Socio-economic predictors of adolescent marriage and maternity in West and Central Africa between 1986 and 2017

BACKGROUND: Early marriage and maternity prevalence rates among adolescent girls remain alarmingly high in West and Central Africa (WCA). This study aims to explore the associations between socio-economic factors and the prevalence of early marriage and maternity, thus contributing to the identification of girls at risk of early pregnancy or marriage. METHODS: We pooled data from national representative surveys (1986 – 2017) for 23 countries in WCA to examine associations between wealth, educational attainment, religious affiliation, and place of residence with adolescent marriage and maternity. We decomposed the wealth and education gradients for individual countries, while controlling for common characteristics of the local environment via the use of primary sampling unit fixed-effects. The pooled sample provides information on 262 721 girls (age 15-19 years). Survey weights and population share weights were used in the estimations. RESULTS: The prevalence of adolescent maternity and marriage exhibited a wealth and education gradient. Prevalence of marriage in the poorest wealth quintile was 41.1% (95% confidence interval (CI) = 38.8%-43.5%) and 10.5% (95% CI = 9.5%-11.6%) in the richest. For maternity it was 38.3% (95% CI = 36.4%-40.3%) in the poorest quintile and 12.7% (95% CI = 11.5%-13.9%) in the richest. Marriage/maternity is three/two times more likely to occur among girls with incomplete primary or no formal education than in those with at least primary. Maternity and marriage among adolescents exhibit a geographical pattern and differences between religious groups. Adolescent marriage prevalence was 34.4% (95% CI = 32.9%-35.8%) in rural areas compared to 13.3% (95% CI = 12.3%-14.2%) in urban areas. Adolescent maternity prevalence was 32.8% (95% CI = 31.7%-33.9%) in rural compared to 16.3% (95% CI = 15.3%-17.3%) in urban areas. Finally, the prevalence of adolescent marriage was substantially higher among Muslims compared to all other religious groups. CONCLUSIONS: Our results highlight the disparities in the prevalence of adolescent marriage and maternity and confirm the existence of wealth and education gradients. These findings can help to improve targeting of vulnerable adolescents and to identify areas for policy implementation.