Reference Case

Reference Case Elements, Origins and Compliance at NICE

Reference cases give economic evaluations prepared for a particular institution or decision process a common methodological starting point, so that differences between submissions reflect the evidence rather than analysts' choices. A widely cited example is the NICE reference case, summarised in table 4.1 of the NICE manual (PMG36, section 4.2.2), which fixes the perspective, outcome measure, discount rate and other methods expected in technology appraisals in England. The NICE reference case moved to the EQ-5D-5L value set from 27 August 2026, and the idea itself runs from the 1996 US Panel on Cost-Effectiveness in Health and Medicine to the iDSI reference case. A reference case is distinct from a base case and from non-reference-case analyses, compliance with it is assessed requirement by requirement, and any departure from it has to be justified. It also differs from a reporting checklist.

A reference case does not perform an economic evaluation or make the final decision. It provides a consistent framework for conducting and reporting the analysis.

Why institutions use reference cases

Two analysts can study the same intervention and reach different results because they select different comparators, perspectives, time horizons, outcome measures or discount rates. A reference case reduces avoidable variation by stating which methodological choices should normally be used for a particular decision process.

Using a common framework makes submissions easier to compare and review. It also helps readers distinguish differences caused by evidence from differences caused by analytical choices.

• A reference case promotes consistency across economic evaluations prepared for the same institution.
• A reference case makes important methodological choices visible to analysts, reviewers and decision-makers.
• A reference case supports comparison across technologies without guaranteeing identical methods or conclusions.
• A reference case should be linked to a named institution, decision context, guidance version and effective date.

Where the reference case idea came from

The reference case was formalised by the Panel on Cost-Effectiveness in Health and Medicine, a non-federal panel convened by the US Public Health Service, which reported in 1996. Its recommendations defined a reference case analysis, a standard set of practices for cost-effectiveness analyses that inform societal resource allocation decisions, on the argument that studies following common methods would gain in quality and comparability. The full report was published as a book edited by Gold, Siegel, Russell and Weinstein, with a summary of the recommendations in JAMA.

The Second Panel on Cost-Effectiveness in Health and Medicine, formed in 2012, published updated recommendations in 2016. It kept the reference case but recommended that every cost-effectiveness analysis report two reference case analyses, one from a health care sector perspective and one from a societal perspective. It also recommended an impact inventory, a structured table of consequences inside and outside the formal health care sector, to make the scope of each analysis explicit.

In the same year the International Decision Support Initiative (iDSI) published a reference case, commissioned by the Bill and Melinda Gates Foundation, to guide economic evaluations in low- and middle-income countries. It draws on earlier work by the World Health Organization, the US Panel and NICE, and comprises 11 key principles, each with methodological specifications and reporting standards. Its authors describe it as an aid to thought that should not be followed mechanically without regard to local context.

What a reference case commonly specifies

A reference case usually covers several connected parts of an economic evaluation. The exact requirements differ, so analysts must use the current guidance that applies to their jurisdiction and decision.

The following elements are commonly addressed, but the list should not be treated as a universal template.

• The reference case identifies the population and relevant subgroups for the decision.
• The reference case identifies the appropriate comparator or comparators.
• The reference case defines the analytical perspective and which costs and outcomes enter the evaluation.
• The reference case identifies the preferred form of economic evaluation.
• The reference case defines the preferred health-outcome measure, such as the QALY where applicable.
• The reference case specifies the appropriate time horizon.
• The reference case specifies discount rates for costs and health outcomes.
• The reference case identifies preferred evidence sources and methods for estimating effectiveness.
• The reference case addresses modelling, uncertainty analysis and validation.
• The reference case specifies how costs, resource use and health outcomes should be measured and valued.
• The reference case establishes reporting and transparency expectations.

The NICE reference case in PMG36

In England, the NICE reference case is set out in section 4.2 of "NICE technology appraisal and highly specialised technologies guidance: the manual" (PMG36, published 31 January 2022, last updated 31 March 2026). Section 4.2.1 states that economic evaluations considered by NICE should include an analysis using reference-case methods, and that additional analyses using different methods must be justified and clearly distinguished from it. Section 4.2.2 summarises the key elements in table 4.1, shown below in condensed form with the manual sections that give the detail.

A note to table 4.1 states that the rows on measuring and valuing health effects, the two data-source rows and equity apply to cost-utility analysis and not to cost-comparison analysis. Section 4.2.9 excludes productivity costs from the reference-case perspective on costs, and section 4.2.3 asks for reasons for not applying reference-case methods to be specified, justified and quantified. The committee then discusses the weight it attaches to any non-reference-case analysis.

The EQ-5D row changed in 2026. The PMG36 overview page states that from 27 August 2026 the EQ-5D-5L descriptive system and the UK value set for the EQ-5D-5L should be used for all reference-case analyses in adults, following NICE's interim methods statement (PMG51), which takes precedence over the manual until the manual is updated. PMG51 amends the table 4.1 wording to name the EQ-5D-5L as the preferred measure, requires EQ-5D-3L data to be mapped to 5L utility values using the Decision Support Unit and EEPRU mapping model, and rules out 3L-based utility values in sensitivity or scenario analyses. The change applies prospectively. For technology appraisals it covers topics with an invitation to participate issued after the statement was published, while ongoing topics continue to map 5L data to 3L utility values. Until the manual text is revised, section 4.3.16 of PMG36 still carries the earlier advice to use the 3L value set.

Reference case, base case and non-reference-case analyses

The three terms describe different things and are easily confused. The reference case is a specification of methods set by an institution, stating which perspective, outcome measure, discount rate and other choices an analysis should use. The base case is the analysis a study presents as its primary result, built on the analyst's central, best-estimate inputs and structural assumptions, around which sensitivity and scenario analyses are run.

Because NICE requires an analysis using reference-case methods, a submission's base case and its reference-case analysis often coincide, but the terms are not interchangeable. A base case can follow every reference-case element and still rest on weak evidence or implausible assumptions. A scenario that changes only a parameter value, such as a treatment effect, remains a reference-case analysis, because the methods have not changed.

A non-reference-case analysis departs from 1 or more methodological elements of the reference case, for example by adopting a wider perspective on costs or a different discount rate. PMG36 gives two examples. Analyses of benefits to government bodies outside the NHS and PSS are non-reference-case analyses that must be presented separately and in disaggregated form (section 4.2.10). A discount rate of 1.5% a year for both costs and health effects is a non-reference-case rate that the committee may consider only when all the criteria in section 4.5.3 are met.

How the reference case shapes an economic evaluation

The reference case influences how an economic evaluation is designed before the model is calculated. It connects the decision problem to a controlled set of methodological choices so the resulting evidence fits the institution's decision process.

Following a reference case does not guarantee that the evidence, model or conclusion is correct. The quality of the analysis still depends on appropriate data, assumptions, implementation, validation and interpretation.

1. Identify the institution and decision process.
2. Confirm the current reference-case guidance and effective date.
3. Translate each applicable requirement into a project-level analytical choice.
4. Record the evidence supporting the project's implementation.
5. Identify any requirement that cannot be followed exactly.
6. Explain each deviation and its expected effect.
7. Test influential alternative choices through sensitivity or scenario analysis.
8. Report compliance, deviations, limitations and results transparently.

How reference cases differ across institutions

Reference cases are created for particular institutions and decision settings. Requirements may differ because institutions have different legal responsibilities, budgets, perspectives, healthcare systems and views about appropriate methods.

An analyst should not combine preferred elements from several reference cases and present the result as compliant with one institution. Cross-country comparisons should identify both the similarities and the differences explicitly.

Important areas to compare are listed below.

• Institutions differ in whose costs and outcomes the evaluation includes, which sets its perspective.
• Institutions define differently the existing practice or alternatives that must be evaluated as comparators.
• Some institutions prefer QALYs as the outcome measure, while others accept natural health units or other outcomes.
• The required time horizon sets how long relevant differences must be captured.
• The discount rates applied to costs and to health outcomes vary between institutions.
• The resources included in costs, the price sources and the price year can all differ.
• Institutions set their own standards for acceptable clinical and observational evidence.
• The expected sensitivity, scenario and probabilistic analyses of uncertainty differ between guidelines.
• Each institution specifies the results, assumptions and deviations that must be disclosed in reporting.

The Reference Case Comparison and Compliance Matrix preserves each institutional profile separately rather than creating an artificial blended reference case.

How to check whether an analysis follows the reference case

Reference-case compliance requires more than selecting yes or no beside a list of requirements. Each judgement should connect the institutional requirement to the project's implementation and supporting evidence.

Some requirements can be checked mechanically, such as whether separate cost and health discount rates match stated values. Other requirements, such as whether the time horizon is long enough, require documented expert judgement.

The compliance record can use six statuses, described below.

• A requirement is compliant when the project follows it and records supporting evidence.
• It is partially compliant when the project follows only part of the requirement or the evidence is incomplete.
• A justified deviation applies when the project departs from the requirement and documents a defensible reason and its expected effect.
• It is not compliant when the project departs from the requirement without an adequate justification.
• It is not applicable when the requirement does not apply to the stated project and the reason is recorded.
• The status review required applies when available information is insufficient for a defensible determination.

A formula may flag a possible mismatch, but it should not make a final judgement when clinical, methodological or institutional interpretation is needed.

When an analyst may depart from the reference case

A reference case is the expected starting point, but some evaluations require justified departures. A deviation may be necessary when the standard method cannot represent the technology, population, evidence or decision problem adequately.

A deviation should never be hidden or chosen merely because it produces a more favourable result. The analyst should explain why the reference-case method is unsuitable, what alternative was used and whether the departure changes the conclusion.

• A deviation record should identify the exact reference-case requirement.
• A deviation record should state the alternative method used.
• A deviation record should explain why the alternative is more appropriate.
• A deviation record should identify the evidence supporting the departure.
• A deviation record should describe the expected direction and size of its effect.
• A deviation record should include a reference-case scenario when feasible.
• A deviation record should identify who reviewed and accepted the judgement.

How sensitivity analysis supports reference-case reporting

Sensitivity analysis can show whether the result depends on a methodological choice that differs from the reference case. It does not convert a non-compliant base case into a compliant one or remove the need to explain a deviation.

Where feasible, analysts should present the institution's required reference case alongside justified alternatives. This allows reviewers to separate the effect of the alternative method from other differences in the evaluation. The NICE manual (PMG36, section 4.7.7) allows uncertainty about the appropriateness of reference-case methods to be explored through sensitivity analysis, provided those analyses are presented separately.

• A reference-case analysis should use the institution's preferred methodological choices.
• A deviation scenario should isolate the effect of an alternative methodological choice where feasible.
• Sensitivity analysis should show whether the ICER, net benefit or preferred alternative changes.
• Structural choices should be tested with explicit scenarios when parameter variation cannot represent them.
• Scenario labels should state exactly which reference-case choice has changed.

How a reference case relates to HTA

Health technology assessment may use economic evaluation as one source of evidence, and a reference case can define how that economic evidence should be produced. The reference case therefore supports the assessment process without determining the final recommendation.

HTA may also consider clinical effectiveness, safety, equity, ethical issues, organisational consequences, patient perspectives and implementation. Compliance with the economic reference case does not guarantee adoption, reimbursement or coverage.

• A reference case standardises selected methods used to produce economic evidence.
• Economic evaluation provides comparative evidence about costs and consequences.
• HTA combines economic evidence with other decision-relevant information.
• The responsible decision-making body retains authority over the final recommendation.

How a reference case differs from a reporting checklist

A reference case tells analysts which methodological choices are normally preferred for a defined decision context. A reporting checklist tells authors what information should be disclosed so readers can understand and assess a study.

The two tools can overlap, but they are not interchangeable. A report can describe a non-reference-case analysis clearly, while an analysis can claim compliance but still report its methods inadequately.

• Reference-case compliance concerns whether the selected methods follow applicable institutional requirements.
• Reporting compliance concerns whether the analysis describes its methods and results adequately.
• Transparent reporting should identify both reference-case compliance and justified deviations.
• A reporting checklist should not be used as evidence that the analytical choices match an institutional reference case.

Using the comparison and compliance workbook

The Reference Case Comparison and Compliance Matrix allows users to select an institutional profile and compare its requirements with the choices used in a project. The workbook separates institutional source information, project implementation, evidence, deviations and reviewer determinations.

Automated checks identify possible inconsistencies in fields such as discount rates. Qualitative requirements, including the adequacy of the time horizon or comparator, remain clearly marked as requiring evidence-based human review.

What should be reported

A transparent economic evaluation should identify the governing reference case and show how the analysis follows it. Readers should also be able to see where the analysis departs from the expected approach and whether those departures influence the result.

1. Report the institution and decision context governing the evaluation.
2. Report the reference-case guidance title, version, effective date and URL.
3. Report how each applicable requirement was implemented.
4. Report the evidence supporting compliance judgements.
5. Report every material deviation and its justification.
6. Report the expected effect of each deviation.
7. Report reference-case and alternative scenarios where feasible.
8. Report who reviewed qualitative compliance judgements.
9. Report unresolved requirements and missing evidence.
10. Report whether deviations change the decision result.

Common mistakes and how to avoid them

Reference-case errors often occur when analysts treat a framework as a generic checklist or rely on remembered requirements from an earlier guidance version. These errors can undermine comparability and make a submission appear compliant when important choices remain unsupported.

• Applying one institution's reference case to another jurisdiction can use the wrong methodological requirements.
• Combining requirements from different institutions can create a hybrid framework that no institution actually endorses.
• Using outdated guidance can produce incorrect comparator, perspective, horizon or discount-rate choices.
• Treating a formula flag as a final compliance decision can replace necessary expert judgement.
• Marking a requirement not applicable without a reason can conceal missing analysis.
• Reporting a deviation without its likely effect prevents reviewers from judging its importance.
• Using sensitivity analysis without explaining the base-case departure does not establish compliance.
• Treating reference-case compliance as proof of model validity overlooks evidence quality, implementation and validation.
• Treating reference-case compliance as an adoption decision ignores the wider HTA process.

Sources

• Gold MR, Siegel JE, Russell LB, Weinstein MC, editors. Cost-Effectiveness in Health and Medicine. New York: Oxford University Press. 1996.
• Weinstein MC, Siegel JE, Gold MR, Kamlet MS, Russell LB. Recommendations of the Panel on Cost-effectiveness in Health and Medicine. JAMA. 1996;276(15):1253-1258.
• Sanders GD, Neumann PJ, Basu A, Brock DW, Feeny D, Krahn M, et al. Recommendations for conduct, methodological practices, and reporting of cost-effectiveness analyses: Second Panel on Cost-Effectiveness in Health and Medicine. JAMA. 2016;316(10):1093-1103.
• Wilkinson T, Sculpher MJ, Claxton K, Revill P, Briggs A, Cairns JA, et al. The International Decision Support Initiative reference case for economic evaluation: an aid to thought. Value in Health. 2016;19(8):921-928.
• 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.
• National Institute for Health and Care Excellence. Interim methods statement: implementing the EQ-5D-5L value set (PMG51). Published 27 August 2026.
• Rowen D, Mukuria C, Bray N, Carlton J, Longworth L, Meads D, et al. A United Kingdom value set for the EQ-5D-5L. Value in Health. 2026;29(5):858-869.
• Drummond MF, Sculpher MJ, Claxton K, Stoddart GL, Torrance GW. Methods for the Economic Evaluation of Health Care Programmes, 4th edition. Oxford: Oxford University Press. 2015.