EXCLUSIVE BREASTFEEDING AMONG WORKING MOTHERS IN GHANA: EVIDENCE FROM THE 2022 GHANA DEMOGRAPHIC AND HEALTH SURVEY

Abstract

ABSTRACT Background Exclusive breastfeeding may protect infants against common infections and support healthy growth and development. Working mothers may face constraints on exclusive breastfeeding arising from work schedules, separation from their infants, and inadequate breastfeeding support. National evidence on the individual, healthcare-related, and contextual factors associated with exclusive breastfeeding among working Ghanaian mothers appears to remain limited. Design Cross-sectional secondary analysis. Setting Nationally representative survey covering urban and rural communities across all 16 administrative regions of Ghana. Participants The analysis included 620 currently working mothers whose youngest living infants were aged 0–5 completed months and lived with them. The complete-case multivariable analysis included 619 mother–infant pairs. Primary outcome measure Current exclusive breastfeeding, defined using the standard 24-hour infant-feeding indicator. Infants were classified as exclusively breastfed when they received breast milk without water, formula, animal milk, other liquids, or solid or semi-solid foods during the preceding day or night. Oral rehydration solution, vitamins, minerals and prescribed medicines were permitted. Aim To estimate the prevalence of exclusive breastfeeding and examine its individual, healthcare-related and contextual correlates among working mothers of infants aged 0–5 months in Ghana. Methods Birth Recode data from the 2022 Ghana Demographic and Health Survey were analysed. Unweighted frequencies and survey-weighted percentages described the study population. Design-adjusted Wald tests assessed bivariate associations. Survey-weighted binary logistic regression estimated adjusted odds ratios (AORs) and 95% confidence intervals (CIs), accounting for sampling weights, primary sampling units, and strata. Results The survey-weighted prevalence of exclusive breastfeeding was 54.3% (95% CI: 49.2–59.3). Ethnicity, mode of delivery, region, and community poverty appeared to be statistically significant in the bivariate analyses. In the adjusted model, region was jointly associated with exclusive breastfeeding (p = 0.004). Mothers in the Northern (AOR = 4.93; 95% CI: 1.50–16.17) and Savannah (AOR = 4.22; 95% CI: 1.08–16.41) regions had higher odds than mothers in the Western Region. Mothers in low-education communities had lower odds than those in high-education communities (AOR = 0.54; 95% CI: 0.30–0.98). Although Guan mothers had higher odds than Akan mothers, the overall association with ethnicity was non-significant, and the estimate appeared imprecise. Maternal age, individual education, religion, parity, wealth, infant sex, antenatal care, postnatal care, and residence were not independently associated with exclusive breastfeeding. Conclusion The prevalence estimate suggests that slightly more than half of working mothers exclusively breastfed their infants. Regional and community differences appeared more pronounced than those associated with most measured individual characteristics. Regionally responsive breastfeeding support and practical community education may contribute to improved coverage. Workplace recommendations require further evidence because employment conditions were not measured directly. Strengths and limitations □ The study used nationally representative data covering all 16 regions of Ghana and applied the survey weights, clusters and strata, and may thereby improve the generalisability of the estimates to working mothers of infants aged 0–5 months. □ Exclusive breastfeeding was reconstructed using the complete standard 24-hour Demographic and Health Survey (DHS) feeding indicator, and mothers who had discontinued breastfeeding remained in the denominator, which may reduce the risk of overestimating prevalence. □ The cross-sectional design did not establish temporal order or causality, while the 24-hour feeding measure may not reflect feeding practices throughout the period from birth to the interview. □ Employment was measured only as current working status. Information on occupation, maternity leave, working hours, breastfeeding breaks, childcare arrangements and facilities for expressing and storing milk was unavailable. □ Small sample sizes in some ethnic and regional categories may have produced wide confidence intervals. Community poverty and education were derived contextual proxies and may not have directly measured community resources, breastfeeding knowledge, or service availability.