Menu
Menu
Menu
Menu
Background: Zambia experienced a major decline in under-five mortality rates (U5MR), with one of the fastest declines in socio-economic disparities in sub-Saharan Africa in the last two decades. We aimed to understand the extent to which, and how, Zambia has reduced socio-economic inequalities in U5MR since 2000. Methods: Using nationally-representative data from Zambia Demographic Health Surveys (2001/2, 2007, 2013/14 and 2018), we examined trends and levels of inequalities in under-five mortality, intervention coverage, household water and sanitation, and fertility. This analysis was integrated with an in-depth review of key policy and program documents relevant to improving child survival in Zambia between 1990 and 2020. Results: The under-five mortality rate (U5MR) declined from 168 to 64 deaths per 1000 live births between 2001/2 and 2018 ZDHS rounds, particularly in the post-neonatal period. There were major reductions in U5MR inequalities between wealth, education and urban–rural residence groups. Yet reduced gaps between wealth groups in estimated absolute income or education levels did not simultaneously occur. Inequalities reduced markedly for coverage of reproductive, maternal, newborn and child health (RMNCH), malaria and human immunodeficiency virus interventions, but less so for water or sanitation and fertility levels. Several policy and health systems drivers were identified for reducing RMNCH inequalities: policy commitment to equity in RMNCH; financing with a focus on disadvantaged groups; multisectoral partnerships and horizontal programming; expansion of infrastructure and human resources for health; and involvement of community stakeholders and service providers. Conclusion: Zambia’s major progress in reducing inequalities in child survival between the poorest and richest people appeared to be notably driven by government policies and programs that centrally valued equity, despite ongoing gaps in absolute income and education levels. Future work should focus on sustaining these gains, while targeting families that have been left behind to achieve the sustainable development goal targets.
Zambia had a population of about 18 million in 2020. It reached lower-middle income country status in 2011 (which was however reversed to lower income in 2022). Income inequality in Zambia is one of the highest in sub-Saharan Africa (and the world) with a Gini index of 57 in 2015, with no improvement over the past two decades [18]. This mixed-methods study integrated quantitative inequality trend analysis with policy and health systems analysis. We used population-representative data from the last four Zambia Demographic and Health Surveys (ZDHS) (2001, 2007, 2013/2014, 2018) to analyze trends in inequalities for mortality, health intervention coverage and socio-economic conditions. The data collection methods for the ZDHS are described elsewhere [19]. Under-five mortality rates were calculated using the syncmrates program in Stata 15. We obtained estimates of the number of deaths among children aged 0–59 months out of 1000 live births, among all live births in the ten years preceding each round of the ZDHS. We also stratified the U5MR in each ZDHS round (2001–2018) by household wealth quintile using cross-tabulations. The wealth index was adopted to examine inequalities, based on DHS’ previously-computed principal component analysis of dwelling materials, access to utilities and household assets. The wealth index is created based on the assets for rural and urban places of residence separately, and divided into quintiles; the first quintile being classified as those within the lowest 20% of wealth index scores and the fifth quintile being those within the highest 20% of wealth index scores [20]. To assess the role of compositional changes in the socio-economic position of women in the poorest and richest wealth quintiles over time, we estimated absolute income levels by quintile for each survey. The calculation of absolute income for each percentile of distribution follows the Fink et al. (2017) definition and includes the Gini index, gross domestic product (in 2011 US dollars, power purchasing parity) and the household expenditure [21]. We then attributed a value in US dollars for the mean income of each wealth quintile (levels over time shown in Supplementary Fig. 3). Absolute education levels were also examined using the proportion of women with at least secondary education within each wealth quintile as another way to assess changes in socio-economic status among the least to most disadvantaged. The direct influence of income or education levels on health is complex, non-linear and multifactorial, and it was not within our aim or scope to uncover their direct causal influence in relation to health intervention coverage or child mortality [22, 23]. Rather, this approach to characterizing wealth groups with absolute socio-economic measures has been proposed previously as valuable to help understand whether there were improvements in a country’s socio-economic growth itself, or if not, whether improvements in health among poorer groups were rather due to intentional policies or programs that overcame the disadvantages of their lower socio-economic status or income [24, 25]. We examined inequality trends in RMNCH, malaria and HIV/AIDS intervention coverage, as well as changes in living conditions such as water and sanitation, and fertility rates between ZDHS 2001 (2007 for HIV/AIDS indicator, the first with disaggregated data) and 2018 by wealth quintile, given their known association with the main U5MR causes that reduced in Zambia in that period [26, 27]. We modified the well-established composite coverage index (CCI) to include malaria prevention as the fifth intervention area [27]. The CCI includes interventions across the continuum of care, where each stage is given equal weight as follows: where: To quantify and compare trends in inequalities over time, we calculated concentration indices (CIX) and slope indices of inequality (SII). CIX is calculated as twice the area between the curve and the line of equality, based on the plot of the cumulative percentage of the sample ranked by the socio-economic variable starting with worst off on x-axis and the cumulative percentage of the health variable on the y-axis. SII is the absolute difference between the predicted outcome value of the individuals with highest and lowest wealth scores, after regressing the mid-point of the cumulative proportion of the sample in each category (using a score from 0 to 1 from most to least disadvantaged) against the outcome estimate for each category [27–29]. The policy and health systems analysis involved in-depth document review of health policy reports, guidelines and strategy documents published and implemented between 1990 to date, obtained from the Zambian Ministry of Health, World Health Organization and United Nations agencies databases. We drew on quantitative health systems data from the WHO Global Health Expenditure Database [30], analysis of the Creditor Reporting System data with the Muskoka2 method [31], the WHO Global Health Database, Ministry of Health data and Zambia’s National Health Facility Census conducted in 2005 and 2017. To assess policies and strategies that may have contributed to reductions in under-five mortality since 2000, we adapted the Countdown to 2015 Policy and Programme Timeline Tool [32]. The Policy and Programme Timeline Tool is useful for identifying health policies, programmes and health systems changes that have been implemented in a country to improve RMNCH indicators and survival over time from 1990 to present. The tool extends across six levels including: national context, macro health systems and governance, health system building blocks, high impact policies specific to RMNCH, high impact research specific to RMNCH, and a cross-cutting component focused on partnerships and convening mechanisms [32, 33]. For this analysis, we focused on three levels that were most relevant to U5MR reduction and with available data or documents to track over time: macro-level governance and health systems environment, specific health system building blocks, and high impact policies specific to RMNCH. For each, we focused on where ‘equity’ was explicitly or implicitly incorporated as a guiding principle or ‘value’ [34, 35].
