About 1,000 extra maternal deaths a year: modelling the USAID cuts in six countries

A modelling study in Health Policy and Planning attempts to put a number on one narrow consequence of the January 2025 suspension and subsequent termination of most United States Agency for International Development programmes: how many additional women die in childbirth in six West and Central African countries when that money stops [s1].

The headline estimate is roughly 1,000 additional maternal deaths in a single year across Burkina Faso, the Central African Republic, Chad, Mali, Niger and Nigeria — a 45% average increase among the populations the study covers [s1]. Understanding what that number does and does not describe requires reading the model's assumptions closely, and the author is unusually direct about them.

What the model does

The study uses a deterministic model built on regional health expenditure elasticities [s1]. In outline: it estimates how the sudden withdrawal of foreign aid changes health spending among populations in humanitarian need, then translates that spending change into a change in the maternal mortality ratio, measured as deaths per 100,000 live births [s1].

Two features of the setup matter more than the arithmetic. First, the analysis is restricted to populations in humanitarian need, not to each country's whole population — these are the groups most dependent on externally financed services and least buffered by anything else. Second, the model assumes that no immediate domestic or external financing substitutes for the lost resources [s1]. That is a deliberate boundary condition, not an oversight. It defines the scenario as "the money stops and nothing fills the hole in the short term."

The findings

Against a baseline of approximately 2,900 maternal deaths predicted across the six countries in 2025 among these populations, the model projects maternal deaths rising by 45% on average, producing roughly 1,000 additional deaths within a single year [s1].

The distribution across countries is uneven in a way that reflects two different things. Niger shows the largest proportional increase, over 90% [s1] — a measure of how concentrated its maternal health provision for these populations was in aid-financed services. Nigeria shows the largest absolute increase, more than 300 additional deaths [s1], which is largely a function of population size.

Sensitivity analyses using alternative elasticity scenarios did not overturn the results [s1].

The limits, stated plainly

The author frames the output as conditional estimates, intended to inform discussions on health financing sustainability, transition planning, and risk mitigation [s1]. That framing is doing real work and should not be skipped over.

A deterministic elasticity model is not a measurement of deaths. It is a projection that says: given an established statistical relationship between health spending and maternal mortality in this region, a spending shock of this size implies a mortality change of roughly this size. The relationship is estimated from historical data across settings, and applying it to an abrupt, discontinuous funding shock assumes the relationship holds under conditions unlike those in which it was measured. Gradual spending declines and a sudden stop may not affect mortality identically.

The no-substitution assumption cuts both ways. If governments, other donors, or non-governmental organisations partially backfill, the real figure is lower than modelled. If the disruption also degrades supply chains, staff retention or referral networks in ways that outlast the funding gap, it could be higher. Neither dynamic is in the model.

And the estimate is bounded to maternal deaths among populations in humanitarian need. It is not a count of all deaths, all maternal deaths, or all consequences.

Why maternal mortality is the sensitive indicator

The study's more general point is that maternal health outcomes in fragile settings are highly sensitive to financing discontinuities [s1]. That sensitivity has a mechanical explanation. Preventing a maternal death typically depends on a service being available at a specific moment — a skilled attendant at delivery, blood for transfusion, a functioning referral to surgical care. Unlike a chronic disease programme, where an interruption may show up in outcomes over months or years, obstetric care fails immediately when the inputs are not there.

That is also why maternal mortality tends to be among the first indicators to move after a financing shock, and why it is a reasonable early-warning metric for the wider effects of aid withdrawal.

What to watch next

This is a projection published roughly fourteen months after the funding decision it models, which means observational data covering the same period will begin to appear over the coming year. Facility-level delivery volumes, obstetric referral counts and confidential enquiry reports from the affected countries are the measurements that will eventually test the projection.

Until then, the appropriate reading of the 1,000 figure is the one the author gives it: a conditional estimate of what a specific financing scenario implies for a specific population, useful for planning, not a body count.

Sources

• Aid cut, lives lost: estimating the impact of USAID's withdrawal on maternal mortality in six African countries, Health Policy and Planning, 10 March 2026