Gender Inequality and Health: Intersections Between Household Headship, Deprivation, and Women’s Empowerment Across Low- And Middle-Income Countries

Abstract

Background: Gender inequality within a country, community, and household plays a vital role in determining the wellbeing of women and children within the household. Generally, female-headed households (FHHs) are regarded as disadvantaged and over-represented among the poor. However, critics have argued that FHHs are heterogenous. The aim of this dissertation is three-fold: i. to propose an FHH typology based on nationally representative household surveys; ii- using the FHH typology, assess the relationship between household headship and multidimensional deprivation while exploring women’s empowerment as an effect modifier; and iii- examine the association between household headship and child health, using child stunting as the outcome, and considering whether household deprivation modifies this relationship.

Methods: This dissertation analysis uses secondary data from nationally representative cross-sectional surveys from low- and middle-income countries (LMICs). The analysis to develop the FHH typology (Chapter 2) was based on DHS and MICS surveys from 103 LMICs. Descriptive analyses were conducted at the household level to generate median proportions of the 16 FHH types (FHH16) defined and their household characteristics. A cluster analysis explored the FHH16 patterns across naturally grouped clusters of countries. For the second objective (Chapter 3), DHS surveys from 46 LMICs were used to assess the relationship between household headship and multidimensional deprivation. Headship was based on a recategorization of the FHH16 typology comprising FHH with husband; with adult women; with adult men (excluding husbands); with women and men; with children only; and women heads alone, plus male-headed households (MHHs). Households were classified as deprived or not deprived using a multidimensional deprivation indicator. The Survey-based Women’s Empowerment Index (SWPER Global) was used to measure empowerment. The multilevel logistic model used included household, sub-national regions, country-level confounders, and sub-national level SWPER as an effect modifier. Finally, in Chapter 4, DHS surveys from 43 LMICs were used to assess the influence of headship on child stunting. We used the same headship categories as in Chapter 3, excluding all households without under-five children, and thus the FHH type where women lived alone. Multilevel analyses were conducted for all children and restricted to children of the heads. Individual, household, sub-national, and country-level confounders were considered. These models were stratified by household deprivation.

Results: Based on the proposed FHH typology, the most common FHH16 types were households where the woman heads lived only with children, where women heads were alone, or lived with men, women, and children but without a husband. Exploratory cluster analysis generated five clusters of countries with similar FHH16 patterns. The clusters had distinct geographic, contextual, and economic characteristics. The results of Chapter 3 showed that the highest proportions of deprivation were found among FHHs with woman heads living alone (63%), followed by FHHs with children only (49%). Among MHHs 33% of households were deprived. Some FHHs had the lowest proportions of deprivation: FHHs with men (not the husband) and FHHs with men and women (25%). The adjusted multilevel model showed that FHHs with men and FHHs with men and women had significantly lower odds of deprivation, 0.61 (95% CI: 0.56-0.65) and 0.63 (95% CI: 0.57-0.70), respectively, than MHHs. FHHs with children alone and FHHs where the woman heads were alone had significantly higher odds of being deprived than MHHs: respectively 1.79 (95% CI: 1.63-1.96) and 3.31 (95% CI: 2.76-3.96). Women’s empowerment significantly modified the effect between multidimensional deprivation and headship. In areas of high sub-national level average empowerment, predicted proportions of deprivation were considerably lower and similar for most headship types compared to subnational areas with the lowest empowerment scores, where higher deprivation was predicted with larger disparities between the headship types. In Chapter 4, the results show that a third of children under-five were stunted. The highest proportion (36%) was recorded in FHHs with children only. Stunting in MHHs was 34%. When analyzing all children, no significant differences in stunting were found between FHH types and MHH. However, restricting the analysis to children of the heads, significant differences in stunting appeared among the non-deprived households. The FHHs with adult men (not the husband) and other adult women, and FHHs with other adult women were protective against stunting (ORs 0.68, 95%CI 0.56-0.83; 0.89, 95%CI 0.81-0.99, respectively) compared to MHHs, controlling for confounders.

Conclusion: Using this FHH typology, exploration of the patterns and differences showed that FHHs are heterogeneous. Furthermore, the ecological analysis emphasized further variation created by different societal and cultural factors. Using this FHH typology, the results highlight the presence of FHHs that are at lower risk of multidimensional deprivation and FHHs that are protective against child stunting in households that are not multidimensionally deprived, compared to MHHs. Moreover, women’s empowerment, which may be a proxy for contextual social norms, is essential in alleviating household deprivation, not only across FHH types but also among MHHs. More emphasis on the intersection between headship, deprivation, and women’s empowerment is needed to understand better how household and individual characteristics impact child health. Interventions need to be better tailored to effectively target the most vulnerable households, which are not necessarily only FHHs.