From Budget to Value: How Health Economic Models Connect

One treatment. More than one question.

Imagine a health plan considering a new treatment. The team asks, "How much more will we spend next year?"

Then someone asks, "What extra health benefit will we get for that money?"

Both questions matter. But they need different analyses.

A budget impact analysis, or BIA, estimates how spending changes when a new option is added. A cost-effectiveness analysis, or CEA, compares the extra cost with the extra health benefit.

One result from a CEA may be an incremental cost-effectiveness ratio, called an ICER. The name is long, but the idea is simple: How much extra do we pay for each extra unit of health gained?

An ICER is a result, not a separate model. A BIA does not automatically turn into an ICER. Still, the two analyses can share useful building blocks.

How did these tools come about?

The story begins with a practical problem. Healthcare has many needs, but money, staff, and time are limited. Decision-makers need ways to compare what different choices cost and what they may achieve.

The methods grew over decades. They did not appear in one neat order, and one did not replace the others.

An early healthcare example was a 1968 study of treatments for chronic kidney disease. It compared treatment costs and outcomes. In 1977, Weinstein and Stason published an influential paper on the foundations of cost-effectiveness analysis in health and medicine.

Researchers also wanted to consider both length and quality of life. Work during the 1960s and 1970s helped shape the quality-adjusted life year, or QALY. This gave analysts a way to combine those two parts of health in one measure.

At the same time, budget holders still needed a practical answer about spending. A treatment could offer good value over many years but require a large payment soon. ISPOR published good-practice guidance for BIA in 2007 and updated it in 2014. These were milestones in setting standards, not the invention of budgeting.

The field grew because each new question needed a suitable way to answer it.

Start with the budget question

Let us use a made-up example. A health plan has 1,000 eligible patients. Assume 200 switch to a new treatment at the start of the year and stay on it for the full year.

For this simple example, the old treatment costs $10,000 per patient per year. The new one costs $13,000. These amounts include all care costs counted in this example, and we assume no other spending changes.

• Before the new treatment: 1,000 patients x $10,000 = $10 million.
• After it is added: 800 patients x $10,000, plus 200 patients x $13,000 = $10.6 million.
• Budget impact: $10.6 million minus $10 million = $600,000 more that year.

That result answers a spending question. It does not yet tell us whether the extra health benefit is worth the cost.

A real BIA would need more detail. The number of eligible patients may change. People may start or stop treatment during the year. The new option may also change hospital use or other care costs. ISPOR's BIA guidance explains why the population, treatment mix, uptake, and costs must fit the budget holder's setting.

Now ask about health value

Suppose we want to know whether the new treatment offers enough added health for its extra cost.

Some work from the BIA may help. We may reuse medicine prices, dosing, visit costs, and estimates of hospital use. But we must check that they fit the new analysis.

Now we also need health outcomes. How does each option affect symptoms, survival, or quality of life? How long do those effects last?

For our example, assume we have sound evidence for average one-year outcomes of 0.70 QALYs with the old option and 0.80 with the new one. These are invented values for teaching.

A QALY combines time lived with a weight for health-related quality of life. One year at a weight of 1 equals one QALY. One year at 0.8 equals 0.8 QALYs. The weights describe health states, not a person's worth.

Using the same one-year costs:

• Extra cost per patient: $13,000 minus $10,000 = $3,000.
• Extra health per patient: 0.80 minus 0.70 = 0.10 QALY.
• ICER: $3,000 divided by 0.10 = $30,000 per QALY gained.

Because we used QALYs, this is a cost-utility analysis, a form of cost-effectiveness analysis. A CEA can also use other health measures, such as life-years gained.

We have now answered two questions with shared cost inputs: the plan spends $600,000 more that year, and the added cost is $30,000 for each added QALY in this example.

The ICER is not a bill for each patient. It is also not an automatic "yes" or "no." Its meaning depends on the decision-maker's criteria, the alternatives, and the uncertainty in the evidence.

What carries over, and what must change?

A shared input is a starting point, not a reason to copy the whole model.

Costs may carry over. Treatment prices and costs per visit can be useful in both analyses. Check the setting, year, and what is included.

Clinical inputs may carry over. Rates of side effects or hospital use may help both analyses if they apply to the same people and treatments.

The population question changes. A BIA needs the number of people likely to receive each option. A CEA often compares expected costs and outcomes per person for the relevant group.

The time period may change. A BIA focuses on the budget's planning period. A CEA may need many years, or a lifetime, to capture important differences. Our one-year example is only suitable if it captures the differences that matter. A longer CEA would need new cost and outcome estimates; we could not simply carry over the $3,000 difference.

The viewpoint also matters: whose costs count? A health plan's budget differs from a wider view that includes patient travel or time away from work. Rules for valuing future costs and benefits must fit the analysis too.

The work can flow the other way. A CEA may estimate future care costs that help build a BIA. The BIA then adds local patient numbers and expected uptake. Neither must always come first.

Other ways to ask the question

The main economic approaches serve different purposes. NICE describes several types. Here is a simple guide.

• Cost-of-illness analysis: What does a condition cost? It describes the burden, such as hospital care and medicines. It may supply cost inputs for other work, but it does not show which treatment is best.
• Cost-minimization analysis: If sound evidence shows the relevant outcomes are equivalent, which option costs less? A study finding no statistically significant difference does not, by itself, prove equivalence.
• Cost-consequence analysis: What are the separate costs and outcomes? It may show spending, hospital stays, symptoms, and patient experience side by side rather than combine them into one number.
• Cost-benefit analysis: What happens when both costs and benefits are expressed in money? This can help compare different kinds of programs, but placing a money value on health needs careful methods and clear assumptions.
• Cost-effectiveness analysis: What is the extra cost for the extra health outcome?
• Cost-utility analysis: What is the extra cost for a gain measured using both length and quality of life, commonly QALYs?
• Budget impact analysis: How will adopting the option change spending for this budget?

These approaches grew to address different needs: describing burden, comparing alternatives, valuing health, and planning spending. They are related tools, not steps that every project must complete.

Where do Markov models fit?

There is another set of names people often hear: decision trees, Markov models, and simulations. These describe how the calculations are built, rather than the economic question being asked.

A decision tree follows a set of branches, such as recovery or a complication after a procedure.

A Markov model follows movement between health states over repeated periods. For example, it may track people whose disease is stable, whose disease has worsened, or who have died.

A simulation can track individual paths or events over time. A disease-transmission model can also account for how an infection spreads between people.

These tools developed because simple calculations could not always capture how care and disease change over time. ISPOR and the Society for Medical Decision Making published a set of modeling good-practice reports in 2012.

The same model structure can support different economic questions. A Markov model, for example, may estimate costs and QALYs for a CEA. A suitably adapted version may help a BIA. The purpose, population, and time period still need to be checked.

Test what could change the answer

What if 400 patients adopt the new treatment rather than 200? Under our simple assumptions, the added budget cost doubles to $1.2 million. The per-patient ICER stays the same because we have not changed per-patient costs or benefits.

What if the QALY gain is only 0.05 rather than 0.10? With the same extra cost, the ICER rises to $60,000 per QALY gained.

These checks show why a single result is not enough. Scenario and sensitivity analyses explore how uncertain inputs or different assumptions affect the answer. They help decision-makers see where more evidence may matter.

Bring the pieces together

At Master Table, we begin with the decision the analysis needs to support. Then we identify the evidence, costs, outcomes, and model structure that fit.

Our work includes budget impact, cost-effectiveness, and other health economic analyses. Shared inputs can save work and improve consistency, but each analysis needs its own checks.

The aim is to explain what a choice may cost, what it may achieve, and how sure we can be of the answer.

To discuss a health economic modeling project, contact hello@mastertable.com.

References

1. Mauskopf JA, et al. (2007). Principles of Good Practice for Budget Impact Analysis. Value in Health, 10(5), 336-347.
2. NICE. Economic evaluation. Technology appraisal and highly specialised technologies guidance: the manual.
3. Klarman HE, Francis JO, Rosenthal GD. (1968). Cost Effectiveness Analysis Applied to the Treatment of Chronic Renal Disease. Medical Care, 6(1), 48-54.
4. Weinstein MC, Stason WB. (1977). Foundations of Cost-Effectiveness Analysis for Health and Medical Practices. New England Journal of Medicine, 296, 716-721.
5. Quantifying life: Understanding the history of Quality-Adjusted Life-Years (QALYs). (2018). Social Science & Medicine.
6. Sullivan SD, et al. (2014). Budget Impact Analysis — Principles of Good Practice: Report of the ISPOR 2012 Budget Impact Analysis Good Practice II Task Force. Value in Health, 17(1), 5-14.
7. NICE. Glossary: economic evaluation. ISPOR. Economic Evaluation topics.
8. Caro JJ, Briggs AH, Siebert U, et al. (2012). Modeling Good Research Practices — Overview. Value in Health, 15(5), 796-803.
9. Briggs AH, et al. (2012). Model Parameter Estimation and Uncertainty Analysis. ISPOR-SMDM Modeling Good Research Practices Task Force, Report 6.

Sources reviewed September 26, 2026. All example prices, population counts, and health outcomes are invented for teaching. They are not estimates for a real treatment or client. Historical milestones are selected examples, not claims about the first use of every method. Local evaluation requirements may differ.