Estimation of district-level under-5 mortality in Zambia using birth history data, 1980-2010

Going deeper with health equity measurement: how much more can surveys reveal about inequalities in health intervention coverage and mortality in Zambia?

BACKGROUND

Although Zambia has achieved notable improvements in reproductive, maternal, newborn and child health (RMNCH), continued efforts to address gaps are essential to reach the Sustainable Development Goals by 2030. Research to better uncover who is being most left behind with poor health outcomes is crucial. This study aimed to understand how much more demographic health surveys can reveal about Zambia's progress in reducing inequalities in under-five mortality rates and RMNCH intervention coverage.

METHODS

Using four nationally-representative Zambia Demographic Health Surveys (2001/2, 2007, 2013/14, 2018), we estimated under-five mortality rates (U5MR) and RMNCH composite coverage indices (CCI) comparing wealth quintiles, urban-rural residence and provinces. We further used multi-tier measures including wealth deciles and double disaggregation between wealth and region (urban residence, then provinces). These were summarised using slope indices of inequality, weighted mean differences from overall mean, Theil and concentration indices.

RESULTS

Inequalities in RMNCH coverage and under-five mortality narrowed between wealth groups, residence and provinces over time, but in different ways. Comparing measures of inequalities over time, disaggregation with multiple socio-economic and geographic stratifiers was often valuable and provided additional insights compared to conventional measures. Wealth quintiles were sufficient in revealing mortality inequalities compared to deciles, but comparing CCI by deciles provided more nuance by showing that the poorest 10% were left behind by 2018. Examining wealth in only urban areas helped reveal closing gaps in under-five mortality and CCI between the poorest and richest quintiles. Though challenged by lower precision, wealth gaps appeared to close in every province for both mortality and CCI. Still, inequalities remained higher in provinces with worse outcomes.

CONCLUSIONS

Multi-tier equity measures provided similarly plausible and precise estimates as conventional measures for most comparisons, except mortality among some wealth deciles, and wealth tertiles by province. This suggests that related research could readily use these multi-tier measures to gain deeper insights on inequality patterns for both health coverage and impact indicators, given sufficient samples. Future household survey analyses using fit-for-purpose equity measures are needed to uncover intersecting inequalities and target efforts towards effective coverage that will leave no woman or child behind in Zambia and beyond.

Accumulating Birth Histories Across Surveys for Improved Estimates of Child Mortality

Producing reliable estimates for childhood mortality rates is essential to monitor progress towards the United Nations Sustainable Development Goals (UN SDGs) and correctly evaluate policies designed to reduce childhood mortality rates. Different model-based approaches have been proposed to assess levels and trends in childhood mortality indicators. In this paper, we propose a design-based complement that accumulates birth histories across different household surveys to increase the precision of childhood mortality rates estimates. We accumulate birth histories across different cross-sectional Demographic Health Surveys/Multiple Cluster Indicator Surveys collected in Senegal and Malawi and estimate pooled childhood mortality rates based on calendar years. We show that accumulating birth histories smoothens fluctuations in time series for national and sub-national mortality rates, establishes more stable and reliable time trends, and results in estimated standard errors of the cumulated rates that are about 50–60% lower than their counterparts from separate surveys.

Modelling chronic malnutrition in Zambia: A Bayesian distributional regression approach

Background

The burden of child under-nutrition still remains a global challenge, with greater severity being faced by low- and middle-income countries, despite the strategies in the Sustainable Development Goals (SDGs). Globally, malnutrition is the one of the most important risk factors associated with illness and death, affecting hundreds of millions of pregnant women and young children. Sub-Saharan Africa is one of the regions in the world struggling with the burden of chronic malnutrition. The 2018 Zambia Demographic and Health Survey (ZDHS) report estimated that 35% of the children under five years of age are stunted. The objective of this study was to analyse the distribution, and associated factors of stunting in Zambia.

Methods

We analysed the relationships between socio-economic, and remote sensed characteristics and anthropometric outcomes in under five children, using Bayesian distributional regression. Georeferenced data was available for 25,852 children from two waves of the ZDHS, 31% observation were from the 2007 and 69% were from the 2013/14. We assessed the linear, non-linear and spatial effects of covariates on the height-for-age z-score.

Results

Stunting decreased between 2007 and 2013/14 from a mean z-score of 1.59 (credible interval (CI): -1.63; -1.55) to -1.47 (CI: -1.49; -1.44). We found a strong non-linear relationship for the education of the mother and the wealth of the household on the height-for-age z-score. Moreover, increasing levels of maternal education above the eighth grade were associated with a reduced variation of stunting. Our study finds that remote sensed covariates alone explain little of the variation of the height-for-age z-score, which highlights the importance to collect socio-economic characteristics, and to control for socio-economic characteristics of the individual and the household.

Conclusions

While stunting still remains unacceptably high in Zambia with remarkable regional inequalities, the decline is lagging behind goal two of the SDGs. This emphasises the need for policies that help to reduce the share of chronic malnourished children within Zambia.

Space-time smoothing of mortality estimates in children aged 5-14 in Sub-Saharan Africa

To meet the SDG requirement of spatial disaggregation of indicators, several methods have been developed to generate estimates of under-five mortality at the sub-national level. The reliability of sub-national mortality estimates in children aged 5-14 with the available survey data has not been evaluated so far. We generate Admin-1 sub-national estimates of the risk of dying in children aged less than five (5q0) and those aged 5 to 14 years old (10q5). We use 96 Demographic and Health Surveys (DHS) in 20 Sub-Saharan countries having at least 3 surveys designed to be representative at a sub-national level. The estimates account for the complex sample design of DHS and HIV-related biases in young children. A Bayesian space-time model previously developed for under-five mortality is used to smooth estimates across space and time in both age groups to reduce problems associated with data sparsity. The posterior distributions of the probability 10q5 are used to compute coefficients of variation and assess precision. Sufficiently precise estimates are retained to study the sub-national relationship between age-specific mortality rates (5q0 and 10q5), accounting for uncertainty in sub-national levels. Out of 1,132 space-time estimates, 62.3% are considered sufficiently precise with high heterogeneity across countries. Across all periods, sub-national estimates of mortality in children aged 0-4 are highly correlated with those in older children and young adolescents but this correlation is largely driven by the mortality decline. Within specific periods of time, it is often impossible to assess the relationship between mortality rates in the two age groups at the sub-national level, except in Nigeria, Ethiopia, Cameroon, Senegal and Zambia. As increased attention is devoted to survival after age 5, more research is needed to ensure that sub-national areas with specific interventions required for older children can be correctly identified.

Geospatial estimation of reproductive, maternal, newborn and child health indicators: a systematic review of methodological aspects of studies based on household surveys

BACKGROUND

Geospatial approaches are increasingly used to produce fine spatial scale estimates of reproductive, maternal, newborn and child health (RMNCH) indicators in low- and middle-income countries (LMICs). This study aims to describe important methodological aspects and specificities of geospatial approaches applied to RMNCH coverage and impact outcomes and enable non-specialist readers to critically evaluate and interpret these studies.

METHODS

Two independent searches were carried out using Medline, Web of Science, Scopus, SCIELO and LILACS electronic databases. Studies based on survey data using geospatial approaches on RMNCH in LMICs were considered eligible. Studies whose outcomes were not measures of occurrence were excluded.

RESULTS

We identified 82 studies focused on over 30 different RMNCH outcomes. Bayesian hierarchical models were the predominant modeling approach found in 62 studies. 5 × 5 km estimates were the most common resolution and the main source of information was Demographic and Health Surveys. Model validation was under reported, with the out-of-sample method being reported in only 56% of the studies and 13% of the studies did not present a single validation metric. Uncertainty assessment and reporting lacked standardization, and more than a quarter of the studies failed to report any uncertainty measure.

CONCLUSIONS

The field of geospatial estimation focused on RMNCH outcomes is clearly expanding. However, despite the adoption of a standardized conceptual modeling framework for generating finer spatial scale estimates, methodological aspects such as model validation and uncertainty demand further attention as they are both essential in assisting the reader to evaluate the estimates that are being presented.

Estimation of under-5 child mortality rates in 52 low-migration countries

Reduction of the under-5 mortality rate is a target of the Sustainable Development Goals. Therefore, this study aimed to estimate under-5 child mortality rates in 52 low-migration countries using population data. The study utilized population data from the US Census Bureau from 1990 to 2015. The method involved first estimating mortality rates for countries with negligible net migration and then applying these rates to countries with matching mortality profiles, where it is reasonable to assume that migration is negligible for children under the age of 5 years. The highest child mortality was concentrated in the African region, followed by Asia and the Western region. However, steady progress in child mortality trends was concentrated in low-income countries. This simple method demonstrated that child mortality has significantly improved in high-income countries, followed by middle- and low-income countries. To reduce the under-5 mortality rates even further in these 52 countries, there is a need to accelerate equitable plans and policies related to child health to promote children's longevity and survival.

Small Area Estimation for Disease Prevalence Mapping

Small area estimation (SAE) entails estimating characteristics of interest for domains, often geographical areas, in which there may be few or no samples available. SAE has a long history and a wide variety of methods have been suggested, from a bewildering range of philosophical standpoints. We describe design-based and model-based approaches and models that are specified at the area-level and at the unit-level, focusing on health applications and fully Bayesian spatial models. The use of auxiliary information is a key ingredient for successful inference when response data are sparse and we discuss a number of approaches that allow the inclusion of covariate data. SAE for HIV prevalence, using data collected from a Demographic Health Survey in Malawi in 2015-2016, is used to illustrate a number of techniques. The potential use of SAE techniques for outcomes related to COVID-19 is discussed.

Mapping 123 million neonatal, infant and child deaths between 2000 and 2017

Since 2000, many countries have achieved considerable success in improving child survival, but localized progress remains unclear. To inform efforts towards United Nations Sustainable Development Goal 3.2-to end preventable child deaths by 2030-we need consistently estimated data at the subnational level regarding child mortality rates and trends. Here we quantified, for the period 2000-2017, the subnational variation in mortality rates and number of deaths of neonates, infants and children under 5 years of age within 99 low- and middle-income countries using a geostatistical survival model. We estimated that 32% of children under 5 in these countries lived in districts that had attained rates of 25 or fewer child deaths per 1,000 live births by 2017, and that 58% of child deaths between 2000 and 2017 in these countries could have been averted in the absence of geographical inequality. This study enables the identification of high-mortality clusters, patterns of progress and geographical inequalities to inform appropriate investments and implementations that will help to improve the health of all populations.

Estimating under-five mortality in space and time in a developing world context

Accurate estimates of the under-five mortality rate in a developing world context are a key barometer of the health of a nation. This paper describes a new model to analyze survey data on mortality in this context. We are interested in both spatial and temporal description, that is wishing to estimate under-five mortality rate across regions and years and to investigate the association between the under-five mortality rate and spatially varying covariate surfaces. We illustrate the methodology by producing yearly estimates for subnational areas in Kenya over the period 1980-2014 using data from the Demographic and Health Surveys, which use stratified cluster sampling. We use a binomial likelihood with fixed effects for the urban/rural strata and random effects for the clustering to account for the complex survey design. Smoothing is carried out using Bayesian hierarchical models with continuous spatial and temporally discrete components. A key component of the model is an offset to adjust for bias due to the effects of HIV epidemics. Substantively, there has been a sharp decline in Kenya in the under-five mortality rate in the period 1980-2014, but large variability in estimated subnational rates remains. A priority for future research is understanding this variability. In exploratory work, we examine whether a variety of spatial covariate surfaces can explain the variability in under-five mortality rate. Temperature, precipitation, a measure of malaria infection prevalence, and a measure of nearness to cities were candidates for inclusion in the covariate model, but the interplay between space, time, and covariates is complex.

Assessing the Contribution of Malaria Vector Control and Other Maternal and Child Health Interventions in Reducing All-Cause Under-Five Mortality in Zambia, 1990-2010

Under-five mortality in Zambia has declined since 1990, with reductions accelerating after 2000. Zambia’s scale-up of malaria control is viewed as the driver of these gains, but past studies have not fully accounted for other potential factors. This study sought to systematically evaluate the impact of malaria vector control on under-five mortality. Using a mixed-effects regression model, we quantified the relationship between malaria vector control, other priority health interventions, and socioeconomic indicators and district-level under-five mortality trends from 1990 to 2010. We then conducted counterfactual analyses to estimate under-five mortality in the absence of scaling up malaria vector control. Throughout Zambia, increased malaria vector control coverage coincided with scaling up three other interventions: the pentavalent vaccine, exclusive breast-feeding, and prevention of mother-to-child transmission of HIV services. This simultaneous scale-up made statistically isolating intervention-specific impact infeasible. Instead, in combination, these interventions jointly accelerated declines in under-five mortality by 11% between 2000 and 2010. Zambia’s scale-up of multiple interventions is notable, yet our findings highlight challenges in quantifying program-specific impact without better health data and information systems. As countries aim to further improve health outcomes, there is even greater need—and opportunity—to strengthen routine data systems and to develop more rigorous evaluation strategies.

Sub national variation and inequalities in under-five mortality in Kenya since 1965

Background

Despite significant declines in under five mortality (U5M) over the last 3 decades, Kenya did not achieve Millennium Development Goal 4 (MDG 4) by 2015. To better understand trends and inequalities in child mortality, analysis of U5M variation at subnational decision making units is required. Here the comprehensive compilation and analysis of birth history data was used to understand spatio-temporal variation, inequalities and progress towards achieving the reductions targets of U5M between 1965 and 2013 and projected to 2015 at decentralized health planning units (counties) in Kenya.

Methods

Ten household surveys and three censuses with data on birth histories undertaken between 1989 and 2014 were assembled. The birth histories were allocated to the respective counties and demographic methods applied to estimate U5M per county by survey. To generate a single U5M estimate for year and county, a Bayesian spatio-temporal Gaussian process regression was fitted accounting for variation in sample size, surveys and demographic methods. Inequalities and the progress in meeting the goals set to reduce U5M were evaluated subnationally.

Results

Nationally, U5M reduced by 61·6%, from 141·7 (121·6–164·0) in 1965 to 54·5 (44·6–65·5) in 2013. The declining U5M was uneven ranging between 19 and 80% across the counties with some years when rates increased. By 2000, 25 counties had achieved the World Summit for Children goals. However, as of 2015, no county had achieved MDG 4. There was a striking decline in the levels of inequality between counties over time, however, disparities persist. By 2013 there persists a 3·8 times difference between predicted U5M rates when comparing counties with the highest U5M rates against those with the lowest U5M rates.

Conclusion

Kenya has made huge progress in child survival since independence. However, U5M remains high and heterogeneous with substantial differences between counties. Better use of the current resources through focused allocation is required to achieve further reductions, reduce inequalities and increase the likelihood of achieving Sustainable Development Goal 3·2 on U5M by 2030.

Electronic supplementary material

The online version of this article (10.1186/s12889-019-6474-1) contains supplementary material, which is available to authorized users.

Changes in the spatial distribution of the under-five mortality rate: Small-area analysis of 122 DHS surveys in 262 subregions of 35 countries in Africa

The under-five mortality rate (U5MR) is a critical and widely available population health indicator. Both the MDGs and SDGs define targets for improvement in the U5MR, and the SDGs require spatial disaggregation of indicators. We estimate trends in the U5MR for Admin-1 subnational areas using 122 DHS surveys in 35 countries in Africa and assess progress toward the MDG target reductions for each subnational region and each country as a whole. In each country, direct weighted estimates of the U5MR from each survey are calculated and combined into a single estimate for each Admin-1 region across five-year periods. Our method fully accounts for the sample design of each survey. The region-time-specific estimates are smoothed using a Bayesian, space-time model that produces more precise estimates (when compared to the direct estimates) at a one-year scale that are consistent with each other in both space and time. The resulting estimated distributions of the U5MR are summarized and used to assess subnational progress toward the MDG 4 target of two-thirds reduction in the U5MR during 1990–2015. Our space-time modeling approach is tractable and can be readily applied to a large collection of sample survey data. Subnational, regional spatial heterogeneity in the levels and trends in the U5MR vary considerably across Africa. There is no generalizable pattern between spatial heterogeneity and levels or trends in the U5MR. Subnational, small-area estimates of the U5MR: (i) identify subnational regions where interventions are still necessary and those where improvement is well under way; and (ii) countries where there is very little spatial variation and others where there are important differences between subregions in both levels and trends. More work is necessary to improve both the data sources and methods necessary to adequately measure subnational progress toward the SDG child survival targets.

Small area estimation of under-5 mortality in Bangladesh, Cameroon, Chad, Mozambique, Uganda, and Zambia using spatially misaligned data

Background

The under-5 mortality rate (U5MR) is an important metric of child health and survival. Country-level estimates of U5MR are readily available, but efforts to estimate U5MR subnationally have been limited, in part, due to spatial misalignment of available data sources (e.g., use of different administrative levels, or as a result of historical boundary changes).

Methods

We analyzed all available complete and summary birth history data in surveys and censuses in six countries (Bangladesh, Cameroon, Chad, Mozambique, Uganda, and Zambia) at the finest geographic level available in each data source. We then developed small area estimation models capable of incorporating spatially misaligned data. These small area estimation models were applied to the birth history data in order to estimate trends in U5MR from 1980 to 2015 at the second administrative level in Cameroon, Chad, Mozambique, Uganda, and Zambia and at the third administrative level in Bangladesh.

Results

We found substantial variation in U5MR in all six countries: there was more than a two-fold difference in U5MR between the area with the highest rate and the area with the lowest rate in every country. All areas in all countries experienced declines in U5MR between 1980 and 2015, but the degree varied both within and between countries. In Cameroon, Chad, Mozambique, and Zambia we found areas with U5MRs in 2015 that were higher than in other parts of the same country in 1980. Comparing subnational U5MR to country-level targets for the Millennium Development Goals (MDG), we find that 12.8% of areas in Bangladesh did not meet the country-level target, although the country as whole did. A minority of areas in Chad, Mozambique, Uganda, and Zambia met the country-level MDG targets while these countries as a whole did not.

Conclusions

Subnational estimates of U5MR reveal significant within-country variation. These estimates could be used for identifying high-need areas and positive deviants, tracking trends in geographic inequalities, and evaluating progress towards international development targets such as the Sustainable Development Goals.

Electronic supplementary material

The online version of this article (10.1186/s12963-018-0171-7) contains supplementary material, which is available to authorized users.

Modelling spatio-temporal patterns of disease for spatially misaligned data: An application on measles incidence data in Namibia from 2005-2014

Background

Making inferences about measles distribution patterns at small area level is vital for more focal targeted intervention. However, in statistical literature, the analysis of originally collected data on one resolution with the purpose to make inferences on a different level of spatial resolution is referred to as the misalignment problem. In Namibia the measles data were available in aggregated format at regional level for the period 2005 to 2014. This leads to a spatial misalignment problem if the purpose is to make decisions at constituency level. Moreover, although data on risk covariates of measles could be obtained at constituency level, they were not available each year between 2005 and 2014. Thus, assuming that covariates were constant through the study period would induce measurement errors which might have effects on the analysis results. This paper presents a spatio-temporal model through a multi-step approach in order to deal with misalignment and measurement error.

Methods

For the period 2005–2014, measles data from the Ministry of Health and Social Services (MoHSS) were analysed in two steps. First, a multi-step approach was applied to correct spatial misalignment in the data. Second, a classical measurement error model was fitted to account for measurement errors. The time effects were specified using a nonparametric formulation for the linear trend through first order random walk. An interaction between area and time was modelled through type I and type II interaction structures.

Results

The study showed that there was high variation in measles risk across constituencies and as well as over the study period (2005–2014). Furthermore, the risk of measles was found to be associated with (i) the number of people aged between 0 and 24 years, (ii) the percentages of women aged 15–49 with an educational level more than secondary, (iii) the percentages of children age 12–23 months who received measles vaccine, (iv) the percentages of malnourished children under 5 years, and (vi) the measles cases for each previous year.

Conclusion

The study showed some of the determinants of measles risk and revealed areas at high risk through disease mapping. Additionally, the study showed a non-linear temporal distribution of measles risk over the study period. Finally, it was shown that ignoring the measurement errors may yield misleading results. It was recommended that group and geographically targeted intervention, prevention and control strategies can be tailored on the basis these findings.

Mapping under-5 and neonatal mortality in Africa, 2000-15: a baseline analysis for the Sustainable Development Goals

BACKGROUND

During the Millennium Development Goal (MDG) era, many countries in Africa achieved marked reductions in under-5 and neonatal mortality. Yet the pace of progress toward these goals substantially varied at the national level, demonstrating an essential need for tracking even more local trends in child mortality. With the adoption of the Sustainable Development Goals (SDGs) in 2015, which established ambitious targets for improving child survival by 2030, optimal intervention planning and targeting will require understanding of trends and rates of progress at a higher spatial resolution. In this study, we aimed to generate high-resolution estimates of under-5 and neonatal all-cause mortality across 46 countries in Africa.

METHODS

We assembled 235 geographically resolved household survey and census data sources on child deaths to produce estimates of under-5 and neonatal mortality at a resolution of 5 × 5 km grid cells across 46 African countries for 2000, 2005, 2010, and 2015. We used a Bayesian geostatistical analytical framework to generate these estimates, and implemented predictive validity tests. In addition to reporting 5 × 5 km estimates, we also aggregated results obtained from these estimates into three different levels-national, and subnational administrative levels 1 and 2-to provide the full range of geospatial resolution that local, national, and global decision makers might require.

FINDINGS

Amid improving child survival in Africa, there was substantial heterogeneity in absolute levels of under-5 and neonatal mortality in 2015, as well as the annualised rates of decline achieved from 2000 to 2015. Subnational areas in countries such as Botswana, Rwanda, and Ethiopia recorded some of the largest decreases in child mortality rates since 2000, positioning them well to achieve SDG targets by 2030 or earlier. Yet these places were the exception for Africa, since many areas, particularly in central and western Africa, must reduce under-5 mortality rates by at least 8·8% per year, between 2015 and 2030, to achieve the SDG 3.2 target for under-5 mortality by 2030.

INTERPRETATION

In the absence of unprecedented political commitment, financial support, and medical advances, the viability of SDG 3.2 achievement in Africa is precarious at best. By producing under-5 and neonatal mortality rates at multiple levels of geospatial resolution over time, this study provides key information for decision makers to target interventions at populations in the greatest need. In an era when precision public health increasingly has the potential to transform the design, implementation, and impact of health programmes, our 5 × 5 km estimates of child mortality in Africa provide a baseline against which local, national, and global stakeholders can map the pathways for ending preventable child deaths by 2030.

FUNDING

Bill & Melinda Gates Foundation.

Global, regional, national, and selected subnational levels of stillbirths, neonatal, infant, and under-5 mortality, 1980-2015: a systematic analysis for the Global Burden of Disease Study 2015

BACKGROUND

Established in 2000, Millennium Development Goal 4 (MDG4) catalysed extraordinary political, financial, and social commitments to reduce under-5 mortality by two-thirds between 1990 and 2015. At the country level, the pace of progress in improving child survival has varied markedly, highlighting a crucial need to further examine potential drivers of accelerated or slowed decreases in child mortality. The Global Burden of Disease 2015 Study (GBD 2015) provides an analytical framework to comprehensively assess these trends for under-5 mortality, age-specific and cause-specific mortality among children under 5 years, and stillbirths by geography over time.

METHODS

Drawing from analytical approaches developed and refined in previous iterations of the GBD study, we generated updated estimates of child mortality by age group (neonatal, post-neonatal, ages 1-4 years, and under 5) for 195 countries and territories and selected subnational geographies, from 1980-2015. We also estimated numbers and rates of stillbirths for these geographies and years. Gaussian process regression with data source adjustments for sampling and non-sampling bias was applied to synthesise input data for under-5 mortality for each geography. Age-specific mortality estimates were generated through a two-stage age-sex splitting process, and stillbirth estimates were produced with a mixed-effects model, which accounted for variable stillbirth definitions and data source-specific biases. For GBD 2015, we did a series of novel analyses to systematically quantify the drivers of trends in child mortality across geographies. First, we assessed observed and expected levels and annualised rates of decrease for under-5 mortality and stillbirths as they related to the Soci-demographic Index (SDI). Second, we examined the ratio of recorded and expected levels of child mortality, on the basis of SDI, across geographies, as well as differences in recorded and expected annualised rates of change for under-5 mortality. Third, we analysed levels and cause compositions of under-5 mortality, across time and geographies, as they related to rising SDI. Finally, we decomposed the changes in under-5 mortality to changes in SDI at the global level, as well as changes in leading causes of under-5 deaths for countries and territories. We documented each step of the GBD 2015 child mortality estimation process, as well as data sources, in accordance with the Guidelines for Accurate and Transparent Health Estimates Reporting (GATHER).

FINDINGS

Globally, 5·8 million (95% uncertainty interval [UI] 5·7-6·0) children younger than 5 years died in 2015, representing a 52·0% (95% UI 50·7-53·3) decrease in the number of under-5 deaths since 1990. Neonatal deaths and stillbirths fell at a slower pace since 1990, decreasing by 42·4% (41·3-43·6) to 2·6 million (2·6-2·7) neonatal deaths and 47·0% (35·1-57·0) to 2·1 million (1·8-2·5) stillbirths in 2015. Between 1990 and 2015, global under-5 mortality decreased at an annualised rate of decrease of 3·0% (2·6-3·3), falling short of the 4·4% annualised rate of decrease required to achieve MDG4. During this time, 58 countries met or exceeded the pace of progress required to meet MDG4. Between 2000, the year MDG4 was formally enacted, and 2015, 28 additional countries that did not achieve the 4·4% rate of decrease from 1990 met the MDG4 pace of decrease. However, absolute levels of under-5 mortality remained high in many countries, with 11 countries still recording rates exceeding 100 per 1000 livebirths in 2015. Marked decreases in under-5 deaths due to a number of communicable diseases, including lower respiratory infections, diarrhoeal diseases, measles, and malaria, accounted for much of the progress in lowering overall under-5 mortality in low-income countries. Compared with gains achieved for infectious diseases and nutritional deficiencies, the persisting toll of neonatal conditions and congenital anomalies on child survival became evident, especially in low-income and low-middle-income countries. We found sizeable heterogeneities in comparing observed and expected rates of under-5 mortality, as well as differences in observed and expected rates of change for under-5 mortality. At the global level, we recorded a divergence in observed and expected levels of under-5 mortality starting in 2000, with the observed trend falling much faster than what was expected based on SDI through 2015. Between 2000 and 2015, the world recorded 10·3 million fewer under-5 deaths than expected on the basis of improving SDI alone.

INTERPRETATION

Gains in child survival have been large, widespread, and in many places in the world, faster than what was anticipated based on improving levels of development. Yet some countries, particularly in sub-Saharan Africa, still had high rates of under-5 mortality in 2015. Unless these countries are able to accelerate reductions in child deaths at an extraordinary pace, their achievement of proposed SDG targets is unlikely. Improving the evidence base on drivers that might hasten the pace of progress for child survival, ranging from cost-effective intervention packages to innovative financing mechanisms, is vital to charting the pathways for ultimately ending preventable child deaths by 2030.

FUNDING

Bill & Melinda Gates Foundation.

Geographical Inequalities and Social and Environmental Risk Factors for Under-Five Mortality in Ghana in 2000 and 2010: Bayesian Spatial Analysis of Census Data

BACKGROUND

Under-five mortality is declining in Ghana and many other countries. Very few studies have measured under-five mortality-and its social and environmental risk factors-at fine spatial resolutions, which is relevant for policy purposes. Our aim was to estimate under-five mortality and its social and environmental risk factors at the district level in Ghana.

METHODS AND FINDINGS

We used 10% random samples of Ghana's 2000 and 2010 National Population and Housing Censuses. We applied indirect demographic methods and a Bayesian spatial model to the information on total number of children ever born and children surviving to estimate under-five mortality (probability of dying by 5 y of age, 5q0) for each of Ghana's 110 districts. We also used the census data to estimate the distributions of households or persons in each district in terms of fuel used for cooking, sanitation facility, drinking water source, and parental education. Median district 5q0 declined from 99 deaths per 1,000 live births in 2000 to 70 in 2010. The decline ranged from <5% in some northern districts, where 5q0 had been higher in 2000, to >40% in southern districts, where it had been lower in 2000, exacerbating existing inequalities. Primary education increased in men and women, and more households had access to improved water and sanitation and cleaner cooking fuels. Higher use of liquefied petroleum gas for cooking was associated with lower 5q0 in multivariate analysis.

CONCLUSIONS

Under-five mortality has declined in all of Ghana's districts, but the cross-district inequality in mortality has increased. There is a need for additional data, including on healthcare, and additional environmental and socioeconomic measurements, to understand the reasons for the variations in mortality levels and trends.

Geographical and socioeconomic inequalities in women and children's nutritional status in Pakistan in 2011: an analysis of data from a nationally representative survey

BACKGROUND

Pakistan has one of the highest levels of child and maternal undernutrition worldwide, but little information about geographical and socioeconomic inequalities is available. We aimed to analyse anthropometric indicators for childhood and maternal nutrition at a district level in Pakistan and assess the association of nutritional status with food security and maternal and household socioeconomic factors.

METHODS

We used data from the 2011 Pakistan National Nutrition Survey, which included anthropometric measurements for 33 638 children younger than 5 years and 24 826 women of childbearing age. We estimated the prevalences of stunting, wasting, and underweight among children and of underweight, overweight, and obesity in women for all 143 districts of Pakistan using a Bayesian spatial technique. We used a mixed-effect linear model to analyse the association of nutritional status with individual and household sociodemographic factors and food security.

FINDINGS

Stunting prevalence in Pakistan's districts ranged between 22% (95% credible interval 19-26) and 76% (69-83); the lowest figures for wasting and underweight were both less than 2·5% and the highest were 42% (34-50) for wasting and 54% (49-59) for underweight. In 106 districts, more women were overweight than were underweight; in 49 of these districts more women were obese than were underweight. Children were better nourished if their mothers were taller or had higher weight, if they lived in wealthier households, and if their mothers had 10 or more years of education. Severe food insecurity was associated with worse nutritional outcomes for both children and women.

INTERPRETATION

We noted large social and geographical inequalities in child and maternal nutrition in Pakistan, masked by national and provincial averages. Pakistan is also beginning to face the concurrent challenge of high burden of childhood undernutrition and overweight and obesity among women of reproductive age. Planning, implementation, and evaluation of programmes for food and nutrition should be based on district-level needs and outcomes.

FUNDING

Bill & Melinda Gates Foundation, Grand Challenges Canada, UK Medical Research Council.

Subnational benchmarking of health systems performance in Africa using health outcome and coverage indicators

National health systems performance (HSP) assessments and benchmarking are critical to understanding how well the delivery of healthcare meets the needs of citizens. Benchmarking HSP has often been done between countries to inform the global public health space. However, its impact is likely to be far greater when implemented sub-nationally to inform actual decisions on resource allocations and performance improvements, especially in high disease burden, low-income countries, where the resource envelope available for health is inadequate. In their study, Roberts and colleagues assemble, analyse and map a minimum set of health intervention and outcome indicators from 1990-2011 to assess and benchmark HSP across the 11 regions of Uganda. This is the first empirical sub-national HSP benchmarking study in the country and the results have potentially important health system policy implications. Please see related research: http://www.biomedcentral.com/1741-7015/13/285.

Benchmarking health system performance across regions in Uganda: a systematic analysis of levels and trends in key maternal and child health interventions, 1990-2011

BACKGROUND

Globally, countries are increasingly prioritizing the reduction of health inequalities and provision of universal health coverage. While national benchmarking has become more common, such work at subnational levels is rare. The timely and rigorous measurement of local levels and trends in key health interventions and outcomes is vital to identifying areas of progress and detecting early signs of stalled or declining health system performance. Previous studies have yet to provide a comprehensive assessment of Uganda's maternal and child health (MCH) landscape at the subnational level.

METHODS

By triangulating a number of different data sources - population censuses, household surveys, and administrative data - we generated regional estimates of 27 key MCH outcomes, interventions, and socioeconomic indicators from 1990 to 2011. After calculating source-specific estimates of intervention coverage, we used a two-step statistical model involving a mixed-effects linear model as an input to Gaussian process regression to produce regional-level trends. We also generated national-level estimates and constructed an indicator of overall intervention coverage based on the average of 11 high-priority interventions.

RESULTS

National estimates often veiled large differences in coverage levels and trends across Uganda's regions. Under-5 mortality declined dramatically, from 163 deaths per 1,000 live births in 1990 to 85 deaths per 1,000 live births in 2011, but a large gap between Kampala and the rest of the country persisted. Uganda rapidly scaled up a subset of interventions across regions, including household ownership of insecticide-treated nets, receipt of artemisinin-based combination therapies among children under 5, and pentavalent immunization. Conversely, most regions saw minimal increases, if not actual declines, in the coverage of indicators that required multiple contacts with the health system, such as four or more antenatal care visits, three doses of oral polio vaccine, and two doses of intermittent preventive therapy during pregnancy. Some of the regions with the lowest levels of overall intervention coverage in 1990, such as North and West Nile, saw marked progress by 2011; nonetheless, sizeable disparities remained between Kampala and the rest of the country. Countrywide, overall coverage increased from 40% in 1990 to 64% in 2011, but coverage in 2011 ranged from 57% to 70% across regions.

CONCLUSIONS

The MCH landscape in Uganda has, for the most part, improved between 1990 and 2011. Subnational benchmarking quantified the persistence of geographic health inequalities and identified regions in need of additional health systems strengthening. The tracking and analysis of subnational health trends should be conducted regularly to better guide policy decisions and strengthen responsiveness to local health needs.

Space-Time Smoothing of Complex Survey Data: Small Area Estimation for Child Mortality

Many people living in low and middle-income countries are not covered by civil registration and vital statistics systems. Consequently, a wide variety of other types of data including many household sample surveys are used to estimate health and population indicators. In this paper we combine data from sample surveys and demographic surveillance systems to produce small area estimates of child mortality through time. Small area estimates are necessary to understand geographical heterogeneity in health indicators when full-coverage vital statistics are not available. For this endeavor spatio-temporal smoothing is beneficial to alleviate problems of data sparsity. The use of conventional hierarchical models requires careful thought since the survey weights may need to be considered to alleviate bias due to non-random sampling and non-response. The application that motivated this work is estimation of child mortality rates in five-year time intervals in regions of Tanzania. Data come from Demographic and Health Surveys conducted over the period 1991-2010 and two demographic surveillance system sites. We derive a variance estimator of under five years child mortality that accounts for the complex survey weighting. For our application, the hierarchical models we consider include random effects for area, time and survey and we compare models using a variety of measures including the conditional predictive ordinate (CPO). The method we propose is implemented via the fast and accurate integrated nested Laplace approximation (INLA).

Benchmarking health system performance across states in Nigeria: a systematic analysis of levels and trends in key maternal and child health interventions and outcomes, 2000-2013

BACKGROUND

Nigeria has made notable gains in improving childhood survival but the country still accounts for a large portion of the world's overall disease burden, particularly among women and children. To date, no systematic analyses have comprehensively assessed trends for health outcomes and interventions across states in Nigeria.

METHODS

We extracted data from 19 surveys to generate estimates for 20 key maternal and child health (MCH) interventions and outcomes for 36 states and the Federal Capital Territory from 2000 to 2013. Source-specific estimates were generated for each indicator, after which a two-step statistical model was applied using a mixed-effects model followed by Gaussian process regression to produce state-level trends. National estimates were calculated by population-weighting state values.

RESULTS

Under-5 mortality decreased in all states from 2000 to 2013, but a large gap remained across them. Malaria intervention coverage stayed low despite increases between 2009 and 2013, largely driven by rising rates of insecticide-treated net ownership. Overall, vaccination coverage improved, with notable increases in the coverage of three-dose oral polio vaccine. Nevertheless, immunization coverage remained low for most vaccines, including measles. Coverage of other MCH interventions, such as antenatal care and skilled birth attendance, generally stagnated and even declined in many states, and the range between the lowest- and highest-performing states remained wide in 2013. Countrywide, a measure of overall intervention coverage increased from 33% in 2000 to 47% in 2013 with considerable variation across states, ranging from 21% in Sokoto to 66% in Ekiti.

CONCLUSIONS

We found that Nigeria made notable gains for a subset of MCH indicators between 2000 and 2013, but also experienced stalled progress and even declines for others. Despite progress for a subset of indicators, Nigeria's absolute levels of intervention coverage remained quite low. As Nigeria rolls out its National Health Bill and seeks to strengthen its delivery of health services, continued monitoring of local health trends will help policymakers track successes and promptly address challenges as they arise. Subnational benchmarking ought to occur regularly in Nigeria and throughout sub-Saharan Africa to inform local decision-making and bolster health system performance.