Introduction and Context
Neglected Tropical Diseases affect the world's poorest populations and carry an outsized burden in sub-Saharan Africa. Two of the most prevalent are Lymphatic Filariasis (LF), which causes permanent disfigurement and disability, and Soil-Transmitted Helminthiasis (STH), which stunts growth and cognitive development in children.
Both diseases are preventable and treatable through Mass Drug Administration (MDA) programs. Africa carries roughly 40% of the global LF burden, and a large share of the school-age children at risk of STH. With coordinated effort, elimination as a public health problem is achievable for both.
This analysis draws on WHO surveillance data and World Bank socioeconomic indicators to test seven hypotheses about what drives national treatment coverage and whether programs are keeping pace with disease burden.
The GDP distribution across LF and STH endemic countries in Africa is sharply right-skewed. Most countries cluster below $2,000 per capita. A handful of outliers sit substantially higher. This concentration of low-income countries shapes every analysis that follows: programs in this region operate under genuine resource constraints, and coverage outcomes must be interpreted in that context.
Elimination is within reach. But it requires high national coverage sustained over multiple years, and the barriers are not purely economic.
Coverage Has Improved Consistently Over Time
National treatment coverage has grown for both diseases since 2000, reflecting the sustained scale-up of WHO-coordinated MDA programs.
The regression of year on national coverage yields a slope of +1.32 percentage points per year for LF (p < 0.001) and +2.28 percentage points per year for STH (p < 0.001). Both trends are statistically robust and hold even after accounting for the disruptions of 2020 and 2021.
The R² values are modest (0.06 for LF, 0.15 for STH), which means time alone explains only a fraction of the variance in coverage. Other contextual factors clearly matter. But the direction of the trend is consistent across both diseases and across the full time span of the data.
The choropleth map of LF burden shows the countries with the largest at-risk populations: Nigeria, the Democratic Republic of the Congo, Mozambique, and Tanzania. These are also the countries where the scale of delivery challenges is greatest.
Some Countries Outperform What Their GDP Would Predict
When coverage is regressed on GDP per capita, the residuals reveal that several low-income countries consistently exceed expectations while some wealthier countries fall short.
Fitting a simple OLS regression of national coverage on GDP produces a predicted coverage for each country-year observation. Countries with large positive residuals are outperforming their economic situation; those with large negative residuals are underperforming relative to peers at the same income level.
Burkina Faso is the clearest overperformer for LF, achieving high coverage at GDP levels where most countries fall far short. Burundi shows a similar pattern for STH. By contrast, Angola and the Central African Republic consistently underperform, suggesting that economic resources have not translated into effective program delivery.
Governance quality, external aid effectiveness, and sustained political commitment appear to matter more than income alone in determining whether an MDA program reaches the people it targets.
Rising GDP Predicts Lower Coverage the Following Year
Higher GDP and improved sanitation in one year are associated with lower national coverage the next. This is a signal of program success, not failure.
A lagged regression of coverage on prior-year GDP yields a coefficient of −0.0026 (p = 0.003). The same negative relationship holds for lagged sanitation access. Countries that crossed disease elimination thresholds in prior years appropriately scaled back mass drug campaigns, which reduces coverage figures even as health outcomes improve.
The interactive chart below allows selection by country, making it possible to trace individual trajectories. The negative slope is consistent across most countries, particularly those where LF has been near-eliminated at the district level.
Coverage Is Shaped by Multiple Socioeconomic Factors
A multivariable OLS model reveals that the key drivers differ between the two diseases, with urbanization critical for LF and sanitation infrastructure critical for STH.
For LF, the strongest positive predictor is urbanization rate (p = 0.001), followed by ODA received per capita (p = 0.005). Health expenditure per capita is not statistically significant for LF (p = 0.297). GDP carries a negative coefficient, consistent with the program-completion interpretation from H3.
For STH, sanitation access is the dominant positive predictor (p < 0.001), and domestic health expenditure also reaches significance (p < 0.001). This reflects how STH control depends heavily on improved infrastructure and school health systems rather than urban density alone.
ODA per capita is significant for both diseases (LF: p = 0.005; STH: p = 0.013), indicating that international aid is a consistent enabling factor regardless of which disease is being treated.
The parallel coordinates plots below visualize the relationship between all predictors and coverage simultaneously. Moving the coverage axis to the rightmost position makes it possible to follow high-coverage profiles back to their associated predictor values.
Drug Combination Significantly Affects LF Coverage
A one-way ANOVA finds statistically significant differences in national coverage rates across MDA drug types (F = 3.764, p = 0.0011).
The Tukey HSD post-hoc test identifies the most meaningful contrast between programs using diethylcarbamazine combined with albendazole (DEC+ALB) and programs that have not yet started. Triple-drug regimens (IVM+DEC+ALB) show higher median coverage but also wider variance, reflecting that more ambitious programs carry more implementation uncertainty.
This result supports the hypothesis that treatment choice is not only a clinical decision. The combination used correlates with programmatic scale, funding access, and the national health infrastructure required to deliver multi-drug regimens safely.
Programme Scale and Burden Responsiveness
Treatment delivery tracks population need closely for LF but lags more for STH, particularly for pre-school-age children outside school-based delivery systems.
H6: Scale and Coverage. The correlation between number of people treated and programme coverage percentage is positive but modest (r = 0.28, R² = 0.078). At low treatment volumes, coverage percentages vary widely. At high volumes, programmes tend to cluster between 75 and 85 percent. STH shows a more stable relationship across scales, consistent with the standardized school-based delivery model used for that disease.
This result highlights that absolute treatment numbers and coverage percentage measure different things. A programme can treat many people while still reaching only a small fraction of the eligible population, particularly in high-burden, low-density settings.
Treatment Is Burden-Responsive but Does Not Fully Close the Gap
For LF, the correlation between population requiring preventive chemotherapy and number of people treated is r = 0.75 (p < 0.001, R² = 0.578). The regression coefficient of 0.333 indicates that for every three people identified as needing treatment, approximately one receives it. Programs are prioritizing high-burden areas, but delivery reaches only a fraction of total need.
For STH, the relationship is weaker (r = 0.60, R² = 0.246), and the gap is larger for pre-school-age children. Children under five fall outside the school-based distribution system and require community-based outreach to reach. That outreach is less consistently funded and harder to scale.
Country Trajectory Clusters
K-means clustering on 20-year national coverage time series groups countries into three distinct program trajectories: strong improvers, stalled programs, and irregular coverage.
Rather than treating all countries as comparable observations in a regression, clustering allows the data to group countries by the shape of their coverage trajectory over time. Three clusters emerge from the LF data. The first group shows consistent year-on-year improvement, often from a low starting point. The second group reached a plateau after initial gains and has not advanced significantly in the past decade. The third group shows irregular or declining coverage, associated with conflict, leadership transitions, or interrupted donor funding.
This classification adds granularity that aggregate trend statistics obscure. A country in the third cluster and a country in the first may show similar average coverage over a 20-year period while having fundamentally different program trajectories. Identifying cluster membership can help prioritize where structural support is most needed.
What the Data Shows
Africa's NTD eradication programs have made real and measurable progress since 2000. Coverage trends are positive for both LF and STH. Treatment delivery is broadly responsive to population need. The programs are scaling up.
The counterintuitive finding, that higher GDP predicts lower coverage, reflects program success rather than failure. Countries that cross elimination thresholds appropriately shift from mass campaigns to targeted surveillance. The negative correlation is a marker of transition, not decline.
Governance and aid effectiveness separate outperformers from underperformers more reliably than income does. Burkina Faso and Burundi demonstrate that strong programmatic commitment can offset resource constraints. Angola and the Central African Republic show the reverse. The critical variable is whether the delivery infrastructure exists and whether the political will to sustain it does too.
The gap between need and delivery remains substantial. For LF, programs reach roughly one in three people who require preventive chemotherapy. For STH pre-school-age children, the fraction is lower. Closing that gap will require investment not just in drugs but in the community health infrastructure that delivers them to people outside the reach of school-based systems. That is where the work remains.
Limitations: WHO data quality varies across countries and years. Linear interpolation smooths real fluctuations. COVID-19 creates artificial dips in 2020–2021. ODA figures are broad country-level aggregates that do not isolate NTD-specific spending.
LF vs STH Coverage Balance by Country
The difference between LF and STH national coverage reveals where each country's program efforts are concentrated and where imbalances remain.
Blue shading indicates countries where LF coverage exceeds STH coverage. Red shading indicates the reverse. A near-zero delta suggests balanced treatment across both diseases. Strong disparities may reflect disease-specific funding streams, different delivery mechanisms, or data quality differences between the two surveillance systems.
Nigeria, the Democratic Republic of the Congo, Mozambique, and Tanzania consistently show large LF-focused differentials, corresponding to their status as the highest-burden LF countries on the continent. Countries with strong school health systems tend toward balanced or STH-favored differentials.