Maternal education and infant survival in Somaliland: a descriptive decomposition of observed survival differences across sequential adjustment models.

BACKGROUND: Infant mortality remains a major public health challenge in Somaliland, a fragile setting with limited health infrastructure. Although maternal education has consistently been associated with improved child survival, it remains unclear whether infant survival differs across maternal education levels and how this association changes after accounting for health service utilization and household environmental conditions. This study examined the association between maternal education and infant survival, with particular attention to whether secondary education was associated with greater infant survival than lower educational attainment. A secondary objective was to descriptively decompose the observed association using the Karlson-Holm-Breen (KHB) method. METHODS: This study was a secondary analysis of nationally representative data from the 2020 Somaliland Health and Demographic Survey (SLHDS). Cox proportional hazards regression models were used to estimate the association between maternal education and infant survival through sequential adjustment for socioeconomic, health service, and environmental factors. The Karlson-Holm-Breen (KHB) method was applied as a descriptive decomposition approach to quantify the observed attenuation in the association between maternal education and infant survival after adjustment for antenatal care (ANC), skilled birth attendance (SBA), improved drinking water, and improved sanitation facilities. RESULTS: Maternal secondary education or higher was associated with a 65.7% lower hazard of infant death in the unadjusted model (HR = 0.343, 95% CI: 0.15-0.81, p < 0.05), whereas primary education was not significantly associated with infant survival. A post-estimation Wald test demonstrated a significant difference between primary and secondary education in the unadjusted model (p = 0.033). After adjustment for socioeconomic, health service, and environmental factors, the association for secondary education was attenuated and was no longer statistically significant (HR = 0.501, 95% CI: 0.20-1.26; Wald test: primary = secondary, p = 0.124). Descriptive decomposition indicated that 20.6% of the observed attenuation in the association was associated with the variables examined, with adjustment for skilled birth attendance contributing the largest share (19.7%). Significant regional differences in infant survival were also observed. CONCLUSION: Maternal secondary education was associated with improved infant survival in unadjusted analyses, whereas primary education was not. This association was attenuated and became statistically non-significant after adjustment for socioeconomic, health service, and environmental factors. Descriptive decomposition indicated that adjustment for skilled birth attendance contributed the largest share of the observed attenuation. Because these findings are based on observational survey data, the decomposition should be interpreted as describing changes in statistical associations rather than causal mechanisms. Policies that promote female secondary education alongside equitable access to skilled obstetric care may help improve infant survival in Somaliland.