Guerilla Evaluators
Bearing Witness to the Half a Million Deaths from Cuts to Aid
New estimates suggest that USAID cuts led to somewhere between 500,000 and 1.6 million preventable deaths in fiscal year 2025.1 Let’s pause. Half a million to 1.6 million men women and children died who shouldn’t have.
The number is so large it stops meaning anything. It becomes statistical, abstract, the kind of thing you read and then move on from. But somewhere in that number are mothers who did not survive childbirth because the program that trained midwives stopped. Children who died of diarrheal dehydration because the supply chain for oral rehydration salts was terminated. People living with HIV who developed AIDS because antiretroviral programs were cancelled. These deaths have names. They have families. They happened while official language insisted life-saving work continued.
The administration said life-saving work would continue. Secretary Rubio repeated it like a mantra in the early months of 2025: “life-saving medicine, medical services, food, shelter, and subsistence assistance” were eligible for waivers. The phrase appeared in briefing after briefing, as if saying it made it true. But programs that deliver some of the most cost-effective interventions we know were terminated anyway. Ready-to-use therapeutic foods, the emergency nutrition packets that prevent children from starving, sat in locked warehouses while children died. HIV vaccine trials ground to a halt midstream, years of research at risk of disappearing. Childhood immunization programs that reached hundreds of thousands of children across Nigeria and Côte d’Ivoire simply stopped. These are interventions backed by decades of rigorous evidence, embedded in global health policy because they work.
There is on-the-ground evidence now. Rising malnutrition mortality in northern Nigeria, Somalia, and in the Rohingya refugee camps. Rising food insecurity in northeast Kenya, linked to the global collapse of therapeutic food supply chains. Spiking malaria deaths in northern Cameroon, linked to breakdown in the global supply of antimalarials. A health crisis across Africa as HIV programs were disrupted. Dr. Atul Gawande, who headed global health at USAID during the Biden administration, stated in October: “Hundreds of thousands of deaths have already occurred.”
Bearing Witness: Closing the Gap
The researchers at the Center for Global Development are doing the work of bearing witness. They are counting what the administration refuses to count. They are naming what official language works hard to obscure.
Secretary Rubio repeated “life-saving work will continue” like a mantra. He was counting on the fact that most deaths would happen quietly, in countries Americans do not think about, from causes that would not show up in headlines for months or years. He was counting on the lag time built into measurement systems. He was counting on the gap between harm and its visibility. The researchers are closing that gap, documenting the scale of preventable death that official language works to obscure.
The Gap Philanthropy Cannot Fill
Project Resource Optimization has mobilized nearly $46 million from private donors to restart some cancelled programs. They estimate this emergency funding will reach over 3.6 million people. This is extraordinary work, done by former USAID staff who saw the collapse coming and built triage systems to identify which programs could save the most lives per dollar if rapidly refunded.
But PRO had to be ruthless in their prioritization. They could only focus on programs that private philanthropy could feasibly pick up. That meant excluding critical infrastructure that only governments can sustain: the Famine Early Warning System, which facilitates collective action on famine data; global procurement of pharmaceuticals and ready-to-use therapeutic food; vaccine supply chains that reach millions of children; research trials that require multi-year funding commitments. These are not things private donors can replace, no matter how much they care. Even the Gates Foundation, which has more resources than most governments, has said clearly: private foundations cannot replace US government funding. Not at this scale. Not for this work.
The gap between what philanthropy can do and what government did is not just about money. It is about the infrastructure of trust, the systems of coordination, the institutional relationships that make large-scale humanitarian work possible. You cannot replace those with emergency grants, no matter how well-intentioned.
The Costs We Will Not Recover
The immediate deaths are only part of the damage. Even if aid budgets are restored in a few years, the chaotic dismantling of USAID will have permanent consequences on the capacity to deliver aid worldwide. Partnerships that took decades to build have been severed. Local organizations that relied on consistent funding have closed their doors. Research trials stopped midstream cannot simply resume. Clinical samples sit in storage, years of data at risk of becoming worthless. The institutional knowledge held by USAID staff, scattered when the agency was gutted, does not reconstitute easily.
This is not just about technical capacity. It is about trust. Why would a government invest in building health systems with US support if that support can be yanked without warning? Why would researchers commit to multi-year trials if funding can disappear overnight? Why would communities collaborate if collaboration makes them vulnerable to the whims of an administration that does not value their lives?
The costs of breaking these relationships will compound for years. Governments will hedge their bets, diversifying away from US partnerships. Researchers will seek more stable funding sources. Communities will remember what happened when they trusted American commitments. We will not get that trust back quickly. Maybe not at all. This is the deeper extraction. Not just the money withdrawn, but the relationships destroyed. Not just the programs cancelled, but the confidence shattered.
What We Do Now
I do not know how to write about half a million preventable deaths without the words feeling inadequate. The number is so large it stops meaning anything. It becomes statistical, abstract, the kind of thing you read and then move on from. But somewhere in that range is a farmer who died of malaria because the distribution network for bed nets collapsed. A teenager who didn’t survive tuberculosis because the treatment program shut down mid-course. An infant who starved because therapeutic feeding centers closed their doors. These deaths have names. They have families. They happened while official language insisted life-saving work continued.
The researchers at the Center for Global Development operated as guerrilla evaluators, refusing to wait for official accounting. They documented harm in real time, using the same rigorous methodology that has always guided cost-effectiveness analysis, but applying it to a crisis as it unfolds. Former USAID staff, scattered when the agency was dismantled, turned their expertise toward counting what the administration refused to count. This is what evaluation can do when it serves truth rather than convenience.
For evaluators and researchers: Use what you know. Document what’s being lost while there’s still time to document it. The infrastructure that took decades to build is being dismantled right now. The partnerships that made large-scale humanitarian work possible are fracturing. If we wait for the perfect data, we will be writing postmortems instead of sounding alarms.
For anyone who cares about this work: Project Resource Optimization is still working to restart cancelled programs. Their urgent and vetted list identifies the most cost-effective, life-saving work that private funding can still rescue. Every program on that list represents lives that can still be saved. Contact your representatives about restoring funding. Don’t let this disappear from view.
For Congress and anyone concerned with accountability: Someone needs to be held accountable for half a million to 1.6 million preventable deaths. These deaths need to be named in the official reco
rd, not buried in the lag time between harm and measurement. The administration cancelled programs that were working. People died. That deserves investigation.
This is the work of remembering versus the pressure to move on. Systems want us to forget. The next crisis will arrive, the next news cycle will turn, and the silence in those kitchens will become permanent. Staying present to what happened, refusing to let it disappear, insisting that these deaths matter: this is what we owe.
Anthralytic is a strategy and evaluation studio that helps mission-driven organizations clarify their impact using human expertise, data, and thoughtful frameworks.
These figures are modeled estimates of what has already happened in fiscal year 2025 (which ended September 30th), not projections of future deaths. The methodology calculates lives saved per dollar spent on specific interventions (HIV treatment, malaria control, tuberculosis programs, humanitarian relief), multiplied by documented funding cuts. The range of 500,000 to 1.6 million reflects two ways of counting the cuts: money that was actively being spent this year (the lower end), versus money that had been committed to future programs before they were cancelled (the higher end). The estimates are corroborated by on-the-ground reports of rising mortality and Dr. Atul Gawande's October 2025 statement that "hundreds of thousands of deaths have already occurred."
