The Economics of Prevention: Cost-Saving Is Not the Same as Cost-Effective

Prevention has an intuitive economic appeal: avoid illness now and avoid treatment costs later. Sometimes that is exactly what happens. A low-cost intervention aimed at a high-probability, expensive outcome can improve health and reduce total spending.

But preventive care is not one intervention, and preventing illness is not free. Screening, counseling, vaccination programs, environmental changes, medication, outreach, follow-up tests, administration, and treatment of findings all use resources. Some programs save money. Many produce health at an added cost. Others offer little health benefit or cause enough harm that they are poor value at any price.

The question you should be asking is therefore not whether prevention "pays for itself" as a category. It is whether a specified intervention, for a specified population, improves outcomes enough to justify its full costs and tradeoffs compared with the best alternative.

Effectiveness comes before economics#

An intervention cannot become good value merely by being inexpensive. It must first have a credible effect on an outcome that matters. A cheap program with no meaningful benefit consumes resources that could be used elsewhere. A modestly costly program can be valuable if it prevents substantial morbidity or mortality.

The causal evidence should match the economic model. If a randomized trial shows an effect only on a laboratory marker, translating that marker into lifetime health requires additional assumptions. If observational evidence is confounded, a model can reproduce the bias with impressive precision. Effect size, durability, harms, adherence, and applicability all need uncertainty ranges.

The comparator matters too. "Do nothing" may be unrealistic when current practice already includes part of the program, and the relevant comparison might be organized outreach versus opportunistic care, a new screening interval versus the existing interval, or a targeted approach versus universal delivery.

Four questions that are often confused#

Cost-saving means that the intervention produces better health outcomes and lowers total costs counted by the analysis. The expense of delivery is more than offset by costs avoided. This is sometimes called a dominant result. True cost-saving prevention is attractive, but it is not the only prevention worth doing.

Cost-effectiveness compares the difference in cost with the difference in health between options. A program may add cost and add healthy life. Decision-makers then judge whether the incremental health is worth the incremental resources under their rules and constraints. Curative treatments and chronic care are assessed the same way; prevention should not face a special requirement to save money.

Budget impact asks what the program will do to actual spending over a defined period; a nationwide program can be cost-effective yet require a large near-term budget because millions are eligible. Conversely, a small program may fit the budget but offer poor value per unit of health.

Return on investment asks whether a named stakeholder receives financial returns relative to its spending: an employer may count reduced absence, a health plan may count medical claims, and a government may count tax revenue or service use. A favorable social return does not guarantee a positive cash return to the organization paying the bill. These four measures answer different questions, and substituting one for another can make a policy look more certain to you than it is.

Not all prevention saves money#

Cohen, Neumann, and Weinstein reviewed the economic literature and argued that statements about prevention saving money should be held to the same standards as statements about treatment. Their central point remains useful: some preventive measures save money, but many improve health at a net cost, and their value varies widely.

Screening illustrates the reason. A program spends money on everyone invited or tested, while only a fraction can benefit, and it can also generate false positives, follow-up procedures, overdiagnosis, and treatment of findings that would never have caused harm. Screening can still be highly valuable when evidence shows a favorable balance, but earlier detection does not guarantee cost savings.

Preventive medicines can have a similar structure. Many people may need years of treatment to avert events in a smaller subset. Benefits can be substantial, especially at higher baseline risk, while drug, monitoring, adverse-event, and adherence costs remain. The economic conclusion depends on the actual numbers.

Public-health measures may spread benefits across entire populations and sectors. Cleaner air, safer roads, or tobacco control can affect health, productivity, and public spending. Their analysis must capture relevant consequences without assuming every benefit becomes cash that can be removed from next year's health budget.

Baseline risk changes value#

Absolute benefit rises when baseline risk is higher, assuming a comparable relative effect. If an intervention reduces relative risk by 20 percent, it prevents far more events when untreated risk is 20 percent than when it is 1 percent; the delivery cost per person may be similar, so the cost per event prevented can differ sharply.

Targeting can improve value by focusing resources on people more likely to benefit. Yet targeting itself has costs. Risk assessment, data systems, outreach, and confirmation can consume resources and introduce errors. A risk rule may also miss people, perform unequally across populations, or intensify barriers for those with incomplete records. Universal programs can be simpler and more equitable in access, while targeted programs can concentrate benefit, so the appropriate choice depends on the risk distribution, the accuracy and fairness of targeting, program capacity, and the consequences of missed eligibility.

Perspective decides which costs count#

An analysis from a health-system perspective may count medical costs borne by the system, a payer may count reimbursed services, and a government perspective can include several public budgets. A patient perspective includes travel, time, copayments, caregiving, and lost income. A societal perspective aims to capture consequences regardless of who bears them.

The Second Panel on Cost-Effectiveness in Health and Medicine recommended transparent reporting of both health-care-sector and societal perspectives, with an impact inventory showing what was included. The principle is broader than any one jurisdiction: you need to know whose costs and benefits appear in the calculation.

Perspective can reverse the apparent result. A clinic may pay for outreach while a hospital avoids admissions. A health plan may invest now while savings arrive after members move to another insurer. You may bear time and transport costs that do not appear in claims. A program is not costless because a particular budget does not record the burden.

The wrong-pocket problem#

The wrong-pocket problem occurs when the organization able to fund prevention is not the organization that receives later financial benefit. Savings can cross budgets, sectors, or time periods. A school, housing authority, employer, public-health department, and health insurer may all affect outcomes while accounting separately.

This is an incentive and governance problem, not proof that the intervention lacks social value. Possible responses include shared financing, bundled budgets, longer contracts, or explicit recognition of cross-sector benefits. Each needs evidence and accountability; labeling a program "prevention" does not resolve who pays or whether promised savings materialize.

The problem also works in reverse. A stakeholder may report a positive return because costs were shifted to patients or another public service. A complete analysis follows both benefits and burdens across pockets.

Time horizon and discounting#

Costs often occur early while benefits arrive later. An analysis needs a horizon long enough to include meaningful differences between options. A one-year model can make a childhood intervention appear all cost and no benefit. A lifetime model may be appropriate for long-term consequences, but it relies on more assumptions about durability, future care, and competing risks.

Discounting gives less present value to costs and health outcomes occurring farther in the future, according to the reference rules used by the decision-maker. It reflects time preference and opportunity cost and supports consistent comparison across programs. Because prevention often has delayed benefits, the discount rate can materially affect results.

Discounting is not inherently hostile to prevention, and simply omitting it can make comparisons inconsistent. Analysts should disclose the rates, follow the relevant reference case, and test plausible alternatives. Results that change dramatically under modest assumptions deserve a more cautious conclusion.

Opportunity cost#

Resources used for one program cannot produce health somewhere else. That lost alternative is the opportunity cost. It may involve another prevention program, treatment, staffing, mental-health service, or basic care infrastructure.

Cost-effectiveness thresholds attempt to represent, with varying methods and limitations, how much health could be produced or displaced by spending. They should not be treated as universal scientific constants. Jurisdictions differ in budgets, prices, disease burden, objectives, and decision rules.

Opportunity cost is why a health-improving program can still be a poor allocation. If the same resources could produce much more health through another feasible option, choosing the lower-value program has consequences even if its own participants benefit.

Implementation costs belong in the model#

An intervention described in a trial may require training, information systems, space, equipment, quality control, administration, outreach, reminders, translation, transportation support, and management. Initial setup and ongoing maintenance are different costs. Programs also need capacity to confirm results and provide the downstream care that detection creates.

Uptake may be lower than in a trial. Unequal access can concentrate participation among people already well served, staff time can displace other work, and false-positive findings and adverse effects can add cost and harm. Scale can reduce some unit costs while revealing bottlenecks elsewhere.

Economic evaluations should therefore model implementation as part of the intervention, not as a free wrapper around it. Scenario analyses can compare optimistic, expected, and constrained delivery. A sensitivity analysis should show what happens when effect size, uptake, duration, unit cost, downstream events, and time horizon vary.

Cost-effectiveness and equity answer different questions#

An average ratio can conceal who benefits. A program may be efficient overall while widening disparities if access favors groups with fewer barriers. Another may look less efficient because it spends more to reach people facing structural obstacles, even though equity is an explicit policy objective.

Equity analysis can examine outcomes and burdens by income, geography, disability, race and ethnicity, language, sex, age, or other relevant dimensions. It can test whether delivery rules exclude groups, whether costs fall on those least able to bear them, and whether reducing a gap is valued beyond the average health total.

This does not require hiding tradeoffs. Decision-makers should report when equity goals use additional resources or when an apparently efficient design has unequal consequences. Transparent plural criteria are more honest than pretending one ratio contains every social value.

Recommendation grades are not economic verdicts#

The US Preventive Services Task Force states that its recommendations are based on the balance of benefits and harms, not on cost-effectiveness. A favorable grade does not establish that an intervention saves money or fits every budget. An insufficient-evidence statement is not a finding that an intervention lacks economic value. Economic evaluation can inform implementation once evidence of net benefit is considered, but it answers a separate question, and keeping the two functions distinct is what stops a clinical-evidence grade from being repurposed as a spending analysis.

How to read a prevention economic claim#

Start with the exact intervention, population, and comparator. Check whether health effects are measured or modeled, whether important harms appear, and whether the effect persists long enough to support projected outcomes. Identify the perspective, currency year, setting, time horizon, and discount rates.

Then ask whether the headline refers to cost-saving, cost-effectiveness, budget impact, or return on investment. Look for implementation costs, baseline-risk assumptions, uptake, and sensitivity analysis. Check who pays and who benefits. Finally, examine distribution rather than relying on an average alone.

Prevention does not need to save money to be worthwhile. It needs credible benefit, a transparent account of cost and harm, and a fair comparison with what those resources could otherwise achieve.

References#

1. Cohen, Neumann, and Weinstein, Does Preventive Care Save Money?
2. Second Panel on Cost-Effectiveness in Health and Medicine
3. NICE economic evaluation methods
4. WHO cost-effectiveness portal
5. WHO health economics overview
6. USPSTF and cost considerations
7. The wrong-pocket problem in preventive health investment