State of inequality in malaria intervention coverage in sub-Saharan African countries

Over the last decade, malaria programs have been transformed with increased investment, new technologies, economic development, and shifting paradigms in global health [1]. Prior to 2008, control efforts in the region focused on groups at highest risk of adverse outcomes, especially pregnant women and children under the age of five [2]. Control programs prioritized distribution channels most likely to reach these populations such as routine vaccination campaigns, antenatal clinics, and social marketing where malaria commodities were subsidized or rendered for free [3, 4]. These channels benefitted most those with access to health facilities, which tended to be the less poor urban households [5, 6]. In 2008, a global call from the UN Secretary General placed the universal coverage of malaria interventions at the center of control efforts [7]. The aim was reflected in the Global Action Plan for Malaria [8, 9], launched the same year by the Roll Back Malaria, laying out a blueprint for control, elimination, and eventual eradication of malaria. It states that a vision of a “malaria free world” is brought about by “achieving and sustaining universal access to and utilization of preventive measures; achieving universal access to case management in the public and private sectors and in the community; and accelerating the development of surveillance systems” [9]. The aim broadened the scope of control efforts in endemic countries to include all ages and led to programs shifting from targeted distribution of insecticide-treated nets (ITNs) to mass distribution campaigns and community delivery supported by routine services. In malaria case management, fast-acting artemisinin-based combination therapies (ACTs) replaced monotherapy (‘non-artemisinin’ drug formulations) as the recommended first-line treatment for uncomplicated malaria in 2004 in response to emerging parasite resistance [10]. Initially costly and scarce, these drugs became increasingly available through the Global Fund and the novel financing mechanisms it enabled such as the Affordable Medicines Facility – malaria [11, 12]. Innovative delivery channels were introduced in many endemic countries to directly tackle distributional failures; these engaged the informal health sector and emphasized community health workers in servicing vulnerable and hard-to-reach populations [13]. As endemic countries adopted and adapted to these new policies, the malaria service gap between the poorest and the least poor began to narrow [14].

Economic gradients and progress toward equity in malaria control and curative interventions have been documented in several broad strands of literature. A large number of studies evaluated ITN coverage following a mass-distribution campaign in narrow geographic areas or within a single country showing improvement in equity after the campaign [15‐18]. Socioeconomic status has also been included as one of covariates explaining coverage and use of malaria interventions; while its relative importance varied by setting and model specification, it has generally been found to be an important predictor with both use and coverage increasing with economic status [19‐21]. Several multi-country analyses included ITN use among children under the age of five into a composite index for maternal and child health to track country and regional progress toward the Millennium Development Goals [22]; these studies found that, although relative inequalities have generally been narrowing, appreciable differences between the poorest and the least poor persist. There have been a number of evaluations that focused on a single malaria intervention (most commonly ITNs) [23, 24]. To the best of our knowledge, only one earlier study [5] has assessed equity across a range of malaria interventions in a large number of endemic African countries. The analysis drew on survey reports from 2006 to 2008 and, as such, was constrained to comparisons between asset-wealth quintiles. It showed that, in 13 out of 25 countries, ITN ownership was equitable while malaria treatment for febrile children and intermittent preventive treatment in pregnant women were predominantly inequitable. There has been no comprehensive assessment of malaria interventions with respect to equity since.

Monitoring equity in intervention coverage takes a new meaning today as malaria control programs mobilize efforts toward elimination [25]. Scale-up of interventions and introduction of new technologies can enhance inequities if implemented without an explicit focus on equity [26]. While wealth-related gradients have been narrowing, these gains were achieved at very low baseline coverages and with extensive donor support [27]; sustaining these gains and scaling-up toward populations at highest risk of malaria might be a challenge for many countries [28]. This study presents an update on the state of inequality in malaria intervention coverage in endemic sub-Saharan African countries through to 2015. We follow a standard framework for assessment of equity in health [29] and interpret the estimates it yields with respect to their relevance to control programs.

Overall, across countries and coverage indicators, malaria programs in the region are inequitable (Table 1). Yet, in 16 out of 30 countries, ITN ownership is distributed equally or favors the poorest; in over half of the countries, ITN use and IRS are distributed equally or favor the poorest; for interventions relying on the formal health sector, such as intermittent preventive treatment in pregnancy (IPTp) and access to formal care, provider inequities are not evident in nearly half of the countries. These categorizations refer to relative differences in coverage of malaria interventions across the asset-wealth continuum within countries and do not take into account the level of intervention. For instance, in Benin, where household access to at least one ITN for every two persons is disproportionately higher among the least poor (CIX = 0.07), the poorest strata have better access to ITNs in absolute terms than the poorest strata in Cote d’Ivoire, where the indicator is equitably distributed (42% and 34%, respectively) (Additional file 2: Table SA3).

Figure 1 illustrates the strength of economic gradients in distribution of malaria interventions qualified above, thereby addressing the question of the existing level of inequality, comparing this across interventions. The CIX for access to formal care – the most consistently inequitable coverage indicator – is plotted against country mean (Fig. 1a). Inequity is highest in settings with lowest access. Most high PfPR_2-10 countries (in shades of red) are at this end of the distribution; Ghana is one notable exception. For all other intervention coverage indicators, we found little correlation between level and degree of inequality (Additional file 3); that is, higher coverage is not necessarily described by a more equitable distribution nor is low coverage consistently inequitable.

Compared to access to formal care, ITN use is distributed more equitably (most country estimates above the equality diagonal) (Fig. 1b). High PfPR_2-10 countries with large inequities in access to formal care (positive values of concentration index on y-axis) appear to be equitable or pro-poor in distribution of ITN use (clustering of points around zero and –10 values of rescaled CIX on x-axis). In Congo, Democratic Republic of the Congo, Togo, and Uganda, access to formal care strongly favors the least poor, while ITN use favors the poorest. On the opposite end of the plane are a handful of lower PfPR_2-10 countries, including Burundi and Malawi, where inequality in ITN use exceeds that in access to formal care.

Anti-malarial treatment is equitable or slightly pro-poor in low PfPR_2-10 countries (shown in shades of blue) (Fig. 1c). In high PfPR_2-10 countries, antimalarial treatment favors the least poor, though the degree of inequality is lower than that estimated for formal care. There are a few exceptions to this general trend. In Democratic Republic of the Congo and Nigeria, both medium PfPR_2-10 countries, the distribution of antimalarial treatment is more inequitable than that of formal care. In Togo, a high PfPR_2-10 country, antimalarial treatment is distributed equitably despite large gradients in formal care favoring the less poor. Distribution of first-line medication is consistent with that of receipt of any antimalarial described above (Additional file 3: Figure SA2). The two indicators are correlated most closely in settings with lower PfPR_2-10, where first-line drugs account for a large fraction of antimalarials dispensed. In higher PfPR_2-10 countries, inequities in first-line treatment are lower than those estimated for any antimalarial medication; however, access to these medications is also significantly lower. For instance, in Nigeria, about 33% of children with fever are treated with an antimalarial, while only about 6% receive the first-line drug; the least poor are more than twice as likely as the poorest to receive either of the medications (Additional file 2: Tables SA8-SA9).

ITN use is the most equitably distributed malaria intervention coverage indicator. Figure 2 illustrates changes in its distribution over the last decade. ITN use expanded in all countries (positive values of annual absolute change on the x-axis) between 2005 and 2015 (Fig. 2a); it increased somewhat faster between 2005 and 2010 compared to the later period. In most countries, these expansions were pro-poor (positive values of excess change on the y-axis). Where expansions favored the least poor (negative values of excess change on the y-axis), the relative gains in coverage were smaller in absolute terms than the excess change in settings where growth was pro-poor (positive values of excess change on the y-axis). It is primarily in these countries where interventions expanded faster in the least poor strata where we see inequities in distribution of ITN use today (in shades of grey). Figure 2b presents a more nuanced picture of intervention expansion path in three countries, showing that failure to sustain a pro-poor focus in expansion yields greater inequality in later periods (Rwanda), that high baseline inequality in service coverage defies pro-poor growth (Tanzania), and that, in some countries, expansion in ITN use consistently prioritized the poorest strata (Ghana).

While relatively smaller, improvements in coverage occurred across all interventions delivered by malaria control programs (Additional file 4). In about half of the countries, these expansions were pro-poor or distributed equally across asset-wealth strata for ITN ownership and access to formal care. Changes in IPTp were mainly driven by improvements in the least poor strata from 2005 to 2010, with somewhat more equitable growth in the later period. Interpretation of changes in use of antimalarial medication with respect to program performance is confounded by changes in malaria epidemiology and the shift toward diagnostically confirmed treatment. Overall use of antimalarials decreased in most countries with a relatively lower reduction in the least poor compared to the poorest strata. There is limited evidence that the direction of change reversed toward expansion in the poorest strata in the most recent period.

By pursuing malaria intervention coverage targets, countries aim at a broader health and health equity agenda. Tabulating malaria prevalence for a subset of DHS/MIS countries that included malaria diagnosis, we found that parasitaemia remains disproportionally concentrated in the poorest; prevalence gradients are substantially higher than what might have been expected given distribution of coverage indicators (Table 2). Even in countries where malaria interventions are equitable or favor the poor (i.e., Ghana) differences in burden between the poorest and least poor are large.

Decisive strides against malaria over the last decade have enabled significant expansion of control programs across endemic countries [47]; in some, these efforts also achieved equity in intervention coverage [5, 22, 24, 48, 49]. More recent data analyzed here show that, not only has the number of countries with equitable or pro-poor ITN ownership and use increased by 2015, but also that the degree of inequality in countries where inequities persisted is modest with the exception of a handful of low PfPR_2-10 settings. Large wealth-related gaps previously shown in access to antimalarial medication [5] narrowed, with 19 out of 30 countries reporting equitable or pro-poor coverage. These include most Affordable Medicines Facility – malaria pilot phase countries, except Niger and Nigeria, where access to antimalarial medication remains substantially inequitable. Important gains have also been made in delivering IPTp to poor women – in 13 out of 27 countries, this preventive intervention is equitable. In Namibia and Senegal, all malaria control interventions are distributed equitably; whereas in Ghana and Liberia and in Sierra Leone, all control interventions, except for IPTp and IRS, respectively, are distributed equally.

Where asset-wealth differences in coverage persist, the relative performance of malaria interventions with respect to equity both within and between countries provides useful benchmarks. The large economic gradients in the distribution of curative interventions across countries highlight the difficulty of delivering routine care to the poor, but differences in relative degree of inequality (i.e., a highly inequitable CIX of 0.27 in distribution of access to formal care in Guinea compared to a CIX of 0.02 in Democratic Republic of the Congo) suggest that it is feasible to substantially reduce these inequities. Benin and Mali achieve coverage and use of ITNs at rates that are approximately 20 and 40 percentage points above those estimated for formal care; in these countries, ITN use is also distributed equitably across asset-wealth strata in contrast to large economic gradients in curative care. The differential performance of interventions within control programs indicates that there is technical capacity to identify and deliver services to the poor; what is needed is implementation insight to transfer efficiency of high-performing interventions to those lagging behind [50]. Review of malaria programs covering health systems, financing, and operational contexts is needed to gather and synthesize institutional expertise from within the region.

Differences in the level and distribution of malaria interventions within endemic countries also impact adoption, prioritization, and deployment of new malaria tools. Our analyses suggest that, in high endemic settings, interventions deployed via campaigns are likely to be more equitably distributed than those requiring routine delivery through the formal sector. These considerations are immediately relevant for countries currently considering introduction of new malaria interventions such as malaria vaccines [51]. The economic gradients estimated here for access to the formal sector are representative of the EPI coverage in endemic settings [52], suggesting that much of the vaccine’s impact is likely to be missed due to distributional failures unless routine immunization is supplemented with periodic intensification and outreach campaigns.

In many high PfPR_2-10 countries, coverage of malaria interventions remains low; how programs scale-up has important implications for equity [24], in agreement with a previous analysis [20] showing that, where ITN use increased, gains were made across the asset-wealth continuum, often with accelerated growth in the poorest stratum. We find this conclusion also holds for other malaria coverage indicators, indicating that expansions in malaria programs in the region have largely circumvented the inverse care law [51]. Yet, despite pro-poor growth, progress toward equitable distribution of malaria interventions in many settings is limited by past high levels of inequality; turning these programs around is a challenge. For these countries, failure to sustain a focus on equity in service delivery leads to greater inequality in subsequent periods, while narrowing the gap between poorest and least poor requires disproportionately faster and sustained growth in the poorest strata.

These dynamics can partially explain systematic differences in the level and distribution of malaria coverage indicators by PfPR_2-10. Most high PfPR_2-10 countries were late to scale-up ITNs; these efforts relied heavily on donor assistance that financed mass distribution campaigns outside the formal health sector, enabling equitable coverage in these settings. In lower PfPR_2-10 countries, ITN delivery relied more on routine channels propagating disparities in access to formal care to ITN coverage. Distributional failures in the formal health sector limit progress toward equity across malaria interventions as programs rely on routine channels to maintain coverage between campaign rounds (i.e., ITN, IPTp), to strengthen access for priority groups, and to inform planning and resource allocation decisions with data sourced from governmental health facilities. Tackling malaria requires a holistic view of the program; to succeed, there should be a focus on strengthening the health systems at large.

Evidence of economic gradients in intervention coverage lends itself to a clear programmatic interpretation – the poor are left behind; but analogous interpretation of equitable or pro-poor gradients may be misleading. First, as dimensions of social and economic disadvantage overlap, analyses limited to broad comparisons between the poorest and the least poor might average over important gradients understating the true degree of disparity [53]. Second, it is not obvious whether gradients defined over an asset-wealth index capture policy-relevant dimensions of economic or social disadvantage in malaria-endemic countries [39, 40, 54]. Finally, geographic and demographic clustering of malaria in low transmission settings might undermine the capacity of conventional tools to identify and monitor high-risk populations [25]. To illustrate, we analyzed ITN use by a number of common equity stratifiers and an interaction of these dimensions with the asset-wealth index for Liberia, a country with a pro-poor or egalitarian distribution of most malaria interventions (Fig. 3). Here, ITN use is equally distributed across the economic strata at the national level (solid and dashed pink lines); it is effectively equal to national average when stratified by place of residence, education, or PfPR_2-10 (colored bars) (Fig. 3a). However, large gradients for Liberia emerge when either of the dimensions is overlapped with the asset-wealth index (Fig. 3a). There are large regional differences around the national average and even stronger economic gradients within these regions. While pro-poor at national level, ITN use is concentrated among the least poor in rural settings, areas of high PfPR_2-10, and in all regions except for the South Central region of Liberia, which includes the country capital.

Equality across wealth quintiles in coverage, or concentration of interventions in the poorest, is likely to be more effective in reducing burden than concentration of interventions in the richer quintiles because of the association of malaria infection with low socioeconomic status (Table 2) [55]. The areas and individuals at highest risk contribute disproportionately to onward infection [56], so targeting them is highly effective if they can be identified [57]. However, identifying them is challenging [58]. Economic and social dimensions have a role in this, and thus need to be considered as important components (alongside epidemiological and geographic data) in new approaches in malaria mapping, in the development of more comprehensive measures of malaria vulnerability [59, 60], and in analytical tools to help malaria control programs direct efforts towards vulnerable sub-populations [61, 62].

By eliminating disparities in proximal determinants (access to preventive and curative interventions) of malaria morbidity and mortality, control programs in endemic countries address only a part of the causal pathway from exposure to disease; the underlying socioecological conditions could manifest through other mechanisms further perpetuating health inequities [58]. The consideration is relevant to evaluation and performance targets for control programs. Differential tracking of coverage indicators by malaria endemicity and/or other context-specific dimensions of social or economic disadvantage, as shown here, can enhance equity analyses [28, 32] by focusing most directly on populations at risk. Expanding the scope of monitoring indicators to also include quality dimensions could help direct efforts toward increasing efficiency of current control tools. To this end, the equity effectiveness framework can be applied to disentangle the impact of various factors on the gap in the effectiveness of interventions across socioeconomic gradients [52]. For instance, compounding of disadvantage along the service delivery path has been shown for malaria case management, where the poor are not only less likely to seek care but also less likely to do so in the formal sector, or to receive appropriate diagnosis and treatment [32].

There are a number of limitations to consider when interpreting our findings. Due to varying survey cycles between countries, there is some inconsistency in reference years for reported statistics. To the extent that expansion in malaria services accelerated closer to 2015 or accelerated differently with respect to equity, estimates based on less recent surveys understate the level of service provision, overstate the degree of inequality in their distribution across the wealth strata, and overstate differences in the two statistics between countries. Timing of campaigns (i.e., IRS or ITN distribution) with respect to survey implementation is another variable affecting the estimated level of malaria services and comparisons thereof between countries. Finally, by design, DHS surveys are to be conducted during low transmission season, while MIS is conducted at peak. Survey implementation in the field, however, spans months averaging over the seasonality pattern in the data collected [30]. To the extent that in some countries the pattern is maintained comparisons will be biased.

The global health and development agenda place a special emphasis on universal health coverage [63‐65]. The Global Technical Strategy for Malaria 2016–2030, in particular, argues for universal access to interventions for malaria prevention, diagnosis, and treatment as a path toward elimination [63]. For endemic countries, given the historically low access among the poorest, adopting this vision would require overall gains in coverage to be accompanied with accelerated improvements in disadvantaged populations. Our analysis testifies to the progress made by countries in the region in narrowing economic gradients in malaria intervention coverage and highlights the scope for continued monitoring of these programs with respect to equity.