic, economic, and health-related determinants. Given Somalia's fragmented health system and high maternal mortality rates, understanding first birth timing is critical for achieving Sustainable Development Goal 3 on maternal and child health. A Bayesian shared frailty model was employed to analyze variations in birth timing, comparing Weibull, log-normal, and log-logistic models using DIC and WAIC values. The results indicate that age at first marriage is the strongest predictor, with a higher age significantly reducing the hazard of first birth (HR = 0.4636, 95% CI: 0.4399-0.4886, p < 0.001). Residence also shows a significant effect, where women in rural areas experience delayed first births (HR = 0.9411, 95% CI: 0.9026-0.9813, p < 0.01). Other factors, including region, education, wealth, contraceptive use, and marital status, were not statistically significant, while the husband's desire for children had a weak association with first-birth timing. The Bayesian log-logistic AFT shared frailty model best predicted age at first birth. Regions like Banadir were linked to earlier births, while higher education, greater wealth, and later age at first marriage were the strongest predictors of delayed childbirth. Contraceptive use, marital status, and media access had minimal impact. Socio-demographic and economic factors, especially age at marriage and education, are key determinants. Finally, the study highlights the influence of social and family dynamics on reproductive health, underscoring the need for targeted interventions to delay early childbearing and improve maternal and child health in Somalia.