Abstract
BACKGROUND: Anemia remains a major global health crisis, affecting over 500 million women of reproductive age, with high burdens in resource-limited regions like Sub-Saharan Africa. Despite ongoing interventions such as iron supplementation programs, 49% of women of reproductive age in Zambia are anemic. Thus, the purpose of this study was to establish the national and subnational prevalence of anemia and identify its determinants among women of reproductive age in Zambia. METHODS: Data were drawn from the 2018 Zambia Demographic and Health Survey (ZDHS), a nationally representative survey employing a stratified two-stage cluster sampling design across 545 enumeration areas. A multilevel mixed-effects logistic regression model was used to identify individual- and community-level factors associated with anemia among women aged 15-49 years (n = 13,055). Four hierarchical models were constructed (null, individual-level, community-level, and full) to assess fixed and random effects, with model selection guided by Akaike Information Criterion (AIC) and Bayesian Information Criterion (BIC) criteria. Spatial analysis was conducted using Quantum Geographic Information System (QGIS), incorporating displaced GPS coordinates in accordance with Demographic Health Survey (DHS) protocols. All analyses applied sampling weights and assessed multicollinearity (Variance Inflation Factor -VIFs < 5). RESULTS: The national prevalence of anemia among women of reproductive age was 31% (95% Confidence Interval [CI]: 29-33%), with the highest rates observed in Western (38%) and Lusaka (36%) provinces, and the lowest in Central Province (24%). In adjusted analyses, pregnancy (Adjusted Odds Ratio [AOR] = 1.76; 95% CI: 1.52-2.03), Human Immunodeficiency Virus (HIV) positivity (AOR = 2.21; 95% CI: 1.97-2.49), and breastfeeding (AOR = 1.15; 95% CI: 1.02-1.30) were significantly associated with increased odds of anemia. Conversely, being married (AOR = 0.78; 95% CI: 0.68-0.90) and age 25-29 years (AOR = 0.84; 95% CI: 0.71-0.97) were protective. Spatial mapping identified Western Province as a high-burden hotspot. Community-level variance was notable (Intraclass Correlation Coefficient [ICC] = 6%, Median Odds Ratio [MOR] = 1.52), with 5% residual clustering persisting after adjusting for both individual and contextual factors, suggesting the influence of unmeasured ecological determinants. CONCLUSION: Anemia remains a significant public health issue among Zambian women of reproductive age, shaped by both individual- and community-level factors. These findings highlight the need for integrated, targeted interventions focusing on high-risk groups in high-prevalence areas. Strengthening clinical services and implementing community-based strategies to address healthcare access and environmental determinants are essential to reducing the burden of anemia in Zambia.