One evidence plan for NICE and the EU joint clinical assessment: where requirements align, where they diverge, and how to build a single SLR and ITC programme

Two assessments, one evidence base

For most new oncology medicines and advanced therapies launching in Europe today, the clinical evidence package will be examined twice in quick succession: once in the EU joint clinical assessment (JCA) under Regulation (EU) 2021/2282, and once by NICE for the NHS in England. The two processes share a vocabulary, systematic reviews, indirect treatment comparisons (ITCs), PICO questions, but they answer different questions.

The JCA is a scientific description of relative effects and their certainty. It "shall not contain any value judgement or conclusions on the overall clinical added value", and the Regulation separates clinical assessment from "non-clinical assessment", which includes cost and economic evaluation^[1]. NICE, by contrast, makes a recommendation about whether a technology is a cost-effective use of NHS resources. Its scope identifies comparators that are established NHS practice, and its reference case is a cost-utility analysis with fully incremental analysis^{[2, 3]}.

The UK is not part of the EU JCA, so the two processes run independently. In our experience, the risk for a global team is not that one submission contradicts the other, but that two parallel workstreams produce two literature reviews, two network meta-analyses and two sets of assumptions that are hard to reconcile later. A single plan avoids that duplication and protects consistency.

Where the requirements align and diverge

Comparators: inclusive in both, but defined differently

Both processes start inclusively. NICE's manual states that at the scoping stage "identifying comparators should be inclusive", and that the scope identifies all potentially relevant comparators that are established practice in the NHS^[2]. The JCA scope must be "inclusive and reflect Member States' needs"^[1].

The difference lies in how comparators are carried into the analysis. NICE requires a fully incremental cost-utility analysis across the relevant comparators^[3], so every comparator must sit in one coherent model. The JCA assesses each PICO separately. The HTACG scoping guidance allows a Member State to request one comparator, several comparators each separately, at least one of several comparators, or an individualised treatment comparator that bundles options chosen by patient characteristics^[5]. A JCA scope may therefore require a comparison against a regimen that NICE would not consider, or against a bundle that has no direct analogue in an NHS model.

Two NICE-specific rules also matter. Technologies recommended with managed access are not considered established practice and are not suitable comparators^[2]. And technologies recommended after applying the severity modifier become relevant comparators for future evaluations only when they are recommended for routine use and represent established practice^[4]. A comparator that is standard in several EU systems may therefore be excluded from the NICE decision problem, and vice versa.

Outcomes, quality of life and surrogate endpoints

The HTACG outcomes guidance asks for final patient-centred outcomes, such as mortality, morbidity and health-related quality of life, to be requested at scoping, states that the JCA report should not rank outcomes, and requires adverse events to be available irrespective of causality^[6]. It notes that generic quality-of-life instruments can have insufficient content validity in some conditions, in which case disease-specific instruments should be considered^[6].

NICE needs quality of life in a different form: as utilities that feed QALYs. Its interim methods statement of 27 August 2026 makes the EQ-5D-5L descriptive system and the new UK value set NICE's preferred approach for adults, applied prospectively to topics that start after publication; ongoing topics continue to map 5L data to 3L utilities^[7]. The practical point for trial design is that both needs can be met from the same trial if it collects a disease-specific instrument for the JCA and the EQ-5D-5L for NICE, with adequate completion.

On surrogates the two frameworks are close. NICE recognises three levels of evidence for surrogate relationships, with level 1 being RCT evidence that the effect on the surrogate corresponds to a commensurate effect on the final outcome, and validation preferably from a meta-analysis of RCTs reporting both^[3]. The HTACG guidance states that level 1 trial-level evidence "can be considered as required" and describes a correlation of at least 0.85 as "high", while noting that no universally accepted threshold exists^[6]. A single surrogate validation analysis, using bivariate meta-analytic methods, can serve both.

Subgroups and populations

NICE scopes may highlight subgroups where clinical effectiveness or value for money might differ^[2], and the economic case can be made separately for them. In the JCA, populations that need different comparators become separate PICOs, while requested subgroup analyses sit within PICOs^[5]. The HTACG Q&A states that all outcomes in the scope should be included in requested subgroup analyses, that subgroup analyses are required only where each subgroup includes at least 10 people (and, for binary or time-to-event outcomes, at least 10 events in one subgroup), and that an interaction test is needed where a subgroup in one PICO overlaps with another PICO's population^[13].

A unified plan therefore needs one subgroup specification: a prespecified list of candidate effect modifiers and biomarker-defined populations, analysed with the same definitions for both audiences. The output tables can then be cut by PICO for the JCA and by decision subgroup for NICE.

Indirect comparison methods: shared principles, different emphasis

Both frameworks prefer randomised, anchored evidence. The HTACG methodological guideline states that "anchored indirect comparisons preserve randomisation and should always be used in preference to unanchored methods", treats unanchored MAIC and STC without a common comparator as "highly problematic", and requires model and covariate selection to be prespecified^[8]. NICE's manual asks for a network meta-analysis where technologies have not been compared in a single RCT, states that ideally the network should contain all technologies identified in the scope, and allows population adjustment when effect modifiers are imbalanced, with individual patient data from at least one trial^[9]. The NICE Decision Support Unit recommendations add that the target population must be explicitly stated and that unanchored comparisons must adjust for all effect modifiers and prognostic variables^[14].

Two differences are worth planning for. First, the JCA practical guideline notes that analysing comparators separately or together can produce different networks and sometimes different estimates, and recommends reporting both a population-level network and any comparator-level networks^[15]. NICE's fully incremental model favours one network containing all comparators. Second, for population-adjusted or non-randomised estimates, the HTACG guideline describes testing against a shifted null hypothesis, so that the confidence interval must lie beyond a prespecified threshold; the threshold is left to Member States^[8].

Worked example: what an indirect comparison costs in precision

Suppose the new treatment A has been compared with B in an RCT, with a hazard ratio (HR) of 0.70 and a standard error of the log HR of 0.10. A comparator C has also been compared with B, with an HR of 0.85 and a standard error of 0.12. The Bucher anchored comparison of A with C is:

log HR(A vs C) = log 0.70 − log 0.85 = −0.357 − (−0.163) = −0.194

SE = sqrt(0.10^2 + 0.12^2) = 0.156

The indirect HR is 0.82, with a 95% confidence interval of 0.61 to 1.12. The randomisation within each trial is preserved, but the interval now crosses 1. For NICE, that uncertainty flows into the probabilistic sensitivity analysis and the committee's view of the ICER. For the JCA, it is reported as relative effect and certainty, and if a population-adjusted or non-randomised estimate were tested against a shifted null, an even larger effect would be needed to exclude the threshold^[8]. The same calculation, documented once, serves both audiences; what differs is how each one uses the uncertainty.

Timing and data cut-offs

The JCA runs in parallel with the European Medicines Agency procedure. The scope is finalised at the latest 10 days after the CHMP list of questions, the dossier is due within 100 days of the first request, and the revised draft reports must be finalised at the latest on the date of the marketing authorisation decision^[11]. The Q&A requires results at the most recent prespecified data cut-off, with the last available overall survival cut where different, and a literature search cut-off no more than three months before dossier submission^[13].

NICE also works close to licensing: its manual states that NICE may issue draft guidance or final draft guidance before the technology receives final UK regulatory approval^[12]. During the appraisal, clarification questions are sent within 21 days of the submission and companies have 14 days to respond^[12].

In practice the two submissions are likely to fall within months of each other, on data cuts that may differ. A single evidence plan should fix one primary data cut for the core analyses, schedule SLR refreshes to satisfy the JCA three-month rule, and document in advance how later cuts will be handled in each process.

The role of economic evidence

Economic evidence is excluded from the JCA by design^[1], but it remains the centre of the NICE decision. Since March 2026, NICE's manual states that below a most plausible ICER of £25,000 per QALY gained the decision is normally based on the cost-effectiveness estimate, that between £25,000 and £35,000 the committee makes explicit reference to factors such as uncertainty and uncaptured benefits, and that above £35,000 an increasingly stronger case is needed^{[4, 16]}. Net health benefit is to be presented at £25,000 and £35,000 per QALY, with and without QALY weighting^[4]. The severity modifier applies a QALY weight of 1.2 where proportional shortfall is 0.85 to 0.95 or absolute shortfall 12 to 18, and 1.7 at or above 0.95 or 18^[4].

For example, a technology with an incremental gain of 0.40 QALYs and an incremental cost of £12,000 has a net monetary benefit of 0.40 × 25,000 − 12,000 = −£2,000 at the lower bound and 0.40 × 35,000 − 12,000 = +£2,000 at the upper bound. If the condition qualifies for a 1.2 weight, the weighted gain is 0.48 QALYs, and the net monetary benefit at £25,000 rises to zero.

Why does this matter for a joint plan? Because the clinical inputs that drive the ICER, the relative effect estimates, the extrapolated survival curves and the utility values, rest on the same SLR and ITC outputs that feed the JCA. A choice made for the JCA, such as which network or covariate set is primary, will be visible to NICE, and the HTD dossier for a JCA is published alongside the report^[17]. The economic model should therefore be built on the same evidence base, with any differences explained.

Consistency

The JCA report, its summary and the developer's dossier are published by the Commission. Any difference between the evidence presented there and the evidence presented to NICE should be deliberate, documented and explainable.

Designing one global SLR and ITC programme

We recommend building the programme around a superset question and modular outputs. The following design principles are our own synthesis of both sets of requirements.

1. Superset PICO: Define eligibility to capture every comparator plausible in any EU system and in the NHS, all outcomes in the HTACG outcomes guidance and the inputs needed for the economic model. Narrow at the analysis stage, not the search stage.
2. One protocol, two appraisal tools: NICE asks for a systematic review using a predefined protocol^[9]. The HTACG validity guidance recommends RoB 1 for RCTs and ROBINS-I for non-randomised studies^[10]. Running RoB 1 alongside RoB 2 at extraction avoids rework.
3. Living update plan: Schedule refreshes so the JCA search cut-off falls within three months of submission^[13], and reuse the same refresh for NICE.
4. Covariate review: The Q&A requires an SLR to identify covariates for adjusted comparisons^[13]. Pair it with structured clinical input, because the practical guideline cautions that statistical tests for effect modification should not be used in isolation^[15].
5. Methods ladder per comparison: For each comparator, prespecify the preferred route (direct, anchored NMA or Bucher, population-adjusted anchored, unanchored, or evidence gap) and the fallback, with statistical analysis plans written before the JCA scope is final.
6. Dual network outputs: Produce both the full network required for NICE's incremental analysis and comparator-level networks for JCA PICOs, with a reconciliation note where estimates differ^{[9, 15]}.
7. Shared code and data: Use a single validated code base, with access to individual patient data secured early for population adjustment^[14].

Governance of evidence across markets

A joint plan needs a governance model that treats evidence as a controlled asset. In our experience the elements that matter most are:

• An evidence council with global HEOR, biostatistics, medical and the lead markets, which signs off the superset PICO, the methods ladder and any change to a primary analysis.
• A single source of truth for SLR data, analysis code and outputs, version-controlled and traceable to each submission.
• A divergence log that records every place where the NICE and JCA submissions differ, the reason, and who approved it.
• Market feedback loops: National processes may refer to the JCA. The G-BA, for example, notes that companies may refer to a JCA already submitted under the Regulation^[18]. Questions raised in one process should be reviewed for their implications in others.
• Method change monitoring: NICE's manual has been updated several times since 2022, including the threshold change in March 2026^[19], and the HTACG continues to publish guidance and Q&A. A named owner should track both.

What this means for evidence planning

• Start from the superset: One broad SLR and one covariate review, designed for the JCA's inclusive scope, will usually also cover NICE's needs.
• Design trials for both audiences: Comparator choice, disease-specific quality of life plus EQ-5D-5L, and prespecified subgroups serve both processes at once.
• Prespecify the ITC programme: Both frameworks prefer anchored comparisons and expect transparent, prespecified methods^{[8, 9]}.
• Keep economics downstream but aligned: The JCA excludes economics, but the NICE model must rest on the same clinical estimates, and differences should be documented.
• Govern for consistency: Published JCA dossiers make cross-market consistency visible. A divergence log and an evidence council are inexpensive protection.

Evidaxia designs and delivers integrated evidence programmes for UK and EU submissions. See our EU JCA dossier service, our support for UK HTA submissions, and our systematic literature review work built for multiple agencies.

References

1. Regulation (EU) 2021/2282 of the European Parliament and of the Council of 15 December 2021 on health technology assessment and amending Directive 2011/24/EU. Official Journal of the European Union. 2021 Dec 22;L458:1-32. Available from: https://eur-lex.europa.eu/legal-content/EN/TXT/HTML/?uri=CELEX:32021R2282
2. National Institute for Health and Care Excellence. NICE health technology evaluations: the manual (PMG36). Section 2: The scope [Internet]. London: National Institute for Health and Care Excellence; 2022 [updated 2026; cited 2026 Oct 30]. Available from: https://www.nice.org.uk/process/pmg36/chapter/the-scope
3. National Institute for Health and Care Excellence. NICE health technology evaluations: the manual (PMG36). Section 4: Economic evaluation [Internet]. London: National Institute for Health and Care Excellence; 2022 [updated 2026; cited 2026 Oct 30]. Available from: https://www.nice.org.uk/process/pmg36/chapter/economic-evaluation
4. National Institute for Health and Care Excellence. NICE health technology evaluations: the manual (PMG36). Section 6: Committee recommendations [Internet]. London: National Institute for Health and Care Excellence; 2022 [updated 2026; cited 2026 Oct 30]. Available from: https://www.nice.org.uk/process/pmg36/chapter/committee-recommendations
5. Member State Coordination Group on HTA. Guidance on the scoping process [Internet]. Brussels: European Commission; 2024 Nov 28 [cited 2026 Oct 30]. Available from: https://health.ec.europa.eu/document/download/7be11d76-9a78-426c-8e32-79d30a115a64_en?filename=hta_jca_scoping-process_en.pdf
6. Member State Coordination Group on HTA. Guidance on outcomes for joint clinical assessments [Internet]. Brussels: European Commission; 2024 Jun 10 [cited 2026 Oct 30]. Available from: https://health.ec.europa.eu/document/download/a70a62c7-325c-401e-ba42-66174b656ab8_en?filename=hta_outcomes_jca_guidance_en.pdf
7. 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/statement-on-using-the-eq-5d-5l-value-set
8. Member State Coordination Group on HTA. Methodological guideline for quantitative evidence synthesis: direct and indirect comparisons [Internet]. Brussels: European Commission; 2024 Mar 8 [cited 2026 Oct 30]. Available from: https://health.ec.europa.eu/document/download/4ec8288e-6d15-49c5-a490-d8ad7748578f_en?filename=hta_methodological-guideline_direct-indirect-comparisons_en.pdf
9. National Institute for Health and Care Excellence. NICE health technology evaluations: the manual (PMG36). Section 3: Evidence [Internet]. London: National Institute for Health and Care Excellence; 2022 [updated 2026; cited 2026 Oct 30]. Available from: https://www.nice.org.uk/process/pmg36/chapter/evidence
10. Member State Coordination Group on HTA. Guidance on validity of clinical studies, version 1.0 [Internet]. Brussels: European Commission; 2024 Sep 19 [cited 2026 Oct 30]. Available from: https://health.ec.europa.eu/document/download/9f9dbfe4-078b-4959-9a07-df9167258772_en?filename=hta_clinical-studies-validity_guidance_en.pdf
11. Commission Implementing Regulation (EU) 2024/1381 of 23 May 2024 laying down, pursuant to Regulation (EU) 2021/2282 on health technology assessment, procedural rules for the interaction during, exchange of information on, and participation in, the preparation and update of joint clinical assessments of medicinal products for human use at Union level, as well as templates for those joint clinical assessments. Official Journal of the European Union. 2024 May 24;L series. Available from: https://eur-lex.europa.eu/legal-content/EN/TXT/HTML/?uri=CELEX:32024R1381
12. National Institute for Health and Care Excellence. NICE health technology evaluations: the manual (PMG36). Section 5: Developing the guidance [Internet]. London: National Institute for Health and Care Excellence; 2022 [updated 2026; cited 2026 Oct 30]. Available from: https://www.nice.org.uk/process/pmg36/chapter/developing-the-guidance
13. Member State Coordination Group on HTA. Questions and answers on general methodological and procedural issues for joint clinical assessments [Internet]. Brussels: European Commission; 2026 May [cited 2026 Oct 30]. Available from: https://health.ec.europa.eu/document/download/b89dc548-10d7-4ef7-9c3b-deea90b639ca_en?filename=hta_20260518_QA_methodological_procedural_questions_HTACG_en.pdf
14. Phillippo DM, Ades AE, Dias S, Palmer S, Abrams KR, Welton NJ. Methods for population-adjusted indirect comparisons in health technology appraisal. Med Decis Making. 2018;38(2):200-11. Available from: https://eprints.whiterose.ac.uk/136613/1/0272989x17725740.pdf
15. Member State Coordination Group on HTA. Practical guideline for quantitative evidence synthesis: direct and indirect comparisons [Internet]. Brussels: European Commission; 2024 Mar 8 [cited 2026 Oct 30]. Available from: https://health.ec.europa.eu/document/download/1f6b8a70-5ce0-404e-9066-120dc9a8df75_en?filename=hta_practical-guideline_direct-and-indirect-comparisons_en.pdf
16. National Institute for Health and Care Excellence. Changes to NICE's cost-effectiveness thresholds. Frequently asked questions [Internet]. London: National Institute for Health and Care Excellence; 2026 [cited 2026 Oct 30]. Available from: https://www.nice.org.uk/what-nice-does/faqs/changes-to-nice-s-cost-effectiveness-thresholds
17. European Commission. Joint clinical assessment report on tovorafenib (Ojemda): report, summary report, factual accuracy annex and health technology developer dossier [Internet]. Brussels: European Commission; 2026 Jun 9 [cited 2026 Oct 30]. Available from: https://health.ec.europa.eu/publications/joint-clinical-assessment-report-tovorafenib-ojemda_en
18. Gemeinsamer Bundesausschuss. The benefit assessment of medicinal products in accordance with the German Social Code, Book Five (SGB V), section 35a. G-BA [Internet]. Berlin: Gemeinsamer Bundesausschuss; [date unknown] [cited 2026 Oct]. Available from: https://www.g-ba.de/english/benefitassessment/
19. National Institute for Health and Care Excellence. NICE health technology evaluations: 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

Sources accessed October 2026. This article is analysis for information and is not legal, regulatory or tax advice.