Background The health of children in female-headed households (FHHs) is commonly regarded as less optimal than the health of their counterparts in male-headed households (MHHs). Recent literature emphasizes the heterogeneity of FHHs and that health of children within these households may not necessarily be worse off than children within MHHs. This study aims to assess the relationship between household headship and stunting among children under five years old, across different levels of household deprivation.

Methods This cross-sectional study used Demographic Health Surveys (DHS) in 48 low- and middle-income countries to explore the relationship between stunting among children under-five years, household deprivation and household headship, identified as six categories: MHH, FHHs with husband; with adult women; with adult men (excluding husbands); with women and men; and with children only. Multilevel logistic models were employed and considered individual, household, sub-national regional and country level confounders. These models were stratified by deprivation levels, which is measured by the Multidimensional Poverty Index (MPI), and classified as ‘most deprived’ if in the top 25% of the MPI score and as ‘less deprived’ otherwise.

Results A third of children under-five were stunted. The proportions ranged from 29.9% (95% CI: 29.3%-30.4%) stunting among children in FHHs with men and women to 36.3% stunting among children in FHHs with children only. The proportion of stunting was 33.7% (95% CI: 35.7%-36.8%) among children in MHHs. The adjusted multilevel analysis, stratified by household deprivation levels, showed that estimates of stunting within the ‘most deprived’ were not significantly different between children in MHHs and in FHH types. Statistically significant differences were present in households categorized as ‘less deprived’, where FHHs with adult women and men and FHHs with children alone had slightly lower odds of having stunted children (0.94, 95% CI: 0.89–0.99; 0.88, 95%CI: 0.82–0.94, respectively) compared to MHHs, after controlling for confounders. Children in the rest of the FHH types did not have statistically significant differences in odds of being stunted.

Conclusion Using a more nuanced FHH typology, our analysis showed that stunting levels were not significantly higher among children within FHHs compared to children in MHHs. In some circumstances where household deprivation is not severe, health of children in FHHs may be better than MHHs. Child health interventions need to be better tailored to effectively target the most vulnerable households, which are not necessarily households headed by women. The health of children in deprived MHHs needs to be taken into consideration as well.