In 2021, OECD countries appeared to discover prevention. According to the OECD’s Health at a Glance 2023, the share of health spending that went to prevention roughly doubled on average compared with 2019, reaching more than 5%. In Austria, Denmark, the Netherlands and the United Kingdom it jumped by more than six percentage points.
Then it went away again. The OECD’s later review of spending trends noted that prevention had fallen back to roughly the share it held before the pandemic.
The spike was not a change of heart. It was testing, surveillance and vaccination, emergency spending that the OECD itself described as time-limited rather than a planned investment in population health. In the UK, the jump was driven largely by the £15 billion allocated to NHS Test and Trace.
That episode tells you more about how health budgets work than any pie chart. The standard complaint is that prevention gets a tiny slice. True, but the size of the slice is not the real problem. The real problem is where the slice sits in the budget: in the part that can be cut without anyone being turned away at a door.
Two kinds of spending
Look at a health budget the way a finance ministry does. Part of it is committed. Hospital salaries, drug contracts, pensions, the debt on the new wing. Cutting that means strikes, closures and headlines about patients on trolleys.
The other part is discretionary. Smoking cessation services, sexual health clinics, school nutrition programmes, a campaign to find undiagnosed hypertension. Cut these and nothing visible happens this year. The consequences arrive five or fifteen years later, spread thinly across the population, and attributed to nobody.
Prevention lives almost entirely in the second category. So when money is tight, it is not that ministers weigh prevention against treatment and choose treatment. It is that treatment was never really on the table.
England shows the pattern clearly. The public health grant that funds local prevention services has been cut by more than a quarter per person in real terms since 2015/16, according to the Health Foundation, even after an increase for 2025/26. Over the same years NHS budgets, under intense pressure, still grew. Nobody decided that sexual health services mattered less than hospitals. The cut simply happened in the one place where it could.
The accounting makes it worse
There is a second problem, which is that health accounts are poor at seeing prevention in the first place.
The OECD’s own methodology notes that preventive work done inside a GP consultation is often hard to separate from ordinary outpatient care. A blood pressure check folded into a visit for a cough is prevention, but it is booked as treatment. So is a statin prescription, which is prevention in pharmacological form.
Meanwhile, some of the most effective prevention a government does never touches the health budget at all. Tobacco and alcohol taxes raise revenue rather than spend it. Road safety rules, clean water and building codes sit in other ministries. When advocates compare “prevention” and “treatment” lines in a health account, they are comparing a narrow and partly mislabelled category with everything else.
This matters because it shapes the ask. Campaigns to raise prevention to some target share of health spending push ministries toward the kind of prevention that is easy to count, which is usually programme spending, and away from regulatory and fiscal measures that are cheaper and often stronger.
Where the deaths actually are
The case for rebalancing is not sentimental. Health at a Glance 2023 estimated that across 26 OECD countries with data for 2020 or 2021, about 2.1 million premature deaths could have been prevented through effective primary prevention and public health measures, against almost one million considered treatable through better and more timely healthcare. The pandemic inflated the preventable figure, but the underlying pattern of chronic disease, tobacco, alcohol and injury was there before it.
So the burden sits mostly on the prevention side. The money sits mostly on the treatment side. And the budget structure guarantees that the gap widens every time there is a squeeze.
How the cut actually happens
It is worth slowing down to see the mechanism, because it is rarely a single decision.
Most health budgets are built incrementally. Each line starts from last year’s figure and is adjusted up or down. The baseline for committed spending is defended by people whose jobs and institutions depend on it: hospital directors, unions, professional bodies, suppliers with contracts. The baseline for a prevention programme is defended by a small public health team and, perhaps, an advocacy group.
Now imagine the finance ministry tells the health ministry to find savings equal to 3% of its budget in the middle of the year. The health ministry’s officials go line by line. Salaries cannot be cut without a negotiation that will take longer than the financial year. Drug contracts are signed. Hospital operating budgets are already overspent. What is left are the lines where spending can stop quickly and quietly: a screening outreach programme that has not started yet, a media campaign, a grant to local government, vacant posts in a public health unit.
Nobody in that room is choosing to damage population health. They are choosing the cuts that can be delivered by the deadline. Those cuts happen to fall on prevention, year after year, because that is where the budget is soft.
The result compounds. A prevention programme cut by a tenth in one year does not get restored the next, because the new baseline is lower. After a decade of such adjustments, the gap between intention and funding is enormous, even though no minister ever announced a policy of cutting prevention.
The strongest counterargument
Prevention advocates often claim that prevention saves money. Sometimes it does. Often it does not, and the people holding the budget know it.
In 2008 the New England Journal of Medicine published a widely cited perspective by Joshua Cohen, Peter Neumann and Milton Weinstein asking whether preventive care saves money. Reviewing the cost-effectiveness literature, they concluded that while some preventive measures do save money, most add to costs, just as most treatments do. Many are still good value, in the sense that they buy health at a reasonable price. But they are not free, and they do not pay for themselves.
There is also a longer-term effect that makes finance officials wary. People who avoid dying of a heart attack at 60 go on to live, and to need care, at 80. Prevention can shift costs later rather than eliminate them.
This counterargument is correct, and prevention advocates damage their own case by ignoring it. If prevention is sold as a cost saving and the saving does not appear, the credibility of the whole agenda suffers. The honest argument is different: prevention buys health, often cheaply compared with the treatments health systems fund without hesitation, and much of that health accrues to people who would otherwise have spent years in illness.
That argument is about value, not savings. It is weaker in a budget negotiation with a finance ministry looking for cuts, which is exactly why the budget design has to protect prevention rather than relying on its return on investment.
Second-order effects
The squeeze on prevention has consequences that ripple into treatment budgets later.
Cutting sexual health services, for instance, tends to show up years later as more late diagnoses and more complex care. Cutting drug and alcohol treatment affects emergency departments, police and courts. Those costs land on other budgets and other institutions, sometimes in other ministries, which is one reason nobody connects them to the original cut.
There is a workforce effect too. Public health professionals watch their programmes being cut first in every downturn and draw their own conclusions about careers. When a crisis hits and preventive capacity is suddenly needed, as it was in 2020, the expertise has thinned out and has to be rebuilt at emergency prices.
And there is a political effect. The more prevention depends on emergency money, the more it becomes associated with crises. Once the crisis passes, as the 2021 spike showed, the money goes with it, and the case for steady investment looks weaker because the programmes being cut were visibly temporary.
A worked scenario for a ring fence
Suppose a ministry wanted to protect a core prevention budget covering tobacco cessation, immunisation outreach, hypertension screening and sexual health. A durable design might look like this.
The protected amount is set in law as a floor, indexed to inflation and population, not as a percentage target that can be met by relabelling spending. Any reduction below the floor requires a vote in parliament, which means someone has to stand up and defend it. The money is committed in three-year envelopes, so a single bad year cannot erase a programme. Part of the revenue from tobacco and alcohol taxes is linked to the floor, not to cover all of it but to create a visible connection between harm and prevention. And the ministry publishes an annual prevention account that includes preventive activity inside primary care and pharmacy, not only programme lines.
None of this guarantees generous funding. It guarantees that cuts are deliberate and visible, rather than the default outcome of a mid-year savings exercise.
What this means for each actor
- For health ministries: stop negotiating prevention budgets on the promise of savings. Negotiate them as value, and push for legal protection that removes them from in-year cuts.
- For finance ministries: recognise that the savings extracted from prevention are often costs deferred to other budgets. Multi-year envelopes make the trade-off visible.
- For funders: when donor money pays for prevention, plan the domestic transition from the start. Programmes that exist only on external funding are the first to vanish when that funding ends.
- For clinicians: preventive work inside consultations is invisible in health accounts unless it is recorded. Coding it properly helps ministries see how much prevention the system already does.
What to watch next
Watch whether countries that increased prevention spending during the pandemic keep any of the increase in their next budgets. Watch whether any government legislates a protected prevention floor. And watch how prevention is described in health budget speeches: as an investment that will save money, or as a commitment to health worth paying for. The second framing is more honest, and in the long run more durable.
Protecting prevention from the squeeze
If the core problem is that prevention is optional, the fix is to make it less optional. That is a question of budget design more than advocacy.
- Ring-fence with real teeth. A protected line that can only be reduced by explicit legislative decision changes the politics. England’s grant is ring-fenced for public health use, but not protected from being shrunk, which is why the ring fence did not stop the decline.
- Multi-year envelopes. Prevention pays back over a decade. Funding it one year at a time means it is permanently exposed to the next crisis.
- Credit the fiscal side. Earmarking part of health tax revenue for prevention has its critics among economists, and for good reasons, but it at least connects the measure to a stable source.
- Count honestly. Report preventive activity inside primary care and pharmacy, not just programme lines, so the debate is about real effort rather than a residual category.
None of this settles the harder question of how much to shift. Treatment spending protects people who are sick now, and every pound moved to prevention is a pound not spent on someone waiting for a hip. That trade-off is real, and the people waiting are also voters.
But the current system does not make that trade-off deliberately. It makes it by default, every time a budget tightens, in the direction that is quietest. The 2021 spike showed that governments can find money for prevention when the threat is immediate. The collapse afterwards showed what happens once it is not.
If you had to protect one prevention budget line from the next round of cuts in your system, which would it be, and how would you do it?
