Effect of single-parent versus dual-parent households on dietary intake and growth among under-five children in Rwanda: an analysis using directed acyclic graph.

Abstract

Introduction: Poor childhood diet and malnutrition are global issues, particularly in low- and middle-income countries. Adequate nutrition during the early years of life is crucial for optimal child growth. Household dynamics influence children's well-being, with single-parent households facing unique challenges that can affect dietary intake and growth. In Rwanda, stunting and being underweight are still prevalent, and the role of household structure has not been studied. This study assessed the relationship of single parent in comparison to dual parent households with children's dietary intake and growth together with the mediating role of wealth in these relationships.

Methodology: This cross-sectional study used secondary data from the 2019/2020 Rwanda Demographics and Health Survey (DHS). The analysis included 3,531 children under five years of age from dual and single parent households at the time of the survey, with a specific focus on 2332 children aged 6-23 months for dietary assessment. To estimate the associations, we performed survey-weighted multivariate logistic regressions using binary outcomes, adjusting for confounders identified by a directed acyclic graph (DAG). Mediation analysis was performed using structural equation modeling (SEM) for continuous outcomes and generalized structural equation modeling (GSEM) for binary outcomes to explore the mediating effect of wealth on the relationships among household structure, the minimum acceptable diet (MAD), and child growth. Both models accounted for the complex sampling design.

Results: Among the 3,531 participants, 26.9% lived in single-parent households.​​​​​​ Stunting was observed in 34% of the children, while 8.3% were underweight, 1% experienced wasting, and only 21.1% of the children aged 6-23 months met MAD. Single-parent households were significantly associated with stunting (OR: 1.22; 95% CI: 1.01-1.46) and wasting (OR: 2.11; 95% CI: 1.11-3.99) but were not associated with underweight or MAD. Household structure indirectly influenced MAD (indirect effect (IE) = -0.007, p = 0.02), height-for-age (IE = -0.068, p < 0.001), and weight-for-age z scores (IE = -0.029, p = 0.001) through wealth, but no mediation effect was observed for height-for-weight z scores.

Conclusion: This study highlights the influence of family structure on children's dietary intake and growth in Rwanda. To effectively combat malnutrition in Rwanda and similar settings, targeted interventions should address household structure, particularly that of single-parent households, and wealth disparities.