Age at first marriage among women in Lesotho: Multilevel survival analysis- insight from 2023-24 demographic and health survey data

Abstract

Background Early marriage remains a significant public health and social issue in Lesotho, with profound implications for women's health, education, and economic empowerment. Despite legal frameworks aimed at curbing the practice, a substantial proportion of women continue to marry before the age of 18, driven by sociocultural norms, economic constraints, and limited access to opportunities. This study investigates the timing and factors associated with first marriage among women in Lesotho using advanced multilevel survival analysis to inform targeted interventions. Methods We analyzed retrospective time-to-event data from the nationally representative, cross-sectional 2023−24 Lesotho Demographic and Health Survey (LDHS), employing a log-logistic accelerated failure time (AFT) model within a multilevel framework to account for individual, household, and community-level factors. The analysis included 6,413 women aged 15–49, with 48.76% having experienced first marriage (events) and 51.24% right-censored. Statistical computations were performed using R statistical software (version 4.4.3), incorporating survey weights to ensure representativeness. Results The median age at first marriage was 24 years (95% CI: 24–25), with significant variation across socioeconomic and contextual factors. Employment (TR = 1.072, 95% CI: 1.043–1.103), middle wealth quintile (TR = 1.060, 95% CI: 1.012–1.110), low community poverty (TR = 1.094, 95% CI: 1.045–1.145), and urban residence (TR = 1.079, 95% CI: 1.033–1.126) were statistically associated with later age at first marriage (p < 0.05). In contrast, primary education (TR = 0.856, 95% CI: 0.749–0.979), younger birth cohorts (1979–2003: TR range 0.774–0.907), richest wealth quintile (TR = 0.927, 95% CI: 0.875–0.981), and rural residence (TR = 0.927, 95% CI: 0.888–0.967) were associated with earlier age at first marriage. The intraclass correlation coefficient (ICC = 0.0029) indicated that only 0.3% of the variation in age at first marriage was attributable to community-level differences. Although statistically significant, the magnitude of these associations is modest. Conclusions The findings show that employment, middle wealth status, low community poverty, and urban residence are associated with a later age at first marriage, while primary education, younger birth cohorts, richest wealth quintile, and rural residence are associated with earlier marriage. The minimal community-level variation suggests that individual and household-level factors are more important determinants of marriage timing. Policymakers could consider prioritizing multisectoral interventions, including expanded educational access, economic empowerment programs, and poverty reduction strategies, with particular attention to rural-urban disparities. Addressing these structural factors may be relevant to achieving sustainable development goals related to gender equality and women's empowerment in Lesotho.