The potential impact of declining development assistance for health on population health in Malawi: A modelling study

Abstract

Background: Development assistance for health (DAH) to Malawi will likely decrease as a fraction of Gross Domestic Product (GDP) in the next few decades. Given the country's significant reliance on DAH for the delivery of its healthcare services, estimating the impact that this could have on health projections for the country is particularly urgent.

Methods and findings: We use the Malawi-specific, individual-based "all diseases-whole health-system" Thanzi La Onse model to estimate the impact that declining DAH could have on health system capacities, proxied by the availability of human resources for health, and consequently on population health outcomes, in the period 2019-2040. We estimate that the range of DAH forecasts considered could result in a 7.0% (95% confidence interval (CI) [5.3, 8.3]) to 15.8% (95% CI [14.5,16.7]) increase in disability-adjusted life years compared to a scenario where health spending as a percentage of GDP remains unchanged. This could cause a reversal of gains achieved to date in many areas of health. The burden due to non-communicable diseases, on the other hand, is found to increase irrespective of yearly growth in health expenditure, assuming current reach, and scope of interventions. Finally, we find that greater health expenditure will improve population health outcomes, but at a diminishing rate. The main limitations of this study include the fact that it only considered gradual changes in health expenditure, and did not account for more severe economic shocks or sharp declines in DAH. It also relied on key assumptions about how other factors affecting health beyond healthcare worker numbers -such as consumable availability, range of services available, treatment innovation, and socio-economic and behavioural factors-might evolve.

Conclusions: This analysis reveals the potential risk to population health in Malawi should current forecasts of declining health expenditure as a share of GDP materialise, and underscores the need for both domestic and international authorities to act in response to this anticipated trend.

Fig 1. Evolution of health burden and life expectancy under different expenditure scenarios.

(a) Total yearly DALYs (averaged over two-year periods) incurred under different expenditure scenarios. (b): Life expectancy (averaged over two-year periods) achieved under different expenditure scenarios. In both plots, solid lines represent mean values, while shaded areas indicate the 95% CIs defined in the Methods section.

Fig 2. Scaling of overall health burden with yearly expenditure growth.

(a) Total DALYs incurred in the period 2019–2040 (inclusive) as a function of the yearly expenditure growth, as well as the normalised total expenditure (NHE) over that period under each scenario (top x-axis), as defined in Eq. 2. The best-fit parameters for the function shown are summarised in the table inside the plot, while the linear best-fits to the first three levels of yearly expenditure growth considered are only included for visual guidance. (b) Percentage DALYs averted in the same period compared to the “GDP growth” scenario, where this captures the current level of health expenditure as a fraction of GDP. The blue shaded area shows the 95% UI in the IHME forecast. In both plots, points represent mean values, while error bars indicate the 95% CIs defined in the Methods section.

Fig 3. Impact on different areas of health.

Average yearly DALYs incurred between 2019 and 2040, grouped into three meaningful categories: HTM, including HIV/AIDS, TB, and malaria (a); RMNCH, including lower respiratory infections, childhood diarrhoea, maternal and neonatal disorders, and measles (b); and NCDs, including COPD, cancers, depression/self-harm, diabetes, epilepsy, heart and kidney disease, and stroke, alongside RTIs (c). In the case of HTM, DALYs in this area are additionally shown broken down by individual causes. In all plots, points represent mean values, while error bars indicate the 95% CI defined in the Methods section. Horizontal lines indicate the yearly DALYs burden for each cause in 2018. This means that any scenario above the respective 2,018 level incurred, on average, a worsening of the health burden due to that cause over the 2019–2040 period, whereas any scenario below the respective 2,018 level incurred an improvement. Finally, the blue shaded area shows the 95% UI in the IHME forecast. A breakdown of the time evolution of the burden due to all of these causes of ill health individually is included in S4 Text.