Cost-Effectiveness Analysis

This page explains when cost-effectiveness analysis is appropriate, which choices define the analysis, and why alternatives must use the same outcome measure. It then shows how incremental costs and outcomes are calculated, how the cost-effectiveness plane and frontier guide interpretation, and why cost-effectiveness does not answer questions about affordability on its own.

Cost-effectiveness analysis, commonly shortened to CEA, is a form of full economic evaluation. It is useful when competing healthcare alternatives can be compared using the same natural measure of outcome, such as life-years gained, infections prevented, hospital admissions avoided, or symptom-free days.

When cost-effectiveness analysis is the appropriate method

The decision question and outcome measure determine whether CEA is the most appropriate analytical method. CEA is particularly useful when decision-makers want to compare the additional cost required to produce one additional unit of a shared health outcome. The result remains specific to the population, alternatives, perspective, time horizon, evidence, and decision context used in the analysis.

Cost-utility analysis is often treated as a specialised form of CEA because it uses preference-weighted health outcomes, most commonly quality-adjusted life years. Cost-benefit analysis differs because it expresses consequences in monetary terms. These methods should not be treated as interchangeable simply because each compares costs and outcomes.

Which choices define a cost-effectiveness analysis

A CEA is shaped by several choices made before the calculations begin. These choices determine which information enters the analysis, which alternatives are compared, and how the result should be interpreted. They should therefore be stated clearly rather than hidden inside a model or technical appendix.

1. Define the healthcare decision and population so that the analysis answers a specific question.
2. Identify the relevant alternatives and comparators so that the analysis represents the choice actually facing the decision-maker.
3. Select the analytical perspective so that the included costs and consequences reflect the intended viewpoint.
4. Choose a sufficient time horizon so that all material differences between the alternatives are captured.
5. Select one shared natural outcome measure that is relevant to every alternative.
6. Estimate resource use, costs, and outcomes consistently using evidence appropriate to the population and decision.
7. Compare the alternatives incrementally rather than relying on separate average ratios.
8. Examine uncertainty and population differences that could change the result.
9. Interpret the findings within the wider decision context rather than treating the calculation as an automatic decision.
10. Report the methods, evidence, assumptions, limitations, and results transparently so that others can scrutinise the analysis.

These choices are connected. For example, changing the comparator changes the incremental calculation, while changing the perspective or time horizon can change which costs and outcomes are counted.

Choosing an outcome that all alternatives share

CEA requires every alternative to use the same natural outcome measure. A shared denominator allows incremental costs to be interpreted in relation to a comparable change in health. The outcome should be clinically meaningful and appropriate to the decision being studied.

• A vaccination CEA might report the cost per infection prevented.
• A screening CEA might report the cost per case detected.
• A hospital programme CEA might report the cost per admission avoided.
• A treatment CEA might report the cost per life-year gained.
• A symptom-management CEA might report the cost per symptom-free day.

Ratios using different outcome measures cannot be compared directly. A cost per hospital admission avoided and a cost per life-year gained answer different questions, even though both may be described as cost-effectiveness ratios.

How resource use becomes cost

CEA separates the quantity of each resource used from the value assigned to it. Resources may include staff time, medicines, tests, hospital stays, equipment, facilities, patient time, travel, or productivity effects, depending on the analytical perspective. Keeping resource quantities and unit costs separate makes the calculations easier to audit and update.

The cost of each resource is generally calculated as:

Resource cost = Quantity of resource used × Unit cost

The total cost for an alternative is calculated by adding the costs of all relevant resources:

Total cost = Σ(Quantity × Unit cost)

Resource use should be measured consistently across alternatives. Analysts should state the source of each quantity and unit cost, the price year, currency, valuation method, and any adjustments for inflation or differences between settings. Costs and outcomes occurring in future years should be discounted at stated rates.

Common errors include omitting important resource consequences, double-counting the same resource, combining quantities and unit costs from incompatible settings, and treating charges or prices as economic costs without explaining the choice.

Calculating the differences between alternatives

CEA compares the change in cost and outcome produced by moving from one alternative to another. These incremental differences are more relevant to a decision than the average cost or outcome of either alternative by itself. The relevant comparator must be identified before the calculation is performed.

For a new intervention compared with an existing alternative:

Incremental cost

ΔC = Cnew − Ccomparator

Incremental effect

ΔE = Enew − Ecomparator

When ratio interpretation is appropriate:

ICER = ΔC ÷ ΔE

The incremental cost-effectiveness ratio, or ICER, reports the additional cost per additional unit of outcome. An ICER should not be interpreted until the signs of incremental cost and incremental effect have been examined.

At a decision threshold λ, incremental net monetary benefit can be calculated as:

INMB = (λ × ΔE) − ΔC

A positive incremental net monetary benefit favours the new intervention on expected value-for-money grounds at the stated threshold. It does not establish that the intervention is affordable or require a decision-maker to adopt it.

Read the cost-effectiveness plane before interpreting the ICER

The cost-effectiveness plane, with incremental effect on the horizontal axis and incremental cost on the vertical axis, shows whether the new intervention costs more or less and produces more or less health than its comparator. Locating the result on the plane prevents a negative ICER from being treated automatically as favourable. The signs of both incremental cost and incremental effect determine what the ratio means.

Lower cost and greater effect

The new intervention costs less and produces more health than its comparator. The new intervention therefore dominates the comparator.

Higher cost and lower effect

The new intervention costs more and produces less health than its comparator. The new intervention is therefore dominated by the comparator.

Higher cost and greater effect

The new intervention produces additional health but also has an additional cost. This trade-off requires a decision threshold or another explicit decision rule.

Lower cost and lower effect

The new intervention saves money but also produces less health. This trade-off requires a decision-maker to assess whether the cost saving justifies the health loss.

The northeast and southwest positions involve trade-offs that require a decision threshold or another explicit decision rule. The southeast and northwest positions establish dominance or being dominated without requiring an ICER threshold comparison.

A negative ICER is ambiguous because it can occur when an intervention is both less costly and more effective or when it is more costly and less effective. When incremental effect equals zero, an ICER should not be calculated. The cost difference and absence of an outcome difference should be reported directly.

Comparing more than two alternatives

A decision involving three or more mutually exclusive alternatives requires fully incremental analysis. Comparing every intervention only with current care can overlook a more efficient adjacent alternative and may leave a dominated strategy in the analysis. The alternatives should therefore be ordered and compared sequentially.

1. Order the alternatives by increasing effectiveness.
2. Remove any dominated alternative, meaning one that costs at least as much as another alternative and is no more effective, while being worse on at least one of the two.
3. Calculate sequential incremental costs, outcomes, and ICERs.
4. Identify any alternative subject to extended dominance.
5. Remove extendedly dominated alternatives and recalculate the sequential results, repeating until the ICERs increase with effectiveness.
6. Interpret the remaining alternatives on the cost-effectiveness frontier.

An alternative is subject to extended dominance when a combination of other strategies can provide greater effectiveness at a lower cost. In practice, this is shown when its ICER is higher than that of the next more effective alternative. Removing it changes the relevant comparisons, so the ICERs must be recalculated rather than copied from the original analysis.

Worked example: cost per life-year gained

Consider a comparison between current care and Intervention A. Both alternatives are measured using life-years, so their outcomes are measured in a common unit. The figures are illustrative and do not represent a real technology or decision.

• Current care: £10,000 per patient and an expected outcome of 8.0 life-years.
• Intervention A: £12,500 per patient and an expected outcome of 8.5 life-years.

Incremental cost = £12,500 − £10,000 = £2,500

Incremental effect = 8.5 − 8.0 = 0.5 life-years

ICER = £2,500 ÷ 0.5 = £5,000 per life-year gained

Intervention A produces additional health at an additional cost. Whether £5,000 per life-year gained represents good value depends on the applicable decision rule, the uncertainty surrounding the estimates, and the wider decision context.

How assumptions and uncertainty affect the conclusion

Costs, outcomes, evidence, and model assumptions are rarely known with certainty. Important assumptions may concern treatment effects, resource use, unit costs, adherence, disease progression, extrapolation beyond observed data, the analytical perspective, the time horizon, and the population expected to receive the intervention. The type of uncertainty being examined should be stated clearly.

• Deterministic sensitivity analysis changes selected inputs or assumptions individually or in defined scenarios.
• Probabilistic sensitivity analysis examines the combined effect of uncertainty in multiple model parameters.
• Scenario analysis tests alternative structural choices, evidence sources, or implementation assumptions.
• Subgroup analysis examines heterogeneity, that is, whether costs and outcomes differ across relevant population groups.

Probabilistic results may be summarised using cost-effectiveness-plane simulations, cost-effectiveness acceptability curves, expected net benefit, or the probability that each option is cost-effective. The probability of cost-effectiveness describes uncertainty about which option performs best, while expected net benefit identifies the option with the greatest expected value at the stated threshold.

A favourable expected result may still be accompanied by substantial uncertainty. Analysts should explain which assumptions most strongly affect the conclusion and whether plausible alternatives change the preferred option.

How CEA relates to budget impact analysis

CEA examines whether the additional outcomes of an intervention justify its additional cost under a stated decision rule. Budget impact analysis examines how adopting the intervention could change total expenditure for a defined budget holder over a specified planning period. The two analyses answer related but different questions.

A cost-effective intervention can still be unaffordable when the eligible population is large, uptake is rapid, per-person expenditure is high, or the available budget is constrained. Budget impact analysis therefore complements CEA by adding information about population size, uptake, expenditure timing, and financial feasibility.

What CEA can and cannot decide

CEA provides evidence about efficiency and comparative value. It does not incorporate every factor relevant to healthcare coverage, reimbursement, or implementation. Decision-makers must combine CEA results with the other evidence required by their institutional process.

• CEA does not determine whether an intervention is affordable within the available budget.
• Standard CEA does not determine how health gains and costs are distributed across population groups.
• CEA does not replace an assessment of clinical effectiveness, safety, or evidence quality.
• CEA does not establish whether implementation is organisationally feasible.
• CEA does not make a reimbursement, coverage, or adoption decision.
• CEA does not make a result from one jurisdiction automatically transferable to another.

Health technology assessment may use CEA alongside clinical, budgetary, ethical, organisational, equity, and implementation evidence. A favourable economic result informs that wider process but does not replace it.

Common mistakes and how to avoid them

A calculation can appear precise while still answering the wrong question. Common errors usually involve the comparator, outcome measure, strategy set, resource inputs, assumptions, or interpretation of an ICER. The following checks help prevent those errors.

• Reporting average rather than incremental ratios obscures the change associated with moving between alternatives.
• Comparing every option only with current care can overlook the relevant cost-effectiveness frontier.
• Failing to remove strict or extended dominance can produce misleading sequential ICERs.
• Treating every negative ICER as favourable ignores the difference between dominance and being dominated.
• Calculating an ICER when incremental effect equals zero produces an undefined or misleading ratio.
• Mixing incompatible outcome units creates ratios that cannot be compared directly.
• Using an irrelevant comparator produces results that do not represent the actual healthcare decision.
• Omitting or double-counting resource use distorts the estimated cost difference between alternatives.
• Combining costs or evidence from incompatible settings can make the results inappropriate for the intended decision.
• Ignoring the cost-effectiveness-plane quadrant encourages mechanical interpretation of the ICER.
• Treating probability of cost-effectiveness as expected value confuses uncertainty with the expected decision.
• Calling an intervention cost-effective without naming the threshold and context presents a conditional result as a universal conclusion.

CEA should therefore be reported as an evidence-based comparison within a defined decision problem. Its conclusions must remain tied to the alternatives, population, evidence, assumptions, outcome measure, and institutional setting used in the analysis.

Sources

• Drummond MF, Sculpher MJ, Claxton K, Stoddart GL, Torrance GW. Methods for the Economic Evaluation of Health Care Programmes, 4th edition. Oxford University Press. 2015.
• Black WC. The CE plane: a graphic representation of cost-effectiveness. Medical Decision Making. 1990;10(3):212-214.
• Cantor SB. Cost-effectiveness analysis, extended dominance, and ethics: a quantitative assessment. Medical Decision Making. 1994;14(3):259-265.
• Stinnett AA, Mullahy J. Net health benefits: a new framework for the analysis of uncertainty in cost-effectiveness analysis. Medical Decision Making. 1998;18(2 Suppl):S68-S80.
• Fenwick E, Claxton K, Sculpher M. Representing uncertainty: the role of cost-effectiveness acceptability curves. Health Economics. 2001;10(8):779-787.
• Briggs AH, Weinstein MC, Fenwick EAL, Karnon J, Sculpher MJ, Paltiel AD. Model parameter estimation and uncertainty analysis: a report of the ISPOR-SMDM Modeling Good Research Practices Task Force Working Group-6. Medical Decision Making. 2012;32(5):722-732.
• Sullivan SD, Mauskopf JA, Augustovski F, Jaime Caro J, Lee KM, Minchin M, et al. Budget impact analysis: principles of good practice. Report of the ISPOR 2012 Budget Impact Analysis Good Practice II Task Force. Value in Health. 2014;17(1):5-14.