Absolute and proportional QALY shortfall

Absolute shortfall subtracts the expected remaining QALYs of people with the condition under current care from those of the matched general population. Proportional shortfall divides the same difference by the general-population QALYs. Both look forward from the point at which the technology would be used, so health already lost does not count, and both are discounted at the reference-case rate.

Signature

AS = Q_h - Q_c; PS = (Q_h - Q_c) / Q_h
Inputs
InputsDefinitionUnit
Q_hExpected remaining discounted QALYs of a general population with the same age and sex distribution, greater than zerodiscounted QALYs per person
Q_cExpected remaining discounted QALYs of people with the condition under established NHS practice, usually the comparator arm of the preferred modeldiscounted QALYs per person
Output
ASFuture QALYs lost by people with the condition compared with the matched general populationdiscounted QALYs per person
PSShare of the general population's remaining QALYs lost with the conditionproportion, normally from 0 to 1

Function

Severity shortfall function

Maps the remaining discounted QALYs expected for an age- and sex-matched general population, and for people with the condition under current NHS care, to absolute and proportional QALY shortfall. NICE uses whichever measure implies greater severity to set a weight on the incremental QALYs of a new technology.

Try this function

Implementations

  • Excel

    Shortfall measures in two cells

    With the general-population QALYs in QALYsGeneral and the condition QALYs in QALYsCondition, the two formulas return absolute and proportional shortfall.

    =QALYsGeneral-QALYsCondition; =(QALYsGeneral-QALYsCondition)/QALYsGeneral

Assumptions

  • Same population, starting point and discounting

    Q_h and Q_c describe the same starting age and sex mix, from the point in the pathway where the technology is considered, over the remaining lifetime, discounted at the reference-case rate of 3.5% a year. The DSU guide applies this rate even where the committee accepts another rate for the cost-effectiveness analysis.

  • Carer QALYs excluded

    Carer quality of life included in the model's QALYs is removed before the calculation, both from Q_c and from the incremental QALYs that attract a weight.

  • Value set for population norms

    For topics starting after 27 August 2026 (in technology appraisals, an invitation to participate issued after that date) the EQ-5D-5L value set is used for both the model utilities and the general-population norms behind Q_h. Ongoing topics started earlier continue to use EQ-5D-3L values, and the cut-offs are unchanged.

Worked examples

  • DSU hypothetical case with EQ-5D-3L norms

    A population aged 62, 33% female, has 11.96 discounted general-population QALYs, the sex-weighted average of 11.86 for males and 12.16 for females, and 6.031 under current care. Absolute shortfall is about 5.93 and proportional shortfall about 0.50, as in section 5.1 of TSD 23. Both fall below the lowest cut-offs, so the weight is 1.

    Q_h = 11.96; Q_c = 6.031; AS = 5.929; PS = 0.496
  • DSU hypothetical case with EQ-5D-5L norms

    With EQ-5D-5L norms the same population has 12.69 general-population QALYs, giving an absolute shortfall of about 6.66 and a proportional shortfall of about 0.52, as in section 5.2 of TSD 23. The weight remains 1.

    Q_h = 12.69; Q_c = 6.031; AS = 6.659; PS = 0.525
  • Young population with a severe condition

    A general population with 20 remaining discounted QALYs and a condition leaving 1.5 gives an absolute shortfall of 18.5 and a proportional shortfall of 0.925, as in the article. The figures are illustrative.

    Q_h = 20; Q_c = 1.5; AS = 18.5; PS = 0.925

Common errors

  • Undiscounted or non-reference-case discounting

    Shortfall calculated from undiscounted QALYs, or at a discount rate other than 3.5%, is on a different scale from the cut-offs of 12 and 18 QALYs and can move a condition into a higher band.

  • Matching on age but not on sex

    Life expectancy and utility norms differ by sex, so the DSU guide computes Q_h separately for males and females and weights them by the proportion female in the patient population. A single-sex or whole-population norm misstates Q_h and both shortfall measures.

Sources

  • NICE definitions of absolute and proportional QALY shortfall

    National Institute for Health and Care Excellence. NICE technology appraisal and highly specialised technologies guidance: the manual (PMG36). Published 31 January 2022, last updated 31 March 2026. Chapter 6 Committee recommendations, sections 6.2.12 to 6.2.17 (severity as future health lost with standard care; absolute and proportional QALY shortfall definitions; general population with the same age and sex distribution; discounting at the reference-case rate).

    View source →

  • DSU guide to calculating severity shortfall

    Wailoo A. NICE DSU Technical Support Document 23: A guide to calculating severity shortfall for NICE evaluations. Sheffield: Decision Support Unit, ScHARR, University of Sheffield; January 2024, updated September 2026. Section 2.1 (AS = Qh minus Qc and PS = (Qh minus Qc) / Qh, forward-looking definition and ranges), section 4.4 (carer quality of life excluded), sections 3.6 and 4.1 (3.5% discount rate, even where another rate is accepted for the cost-effectiveness analysis), section 5 worked examples with EQ-5D-3L and EQ-5D-5L, and appendix 1 step iv (weighting by sex).

    View source →

  • EQ-5D-5L value set for shortfall calculations

    National Institute for Health and Care Excellence. Interim methods statement: implementing the EQ-5D-5L value set (PMG51). Published 27 August 2026. Statement section 1.12 (the EQ-5D-5L value set is used for general-population utilities, including absolute and proportional QALY shortfall), process considerations sections 1.17 and 1.18 (new topics only; ongoing topics continue with 3L values) and appendix C, questions 21 and 22 (5L utilities and 5L population norms for severity, with the table 6.1 cut-offs unchanged).

    View source →

Canonical Identity

Stable URI · Machine-readable · Resolvable · CC BY 4.0