The NICE severity modifier explained: QALY shortfall, the 1.2 and 1.7 weights, and what has changed in 2026
Since 2022 NICE has weighted QALYs for severe conditions using absolute and proportional QALY shortfall. Two changes in 2026, a higher threshold range and the move to the EQ-5D-5L value set, make it worth revisiting how the modifier works and where submissions win or lose a weight.
What the severity modifier is
NICE's technology appraisal manual (PMG36) asks committees to consider the severity of the condition. It defines severity as the future health lost by people living with the condition under standard NHS care, and notes that the extent of unmet need is reflected within this definition.[1] In the reference case all QALYs carry equal weight. The severity modifier is one of the decision modifiers that allow committees to depart from this, quantitatively, through a QALY weight.[1]
The modifier applies to technology appraisals of medicines. It is not applied in highly specialised technologies, where severity is taken to be captured in topic selection. It is unlikely to be relevant to diagnostics. It is not yet applied to technology appraisals of HealthTech, where NICE says it is exploring approaches.[1]
NICE introduced the modifier in 2022 to replace the end-of-life modifier. NICE describes the change as "opportunity cost neutral": designed so that it would not displace more NHS activity than end of life had done.[2]
Absolute and proportional QALY shortfall
Severity is measured by comparing the QALYs that people with the condition can expect under current NHS care with the QALYs that the general population of the same age and sex can expect over their remaining lifetime.[1]
Absolute shortfall (AS) = general population QALYs - QALYs with current treatmentProportional shortfall (PS) = AS / general population QALYs
The QALYs with current treatment are the total QALYs for established NHS practice, taken from the economic model for the population in which the technology will be used.[1] Both estimates are discounted at the reference-case rate.[1]
| QALY weight | Proportional QALY shortfall | Absolute QALY shortfall |
|---|---|---|
| 1 | Less than 0.85 | Less than 12 |
| x1.2 | 0.85 to 0.95 | 12 to 18 |
| x1.7 | At least 0.95 | At least 18 |
Source: NICE PMG36, table 6.1.[1]
Two rules decide how the table is read. The weight is set by whichever measure implies the greater severity. And a value that falls exactly on a cut-off takes the higher level.[1] The weight is applied to incremental QALYs. The DSU guidance adds that it applies only to patients' incremental QALYs, not carers'.[3]
The two measures do different jobs. AS favours younger populations, because they have more QALYs to lose. PS favours conditions that remove most of a person's remaining expected health, at any age. Having both, with the more favourable one applied, gives each type of severity a route to a weight.
Estimating general population QALYs
PMG36 asks only that population EQ-5D and survival data come from "a recent and robust source".[1] The detail is in NICE Decision Support Unit Technical Support Document 23 (TSD 23), first published in January 2024 and updated in September 2026.[3] In summary:
- Mortality: ONS period life tables for England, 2017 to 2019, weighted by the sex mix of the relevant population. TSD 23 avoids the 2020 data because of uncertainty about the long-term effect of COVID-19.[3]
- Population at the point in the pathway: The relevant population is people with the condition at the point in the pathway where the technology would be used, not the lifetime course of the disease.[3]
- Utility norms: For EQ-5D-3L, modelled values from the 2014 Health Survey for England. For EQ-5D-5L, modelled values from the 2017 and 2018 Health Survey for England, valued with the new UK EQ-5D-5L value set.[3]
- Discounting: Always 3.5%, even where the committee accepts 1.5% for the cost-effectiveness analysis, so that severity is assessed consistently across appraisals.[3]
- Other adjustments: Exclude carer quality of life. Present AS and PS for each comparator where several reflect NHS practice. Extend the model to a lifetime horizon for the comparator arm where the base case uses a shorter one.[3]
The move to EQ-5D-5L
The largest recent change is the value set. NICE published its interim methods statement on implementing the EQ-5D-5L value set (PMG51) on 27 August 2026. The 5L value set must be used for technology appraisals with an invitation to participate issued after that date. Ongoing topics continue to map to 3L.[4] For severity, NICE says AS and PS should then be calculated using 5L utilities and 5L population norms. The cut-offs in table 6.1 are unchanged.[4] Both the DSU Excel calculator and the R-based calculator from McNamara and colleagues have been updated for 5L.[4] The R-based calculator builds on work by McNamara and colleagues, who estimated quality-adjusted life expectancy norms for England from 14,412 Health Survey for England respondents and ONS 2017 to 2019 life tables.[5]
The change matters because 5L population norms are higher at a given age. In the TSD 23 hypothetical example (mean age 62, 33% female, comparator 6.031 QALYs), discounted general population QALYs are 11.96 with 3L norms and 12.69 with 5L norms. AS rises from 5.93 to 6.66 and PS from 0.50 to 0.52.[3] That example holds the comparator QALYs constant for illustration. In a real 5L submission, comparator utilities would also change, so the net effect on shortfall cannot be assumed in advance.
A worked example
Method noteGeneral population values below are our calculations, reproducing the NICE DSU severity calculator (2026 version) with its EQ-5D-5L option, a 50% female population and 3.5% discounting. The patient and cost figures are hypothetical.
Step 1: how much comparator QALYs can be before a weight is lost
Because the general population benchmark falls with age, the same comparator QALYs mean very different severity at different ages. The table shows the highest comparator QALYs that would still earn each weight.
| Mean age | General population QALYs (5L, discounted) | Max comparator QALYs for x1.2 | Max comparator QALYs for x1.7 | Binding measure |
|---|---|---|---|---|
| 30 | 21.48 | 9.48 (AS) | 3.48 (AS) | Absolute |
| 45 | 18.07 | 6.07 (AS) | 0.90 (PS) | Mixed |
| 60 | 13.49 | 2.02 (PS) | 0.67 (PS) | Proportional |
| 75 | 7.96 | 1.19 (PS) | 0.40 (PS) | Proportional |
For a 30-year-old population, a weight of 1.2 is available even where patients can expect more than 9 discounted QALYs on current care. At age 75, AS cannot reach 12 at all, and the full test falls on PS.
Step 2: applying the weight
Take a population with mean age 60 (general population QALYs 13.49) whose comparator arm yields 1.20 discounted QALYs:
AS = 13.49 - 1.20 = 12.29(12 to 18, so x1.2)PS = 12.29 / 13.49 = 0.911(0.85 to 0.95, so x1.2)
Suppose the new technology adds 0.80 QALYs at an incremental cost of £32,000:
- Unweighted ICER:
£32,000 / 0.80 = £40,000 per QALY - Weighted ICER:
£32,000 / (0.80 x 1.2) = £33,333 per QALY - Net health benefit at £35,000:
0.80 - 32,000/35,000 = -0.114unweighted,0.96 - 0.914 = +0.046weighted - Net health benefit at £25,000:
0.80 - 1.28 = -0.48unweighted,0.96 - 1.28 = -0.32weighted
PMG36 requires net health benefit to be presented in this way, at £25,000 and £35,000 and with and without the weight. A technology with negative unweighted net health benefit may still be recommended once a modifier applies.[1]
Now change one assumption. If a different survival extrapolation gave the comparator arm 2.10 QALYs, AS would be 11.39 and PS 0.844. The weight would be 1, and the ICER of £40,000 would sit above the range. If the comparator arm gave 0.60 QALYs, PS would be 0.956, the weight would be 1.7 and the weighted ICER would be about £23,500. Shifts of around one discounted QALY in the comparator arm, in either direction, move the decision-relevant ICER across the whole threshold range.
Calculate the weight for your population
The calculator below uses the same method and data as the NICE DSU severity calculator. Enter your population and comparator QALYs to see the shortfall, the weight and the margin to each cut-off. It is also available as a standalone NICE severity modifier calculator.
How it differs from the end-of-life modifier
| End of life (until 2022) | Severity modifier (2022 onwards) | |
|---|---|---|
| Eligibility | Life expectancy normally under 24 months and an extension to life normally of at least 3 months[6] | AS or PS against age- and sex-matched general population QALYs[1] |
| Weight | Up to a maximum of 1.7, set by the weight needed to bring the ICER into range[6] | Fixed steps: 1, 1.2 or 1.7[1] |
| Disease scope | Often cancer treatments[2] | Any severe condition; NICE cites examples in hepatitis D, cystic fibrosis, glioma and leukaemia[2] |
| Quality of life | Not part of the criteria | Built into the shortfall calculation |
| Status as a future comparator | Not necessarily accepted as a standard comparator[6] | Relevant comparator if recommended for routine use[1] |
Three differences matter in practice. The severity modifier counts quality of life, so a chronic condition with poor quality of life can qualify. It treats each technology as a step function rather than a sliding scale up to 1.7. And by measuring shortfall from the point in the pathway, it ties the weight closely to the comparator arm of the economic model.
How the modifier has been applied since 2022
Before implementation: Njoroge and colleagues applied the severity criteria to 132 NICE appraisals published from 2020 to 2022. End of life had been applied in 46 (33%), and 57 (39%) would have qualified for a severity weight. Of those, 79% were in oncology. They concluded that the modifier is harder to achieve and less generous than end of life, with only a fraction of appraisals reaching the highest weight.[7] That analysis used the former £20,000 to £30,000 range.
The DSU review: A NICE-funded DSU report by Wailoo (September 2024) compared 68 decisions from 47 appraisals under the new manual, to 31 March 2024, with 464 earlier decisions used to set the cut-offs. The weight of 1.2 applied in 20.6% versus 29.3% of decisions, and 1.7 in 8.8% versus 8.2%. The average weight was 1.103 versus 1.116, and the overall distributions were not significantly different. The author noted that it "may be too early to draw conclusions". He also cautioned that changes in data sources and methods, especially for AS, make samples hard to compare.[8]
NICE's position: In September 2024 the NICE Board concluded that the modifier was working as intended and that no change was required. It agreed to keep monitoring its impact and to commission research on societal preferences for severe conditions.[2] PMG51 restates both commitments: NICE will monitor the cumulative mean QALY weight, and research on preferences "may inform changes to the severity modifier in the future".[4]
Industry monitoring: The ABPI's October 2025 CONNIE report, based on company feedback, analysed 100 topics. About 72% received no weight, about 23% a weight of 1.2 and about 6% a weight of 1.7, with an average weight of 1.085. Among 62 oncology topics, 32% received 1.2 and 9% received 1.7.[9] Methods and denominators differ from the DSU review, so the two sets of figures should not be compared directly.
Assessment-group experience: In a 2026 Value in Health paper, an evidence assessment group described one appraisal in which its preferred utility values reduced the severity weight from 1.7 to 1.2. Starting age and the survival extrapolation for standard care also moved the result. The authors recommend routinely reporting AS, PS and the resulting weight under alternative assumptions.[10]
Interaction with the cost-effectiveness threshold
Since March 2026 PMG36 has used £25,000 and £35,000 per QALY as the reference points for decision making, replacing £20,000 and £30,000. NICE said the new range applies to new appraisals and those under way.[11, 12] The severity cut-offs and weights were not changed.[1, 4]
Because the weight multiplies incremental QALYs, its effect can be expressed as an equivalent unweighted ICER. This is simple arithmetic, not a NICE-published threshold:
| Weight | Equivalent unweighted range, £20,000 to £30,000 basis | Equivalent unweighted range, £25,000 to £35,000 basis |
|---|---|---|
| 1 | £20,000 to £30,000 | £25,000 to £35,000 |
| 1.2 | £24,000 to £36,000 | £30,000 to £42,000 |
| 1.7 | £34,000 to £51,000 | £42,500 to £59,500 |
Two cautions apply. NICE does not use a precise maximum acceptable ICER, and decisions in the upper part of the range depend on uncertainty, uncaptured benefits and health inequalities.[1] And the move to 5L norms changes the size of the shortfall, so the share of appraisals reaching each weight may change too. NICE's monitoring of the mean weight is the mechanism through which that would be seen.[4]
Our viewUnder the higher threshold, a weight of 1.2 is worth roughly £5,000 to £7,000 per QALY at the edges of the range, and 1.7 is worth considerably more. In our experience, the difference between 1 and 1.2 is often decided by modelling choices in the comparator arm, not by the clinical picture alone.
Implications for evidence and modelling
- The comparator arm is now a pricing input: Survival extrapolation, utilities and the treatment mix for standard care drive both the ICER and the weight. The EAG case study above shows a full category change from these choices.[10]
- Baseline age and sex need care: Lee and colleagues found that using a single mean age rather than the age distribution changed predicted quality-adjusted life expectancy by up to 10.4% (1.01 QALYs). The severity category changed only at the extremes of age or with very high variance.[13] Our step 1 table shows why mean age near a boundary deserves close attention.
- Population definition: Because severity is assessed at the point in the pathway, subgroups with different comparators need their own AS and PS.[3] Any subgroup case should be clinically grounded, not built around the cut-offs.
- 5L transition: For topics invited after 27 August 2026, both the cost-effectiveness model and the shortfall calculation use 5L values. NICE says 3L-based values should not appear in scenario analyses.[4] Utility evidence plans written for 3L mapping need revisiting.
- Carers: Carer QALYs may support cost effectiveness but are excluded from AS and PS and do not attract the weight.[3]
What this means for evidence planning
- Estimate AS and PS early, from the planned comparator model, using the DSU or McNamara calculator with 5L norms. Then run the same calculation across plausible extrapolations and utility sources.
- Find the margin to each cut-off: If the base case sits within a few tenths of a QALY of a boundary, treat the weight as uncertain in pricing scenarios.
- Strengthen standard-care evidence: Natural history data, registry survival and quality of life on current NHS care are as important to the weight as trial data on the new technology.
- Report transparently: Present AS, PS and the weight for each scenario, and net health benefit at £25,000 and £35,000 with and without weighting, as PMG36 requires.[1]
- Link the weight to commercial planning: Model the price needed under each plausible weight before committing to a commercial position.
Evidaxia builds and adapts cost-effectiveness models for NICE submissions, including severity shortfall analyses under the EQ-5D-5L value set. See HTA submissions in the UK and model adaptation.
References
- National Institute for Health and Care Excellence. NICE technology appraisal and highly specialised technologies guidance: the manual (PMG36), chapter 6: committee recommendations [Internet]. London: National Institute for Health and Care Excellence; 2022 Jan 31 [updated 2026 Oct 13; cited 2026 Oct 30]. Available from: https://www.nice.org.uk/process/pmg36/chapter/committee-recommendations
- National Institute for Health and Care Excellence. What is NICE's severity modifier? NICE in depth [Internet]. London: National Institute for Health and Care Excellence; [date unknown] [cited 2026 Oct 30]. Available from: https://indepth.nice.org.uk/what-is-nices-severity-modifier/index.html
- Wailoo A. A guide to calculating severity shortfall for NICE evaluations. NICE DSU Technical Support Document 23 [Internet]. Sheffield: NICE Decision Support Unit, University of Sheffield; 2024 Jan [updated 2026 Sep; cited 2026 Oct 30]. Available from: https://sheffield.ac.uk/nice-dsu/tsds/severity-shortfall-tsd
- National Institute for Health and Care Excellence. Interim methods statement: implementing the EQ-5D-5L value set (PMG51) [Internet]. London: National Institute for Health and Care Excellence; 2026 Aug 27 [cited 2026 Oct 30]. Available from: https://www.nice.org.uk/process/pmg51/chapter/introduction
- McNamara S, Schneider PP, Love-Koh J, Doran T, Gutacker N. Quality-adjusted life expectancy norms for the English population. Value Health. 2023;26(2):163-9. Available from: https://pubmed.ncbi.nlm.nih.gov/35965226/
- National Institute for Health and Care Excellence. Guide to the methods of technology appraisal 2013 (PMG9), section 6: the appraisal of the evidence and structured decision-making [Internet]. London: National Institute for Health and Care Excellence; 2013 Apr 4 [cited 2026 Oct 30]. Available from: https://www.nice.org.uk/process/pmg9/chapter/the-appraisal-of-the-evidence-and-structured-decision-making
- Njoroge MW, Walton M, Hodgson R. Understanding the National Institute for Health and Care Excellence severity premium: exploring its implementation and the implications for decision making and patient access. Value Health. 2024;27(6):730-6. Available from: https://pubmed.ncbi.nlm.nih.gov/38447743/
- Wailoo A. Severity weights in NICE technology appraisals [Internet]. Sheffield: NICE Decision Support Unit, University of Sheffield; 2024 Sep 5 [cited 2026 Oct 30]. Available from: https://sheffield.ac.uk/media/76321/download
- Association of the British Pharmaceutical Industry. Reviewing implementation in practice of the NICE health technology evaluation manual: CONNIE [Internet]. London: Association of the British Pharmaceutical Industry; 2025 Oct [cited 2026 Oct 30]. Available from: https://www.abpi.org.uk/publications/reviewing-implementation-in-practice-of-the-nice-health-technology-evaluation-manual-connie-october-2025/
- Abdelsabour A, Perks J, Kelman C, Robinson S, Farmer C, Lee D. The implementation of National Institute for Health and Care Excellence severity modifier in the single technology appraisal of mirvetuximab soravtansine: evidence assessment group perspective. Value Health. 2026. doi:10.1016/j.jval.2026.08.009 Available from: https://pubmed.ncbi.nlm.nih.gov/42767341/
- National Institute for Health and Care Excellence. Changes to NICE's cost-effectiveness thresholds confirmed [Internet]. London: National Institute for Health and Care Excellence; 2025 Dec 1 [cited 2026 Oct 30]. Available from: https://www.nice.org.uk/news/articles/changes-to-nice-s-cost-effectiveness-thresholds-confirmed
- National Institute for Health and Care Excellence. NICE technology appraisal and highly specialised technologies guidance: the manual (PMG36). Update information [Internet]. London: National Institute for Health and Care Excellence; [date unknown] [updated 2026 Oct 13; cited 2026 Oct 30]. Available from: https://www.nice.org.uk/process/pmg36/chapter/update-information
- Lee D, Hart R, Burns D, McCarthy G. The impact of the approach to accounting for age and sex in economic models on predicted quality-adjusted life-years. Appl Health Econ Health Policy. 2025;23(1):131-40. Available from: https://pubmed.ncbi.nlm.nih.gov/39320707/
Sources accessed October 2026. This article is analysis for information and is not legal, regulatory or tax advice.
One evidence package, thirteen answers.
Our 21-page guide compares thirteen HTA systems across the UK, Europe and Canada, with 16 exhibits and full references. Subscribe to Evidaxia Briefing, our sourced analysis of market access, HTA decisions, evidence and health policy, and the guide downloads the moment you subscribe.
What is the NICE severity modifier?
It is a decision modifier in NICE technology appraisals that applies a QALY weight of 1.2 or 1.7 to incremental QALYs when the condition is severe, measured by absolute or proportional QALY shortfall under current NHS care. It replaced the end-of-life modifier in 2022.
What are the thresholds for the 1.2 and 1.7 severity weights?
A weight of 1.2 applies for proportional shortfall of 0.85 to 0.95 or absolute shortfall of 12 to 18 QALYs. A weight of 1.7 applies for proportional shortfall of at least 0.95 or absolute shortfall of at least 18. The measure giving the higher weight is used, and values on a cut-off take the higher level.
How is QALY shortfall calculated?
Absolute shortfall is general population QALYs for the same age and sex minus the QALYs expected with current NHS treatment. Proportional shortfall is absolute shortfall divided by general population QALYs. Both are discounted at 3.5%.
Which data does NICE use for general population QALYs?
NICE DSU TSD 23 recommends ONS 2017 to 2019 life tables with Health Survey for England utility norms. Since NICE adopted the EQ-5D-5L value set in August 2026, new topics use 5L norms. The DSU and McNamara calculators have been updated accordingly.
Did the 2026 threshold change alter the severity weights?
No. NICE moved to a £25,000 to £35,000 per QALY range, and the EQ-5D-5L interim statement confirms that the severity cut-offs are unchanged. The weights now apply on top of the higher range.
How does the severity modifier differ from end of life?
End of life applied to treatments for people with short life expectancy, normally under 24 months, and allowed a weight of up to 1.7. The severity modifier uses QALY shortfall, includes quality of life, covers any condition and applies fixed weights of 1, 1.2 or 1.7.
Related analysis
Close to a severity cut-off? Let us test the shortfall before you submit.
Start with a free, no-obligation 30-minute call with a senior consultant.