Incremental Cost-Effectiveness Ratio (ICER)

The incremental cost-effectiveness ratio (ICER) is the main summary result of a cost-effectiveness analysis and helps analysts compare the additional cost of one healthcare option with the additional health it produces relative to a specified comparator. The sections below explain how the ICER is calculated, how its position on the cost-effectiveness plane affects interpretation, and why dominance, uncertainty and fully incremental analysis must be considered before applying a threshold. The abbreviation ICER is also used by the Institute for Clinical and Economic Review, an independent US non-profit organisation that assesses the value of medicines and other healthcare interventions; this page covers the ratio.

An ICER is a comparative result rather than a property of an intervention. It should always be reported with the underlying incremental costs and effects because the ratio alone can conceal the direction, scale and uncertainty of the differences being compared.

What an ICER measures

An ICER expresses the additional cost required for one additional unit of health outcome when moving from a comparator to an evaluated intervention. Common outcome units include quality-adjusted life-years (QALYs), life-years gained, cases prevented and clinical events avoided.

The numerator and denominator must refer to the same population, perspective, time horizon and comparison. Changing the comparator or reversing the subtraction changes the incremental values and can change the meaning of the result.

• An ICER compares an intervention with a clearly identified comparator.
• An ICER expresses incremental cost per incremental unit of health outcome.
• An ICER does not show the total cost or total health outcome of either option.
• An ICER cannot be interpreted correctly without examining incremental cost and incremental effectiveness separately.

How the ICER is calculated

The ICER divides the difference in expected cost by the difference in expected health outcome. Both differences should use the evaluated intervention minus the comparator so that the direction of the comparison remains consistent.

Let ΔC represent incremental cost and ΔE represent incremental effectiveness. The calculation is:

ΔC = Cost of intervention − Cost of comparator

ΔE = Effect of intervention − Effect of comparator

ICER = ΔC ÷ ΔE

• Incremental cost must be reported in a stated currency and price year.
• Incremental effectiveness must be reported in a stated health-outcome unit.
• The ICER unit combines the two measures, such as £ per QALY gained.
• The same analytical perspective and time horizon must be used for the costs and effects being compared.

A worked two-option example

Suppose a new medicine costs £18,000 per patient and produces 4.5 QALYs, while current care costs £12,000 and produces 4.2 QALYs. The new medicine therefore costs £6,000 more and produces 0.3 additional QALYs.

Dividing the incremental cost by the incremental health gain gives an ICER of £20,000 per QALY gained. If an illustrative threshold of £30,000 per QALY is applied, the new medicine would be considered cost-effective under the stated assumptions, although affordability and other decision considerations remain separate.

ΔC = £18,000 − £12,000 = £6,000

ΔE = 4.5 QALYs − 4.2 QALYs = 0.3 QALYs

ICER = £6,000 ÷ 0.3 QALYs = £20,000 per QALY gained

• The example uses synthetic values and does not describe a real medicine or reimbursement decision.
• The conclusion depends on the stated comparator, analytical assumptions and cost-effectiveness threshold.
• The ICER does not show whether the total financial effect of adopting the medicine is affordable.

Why the cost-effectiveness-plane quadrant matters

The signs of incremental cost and incremental effectiveness determine where a comparison falls on the cost-effectiveness plane. The quadrant must be identified before interpreting the ICER because the same ratio sign can arise from very different combinations of costs and outcomes.

A negative ICER is particularly ambiguous. It can mean that the intervention is less costly and more effective, which is favourable, or more costly and less effective, which is unfavourable.

• More effective and more costly: The ICER can be compared with the relevant cost-effectiveness threshold.
• More effective and less costly: The intervention strictly dominates the comparator, so a threshold comparison is unnecessary.
• Less effective and less costly: The savings must be compared with the health forgone using an appropriate threshold or net-benefit calculation.
• Less effective and more costly: The intervention is strictly dominated by the comparator.

How the threshold informs interpretation

A cost-effectiveness threshold represents the value or opportunity cost associated with an additional unit of health in the relevant decision system. In the northeast quadrant, an ICER below the applicable threshold generally supports the evaluated intervention on cost-effectiveness grounds, while an ICER above the threshold generally supports the comparator.

The threshold must not be treated as a universal constant. Its meaning, value, units, jurisdiction, applicable date and institutional role should be stated whenever an ICER is interpreted.

• An ICER is not inherently favourable or unfavourable without a relevant threshold and decision context.
• A threshold comparison provides a cost-effectiveness conclusion rather than a complete reimbursement decision.
• Affordability, equity, severity, feasibility and implementation may affect the final decision even when an ICER is below the threshold.
• Threshold scenario analysis should be reported transparently rather than used to conceal an unfavourable base-case result.

How to compare more than two alternatives

When several mutually exclusive alternatives are available, separate comparisons against a common baseline can identify the wrong preferred option. Fully incremental analysis instead compares each efficient option with the next relevant alternative after dominated options have been removed.

The process creates a cost-effectiveness frontier and produces sequential ICERs between the remaining alternatives. The preferred option is the most effective alternative whose sequential ICER remains below the relevant threshold, assuming the comparison is otherwise valid.

1. Order the alternatives from least to most effective.
2. Remove any alternative that is dominated by another option, including ties: it costs at least as much, is no more effective, and is worse on at least one of the two.
3. Calculate incremental costs, incremental effects and ICERs sequentially.
4. Remove any alternative subject to extended dominance, meaning its ICER is higher than that of the next more effective alternative.
5. Recalculate the sequential ICERs after every removal.
6. Apply the relevant threshold to the final cost-effectiveness frontier.

Why net benefit is often easier to use

Net-benefit methods express costs and health outcomes on one scale and avoid many of the mathematical difficulties associated with ratios. Incremental net monetary benefit produces a result whose sign can be interpreted consistently across the cost-effectiveness plane.

At threshold λ, incremental net monetary benefit is calculated as the monetary value of incremental health minus incremental cost. A positive result supports the evaluated intervention on cost-effectiveness grounds, while a negative result supports the comparator.

Incremental net monetary benefit = (λ × ΔE) − ΔC

• Net benefit remains interpretable when incremental effectiveness is close to zero.
• Expected net benefit allows several alternatives to be ranked directly.
• Net benefit can be averaged and analysed more reliably in probabilistic analysis than a collection of simulated ICERs.
• Net benefit and a correctly interpreted ICER give equivalent decisions when the same evidence and threshold are used.

How uncertainty affects the ICER

Estimated costs and health outcomes are uncertain because they depend on sampled data, model inputs and assumptions. An ICER reported as a single point estimate does not show how uncertainty could change the preferred option.

Probabilistic sensitivity analysis should therefore examine uncertainty in incremental costs and effects jointly. Cost-effectiveness planes, cost-effectiveness acceptability curves and expected net benefit can communicate different aspects of that uncertainty without treating the mean of simulated ratios as the decision metric.

• A cost-effectiveness plane shows the joint distribution of incremental costs and effects.
• A cost-effectiveness acceptability curve shows the probability that an option is cost-effective across thresholds.
• Expected net benefit identifies the preferred option under the stated threshold and model.
• The probability of being cost-effective should not replace expected net benefit as the decision rule.
• The ratio of mean incremental cost to mean incremental effect is not the same as the mean of simulation-specific ICERs.

Common interpretation errors

ICER errors often arise when analysts focus on the numerical ratio without checking the underlying comparison. These errors can reverse the apparent conclusion or make a result seem more precise and transferable than the evidence allows.

The main safeguard is to report the incremental components, quadrant, comparator and threshold alongside the ratio. Fully incremental analysis and net benefit should be used when a simple pairwise ratio does not provide a stable or complete interpretation.

• Interpreting every negative ICER as favourable confuses dominance with being dominated.
• Dividing by an incremental effect close to zero can produce an extremely large or unstable ICER.
• Comparing every alternative with one common baseline can miss strict or extended dominance.
• Averaging ICERs across probabilistic simulations produces a misleading summary.
• Reporting an ICER without its comparator makes the result incomplete.
• Treating an ICER below a threshold as proof of affordability confuses efficiency with budget impact.
• Reporting excessive decimal places can imply unsupported precision.

What should be reported with an ICER

A transparent report should allow readers to reconstruct the calculation and understand how the result informs the decision. Reporting only the final ratio is insufficient because it hides the magnitude and direction of the underlying cost and health differences.

The information below should accompany the ICER so that the comparison is auditable and reusable. Jurisdiction-specific thresholds and guidance should include the applicable source, version and date.

• The evaluated intervention and its comparator.
• The population, perspective, time horizon and outcome measure.
• Expected costs and health outcomes for every relevant alternative.
• Incremental costs and incremental effects, shown separately.
• The ICER with its currency, price year and health-outcome unit.
• The cost-effectiveness-plane quadrant and any dominance findings.
• The threshold, jurisdiction, institutional context and source.
• The methods and results used to characterise uncertainty.
• Fully incremental comparisons when more than two alternatives are relevant.
• Budget impact and wider decision considerations, kept separate from the ICER.

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.
• National Institute for Health and Care Excellence. NICE technology appraisal and highly specialised technologies guidance: the manual (PMG36). Published 2022; updated 2026.
• 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.
• Claxton K. The irrelevance of inference: a decision-making approach to the stochastic evaluation of health care technologies. Journal of Health Economics. 1999;18(3):341-364.
• Fenwick E, Claxton K, Sculpher M. Representing uncertainty: the role of cost-effectiveness acceptability curves. Health Economics. 2001;10(8):779-787.