This summer, colleagues and I published a paper in BMJ Global Health examining a question that sits at the heart of global health funding: How should we evaluate programs designed to improve child survival? Our conclusion may seem counterintuitive, especially coming from a researcher like me publishing studies on this topic. While under-five mortality remains one of the most important indicators in global health, relying on it alone to evaluate how well a program was implemented or inform funding decisions is problematic.
The timing of this conversation matters. As development budgets tighten and every investment faces greater scrutiny, funders are being asked to make increasingly consequential decisions with finite resources. The quality of those decisions depends not only on having evidence, but on asking whether we are looking at the right evidence.
At Integrate Health, we know the value of rigorous mortality research. Last year, we published evidence showing that children under five living in the areas where our integrated primary care model was implemented were 29 percent less likely to die than children in comparable areas. It is one of the findings I am most proud of because it represents what matters most: children’s lives saved.
Yet if that under-five mortality result were the only evidence available, I would still hesitate to use it as the basis for deciding whether our organization deserved additional investment. Not because under-five mortality is the wrong measure, but because it arrives too late to steer the program, cannot tell you whether the gains will outlast the funding, and, as a single study, cannot tell you whether they will hold anywhere else.
Under-five mortality is important, but it is only part of the story
The appeal of under-five mortality is evident. It is meaningful. It reflects the outcome we are all working toward. In a sector that can sometimes drown in indicators and reporting requirements, it offers some clarity.
That clarity has also made under-five mortality an influential metric in global health funding. Large impact evaluations are often designed around it, and their findings help shape decisions about which interventions should be expanded and which organizations should receive continued support.
There is nothing inherently wrong with this. We should absolutely seek rigorous evidence that programs improve child survival. The challenge is that one measure in isolation is inadequate to evaluate implementation in the real world, and optimal implementation is what funders should care about.
That distinction matters because under-five mortality tells us what happened. It rarely tells us why.
Imagine two organizations reporting similar reductions in under-five mortality. On paper, they appear equally successful. But one achieved those results by consistently delivering high-quality care through a resilient primary healthcare system that continues improving over time. The other benefited from contextual factors that may not exist elsewhere, from support that will not persist once external funding ends, or from parallel systems that cannot survive once that support is withdrawn.
Under-five mortality alone cannot distinguish between those two stories.
It cannot tell us whether children received the correct treatment at the right time. It cannot tell us whether health workers were well supervised, whether medicines were consistently available, whether care reached the most hard-to-reach communities, or whether improvements were driven by stronger health systems or by factors outside the program’s control.
Nor does it help organizations evaluate, iterate, and improve quickly. High-quality under-five mortality evaluations often take three to five years to complete before meaningful conclusions can be drawn. By then, implementation has evolved, policy factors have changed, funding priorities have shifted, and opportunities to strengthen programs in real time have already passed.
Better funding decisions require better approaches
If under-five mortality should not stand alone, what should accompany it? The answer is not more indicators for the sake of measurement. It is better insight into how impact is achieved.
Funders ultimately need confidence that an organization is not only improving child survival today, but building systems capable of sustaining those gains tomorrow. That requires understanding what happens long before mortality changes become visible. In the paper, we call these intermediate outcomes: the signals, measurable long before mortality shifts, that a program is working as intended.
- Are children receiving the right treatment when they need it?
- Is care reaching the communities furthest from the health system?
- Is access to care equitable?
- Is quality of care improving across facilities and Community Health Workers?
- Are providers consistently delivering evidence-based care as intended?
- Are supply chains functioning reliably?
- Are programs identifying problems and adapting workflows based on data?
These questions may seem less compelling than a mortality estimate, but they are often far more useful for driving impact. They reveal whether a program is moving towards optimization. They explain why child survival improves. And they provide insights that governments and implementing partners can actually replicate elsewhere.
Our recent paper also argues that no single study, however rigorous, should determine whether an implementation organization deserves continued investment. A systematic review found that of sixteen trials run at one site and then repeated at multiple sites, only one still detected a significant mortality reduction. That is not a critique of the interventions. It is a reminder that context shapes outcomes, and a single result, however rigorous, is a thin foundation for a funding decision.
Measuring what matters while improving what matters
This philosophy is why Integrate Health has invested in Measurement, Evaluation, Quality, and Improvement (MEQI). We do not see measurement as something that happens after implementation. We see it as part of implementation itself (visit Integrate Health’s Dashboard to learn more).
MEQI creates a continuous learning system. Rather than waiting years for an under-five mortality evaluation to tell us whether we succeeded, it helps us understand, throughout implementation, whether care is reaching patients, whether quality is improving, where bottlenecks exist, and how programs should adapt.
This serves two complementary purposes. First, it improves care today by identifying opportunities to strengthen implementation while programs are still underway. Second, it provides the explanatory evidence that helps us understand how improvements in child survival were achieved, whether those gains are likely to last, and whether similar approaches are likely to succeed elsewhere.
This does not replace under-five mortality. It makes under-five mortality more meaningful. Because when we understand the mechanisms behind impact, we are better equipped to reproduce it, strengthen it, and ultimately save more lives.
Raising the standard for evidence
None of this argues for measuring under-five mortality less. Quite the opposite. Child survival should remain the outcome by which we ultimately judge ourselves. But one indicator should not carry the weight of every funding decision. Nor can an indicator like child mortality be meaningfully measured on the annual cycle many funders require.
The strongest investments will come from looking at under-five mortality alongside the evidence that explains how those gains were achieved, whether they can be sustained, and whether they can be replicated elsewhere. In practice, that means asking implementers not only what their mortality data show, but how they know care is reaching the right children, how they measure quality, and how quickly they turn that evidence into course corrections.
In a time of constrained resources, global health does not need less evidence. It needs richer evidence. Because the goal is not simply to identify interventions that have saved children’s lives. It is to understand which ones can keep saving them.