Incremental Cost-Utility Ratio

Incremental Cost-Utility Ratio with a Worked Example

The incremental cost-utility ratio, or ICUR, is the name several literatures give to an incremental cost-effectiveness ratio whose health effect is measured in QALYs or another generic measure of health-related utility. It is calculated as $\Delta C / \Delta Q$, the difference in expected cost divided by the difference in expected QALYs, and is read in the same way as a cost per QALY gained. It is the result that most cost-utility analyses report. New Zealand reports an inverse form of the ratio, and any ICUR is read against a threshold. The errors that most often distort an ICUR are negative ratios, near-zero QALY differences, the wrong comparator and averaging ratios across simulations. General ICER mechanics, league tables and threshold policy are covered on the linked pages.

The formula as expected incremental cost over expected incremental QALYs

The ICUR compares two options, an intervention and a comparator, over the same population and time horizon. Its numerator is the incremental cost and its denominator the difference in expected QALYs. Formally, it is written as follows.

$$\text{ICUR} = \frac{\bar{C}_1 - \bar{C}_0}{\bar{Q}_1 - \bar{Q}_0} = \frac{\Delta C}{\Delta Q}$$

where $\bar{C}_1$ and $\bar{C}_0$ are the expected total costs per patient of the intervention and the comparator, $\bar{Q}_1$ and $\bar{Q}_0$ are their expected QALYs per patient, $\Delta C$ is the incremental cost and $\Delta Q$ is the incremental QALY gain. The result is expressed in currency per QALY gained. The NICE manual (PMG36) defines the ratio as expected additional total cost over expected additional QALYs, which fixes the order of operations. Costs and QALYs are averaged first and the ratio is taken last.

New Zealand's Pharmaceutical Management Agency (PHARMAC) prefers to report the inverse, which it calls the incremental utility-cost ratio (IUCR), as QALYs gained per NZD 1 million of net expenditure. The inverse is written as follows.

$$\text{IUCR} = \frac{\Delta Q}{\Delta C} \times 1{,}000{,}000$$

where $\Delta Q$ and $\Delta C$ are defined as above and the multiplier expresses the result per one million units of currency. The PHARMAC manual describes ICURs as in effect the inverse of IUCR results and asks for both to be reported. For a gain in QALYs at extra cost against the same comparator, a higher IUCR and a lower ICUR both indicate better value, so the direction of a reported figure has to be checked before comparison.

One ratio under several names

In 1977 Weinstein and Stason set out the logic that, in cost-effectiveness analysis, the ratio of net health-care costs to net health benefits gives an index for setting priorities, and quality-of-life effects can be built into the benefit measure as adjustments to life expectancy. When the benefit measure is the QALY, the analysis is a cost-utility analysis and its incremental ratio is what some authors call the ICUR. The Canadian National Advisory Committee on Immunization (NACI) defines cost-utility analysis as valuing health outcomes in QALYs or another generic measure of health-related utility, and describes its summary measure as an ICER, sometimes called an incremental cost-utility ratio.

Usage differs by setting rather than by meaning. The PHARMAC Prescription for Pharmacoeconomic Analysis (version 2.2, 2015) refers to "ICURs or ICERs, the incremental cost per QALY" and calls this the unit typically reported for most cost-utility analyses internationally. A review of cost-utility studies in trauma and orthopaedic surgery by Kamaraj and colleagues noted that many studies use the two terms interchangeably. The NICE manual (PMG36) asks for results as ICERs and describes the analysis itself as incremental cost-utility analysis. When health is measured in QALYs, an ICUR, an ICER and a cost per QALY gained figure are the same number under different labels.

Choices behind the numerator and the denominator

The arithmetic is simple, so the value of an ICUR depends almost entirely on the inputs. The comparator, time horizon, costing perspective, discount rate and source of health state utility values all change $\Delta C$ or $\Delta Q$, and a small change in $\Delta Q$ can move the ratio a long way. Agencies fix many of these choices in a reference case; the NICE manual (PMG36), for example, applies the same annual discount rate, currently 3.5%, to costs and health effects.

Kamaraj and colleagues, adapting EUnetHTA guidance, state that the comparator should ideally be the reference treatment in up-to-date, high-quality clinical guidelines, and warn readers to watch for inappropriate, non-cost-effective comparators. Both arms must also be costed and valued on the same basis, since an ICUR built from arms costed in different price years, or valued with different utility tariffs, mixes method differences into the result.

Worked example of a new treatment against usual care

The figures are illustrative. A new treatment for a chronic condition has an expected cost of GBP 24,600 and yields 6.42 discounted QALYs per patient. Usual care costs GBP 16,200 and yields 6.02 QALYs.

1. Incremental cost. 24,600 − 16,200 = 8,400, so the new treatment costs GBP 8,400 more per patient.

2. Incremental QALYs. 6.42 − 6.02 = 0.40, so it adds 0.40 QALYs per patient, a little under five months in full health.

3. The ratio. Dividing the cost difference by the QALY difference gives the ratio below.

$$\text{ICUR} = \frac{8{,}400}{0.40} = 21{,}000$$

where the result is GBP 21,000 per QALY gained, the price at which this treatment buys extra health relative to usual care.

4. The inverse form. The IUCR is 0.40 / 8,400 × 1,000,000, or about 47.6 QALYs gained per GBP 1 million of extra spending.

5. Net benefit at a threshold. At a value of GBP 25,000 per QALY, the net monetary benefit is 25,000 × 0.40 − 8,400 = 1,600, or GBP 1,600 per patient, and the net health benefit is 0.40 − 8,400/25,000 = 0.064 QALYs per patient. Both are positive, which is the same conclusion as an ICUR of GBP 21,000 sitting below GBP 25,000.

The ratio is meaningful here because both differences are positive and the comparator is the current standard. The examples below change one of those conditions at a time.

Reading an ICUR against a threshold

An ICUR becomes a decision aid only when it is set against a willingness-to-pay threshold or a cost-effectiveness threshold that represents the value placed on a QALY. A ratio below the threshold indicates value for money, not affordability, which is assessed through budget impact analysis. When the intervention adds QALYs, a ratio below the threshold is the same statement as a positive net monetary benefit. In symbols, the two statements are linked as follows.

$$\frac{\Delta C}{\Delta Q} < \lambda \iff \lambda \, \Delta Q - \Delta C > 0 \quad \text{when } \Delta Q > 0$$

where $\lambda$ is the threshold value of a QALY and $\Delta C$ and $\Delta Q$ are defined as above. Multiplying both sides by $\Delta Q$ preserves the inequality only when $\Delta Q$ is positive. When the intervention saves money but loses QALYs, the rule reverses. Paulden notes that a positive ICER below the threshold indicates cost-effectiveness in the north-east quadrant of the cost-effectiveness plane but not in the south-west quadrant.

Under the NICE manual (PMG36, as last updated 13 October 2026), a recommendation below a most plausible ICER of GBP 25,000 per QALY gained is normally based on the cost-effectiveness estimate and the acceptability of the technology as an effective use of NHS resources, and the illustrative GBP 21,000 would fall below GBP 25,000.

Negative ratios and near-zero QALY differences

A ratio loses information about the signs of its two parts, and it becomes unstable when its denominator approaches zero. Both problems arise in cost-utility analysis, where QALY differences between active treatments can be small. The table holds the incremental cost from the worked example at GBP 8,400 and varies the QALY gain; the figures are illustrative.

Moving the QALY gain from 0.40 to 0.05, a change of 0.35 QALYs, multiplies the ratio by eight, and a further 0.10 QALYs flips its sign. A negative ratio can also come from a treatment that saves money and gains QALYs, and Paulden notes that a negative ICER is cost-effective in the south-east quadrant but not in the north-west quadrant, so the quadrant has to be reported with the ratio.

For this reason an ICUR should never be reported without $\Delta C$ and $\Delta Q$ themselves. O'Brien and Briggs observe that confidence intervals for cost-effectiveness ratios, whether by bootstrapping or Fieller's method, can be problematic when the numerator or denominator can be zero. The NICE manual (PMG36) asks for expected net health benefits alongside ICERs and notes that they may be particularly informative when differences in costs or QALYs are small or when technologies provide less health benefit at lower cost.

Worked example where the wrong comparator flatters the ratio

The figures are illustrative. In a variant of the example, an older drug is still listed for the same condition, with an expected cost of GBP 18,000 and 5.90 QALYs per patient. Usual care, at GBP 16,200 and 6.02 QALYs, is both cheaper and more effective, so the older drug is dominated.

1. Against the dominated option. The new treatment adds 24,600 − 18,000 = 6,600 in cost and 6.42 − 5.90 = 0.52 QALYs, giving an ICUR of about GBP 12,692 per QALY gained.

2. Against the relevant option. Against usual care the ICUR is GBP 21,000 per QALY gained, as calculated above.

The choice of comparator alone has cut the reported ratio by about 40%. The NICE manual (PMG36) requires results in a fully incremental analysis with dominated and extendedly dominated technologies removed, and in this example that rule removes the older drug before any ratio is formed.

The ICUR beside nearby measures

Several measures share the ICUR's inputs but answer different questions. The distinctions matter mainly when figures from different studies or agencies are put side by side.

The IUCR is the inverse of the ICUR, expressed in QALYs per unit of spending, and is the form PHARMAC prefers. Net monetary benefit and net health benefit combine $\Delta C$, $\Delta Q$ and the threshold; because they are linear in $\Delta C$ and $\Delta Q$, their sign means the same thing in every quadrant, and they can be averaged across simulations and used to rank options.

Misreadings of an ICUR

Most of the errors below come from treating a ratio as if it carried the same information as the two differences behind it. Both affect how results are summarised and compared.

Averaging ratios across simulations. In probabilistic sensitivity analysis the reported ICUR is the ratio of mean incremental cost to mean incremental QALYs. With two equally likely illustrative draws, each with an incremental cost of GBP 8,400 and QALY gains of 0.40 and 0.02, the draw-level ratios are GBP 21,000 and GBP 420,000 and their mean is GBP 220,500. The mean incremental cost is GBP 8,400 and the mean QALY gain 0.21, so the ratio of means is 8,400/0.21 = 40,000, or GBP 40,000 per QALY. Paulden adds that, because the ICER is a discontinuous measure, a meaningful credible interval may be impossible to report.

Ranking options by their ICURs. A lower ICUR is not always the more cost-effective choice. Paulden notes that a higher or lower ICER does not necessarily imply that a strategy is more or less cost-effective than another, so ratios from different comparisons cannot be ranked to find the better buy.

Sources

• Drummond MF, Sculpher MJ, Claxton K, Stoddart GL, Torrance GW. Methods for the Economic Evaluation of Health Care Programmes. 4th ed. Oxford: Oxford University Press; 2015.
• Kamaraj A, Agarwal N, Seah KTM, Khan W. Understanding cost-utility analysis studies in the trauma and orthopaedic surgery literature. EFORT Open Reviews. 2021;6(5):305-315. doi:10.1302/2058-5241.6.200115
• National Advisory Committee on Immunization (NACI). Interpretation Guide: Health Economics. Version 1.0. Ottawa: Public Health Agency of Canada; March 2024.
• National Institute for Health and Care Excellence (NICE). NICE technology appraisal and highly specialised technologies guidance: the manual (PMG36). London: NICE; 2022, last updated 13 October 2026.
• O'Brien BJ, Briggs AH. Analysis of uncertainty in health care cost-effectiveness studies: an introduction to statistical issues and methods. Statistical Methods in Medical Research. 2002;11(6):455-468. doi:10.1191/0962280202sm304ra
• Paulden M. Why it's time to abandon the ICER. PharmacoEconomics. 2020;38(8):781-784. doi:10.1007/s40273-020-00915-5
• Pharmaceutical Management Agency (PHARMAC). Prescription for Pharmacoeconomic Analysis: Methods for Cost-Utility Analysis. Version 2.2. Wellington: PHARMAC; 2015.
• Weinstein MC, Stason WB. Foundations of cost-effectiveness analysis for health and medical practices. New England Journal of Medicine. 1977;296(13):716-721. doi:10.1056/NEJM197703312961304