Consider a scenario that plays out with regularity in HTA review cycles: a Phase III trial demonstrates a meaningful improvement in progression-free survival in a biomarker-selected population. The data is robust, the p-value is strong, and the regulatory submission sails through. Then the HTA body asks: what happens in the broader real-world patient population? What does long-term overall survival look like beyond the two-year trial follow-up? What are the actual treatment patterns and discontinuation rates once the product is used outside a controlled trial setting? The sponsor cannot answer these questions from trial data alone. Without credible real-world evidence to fill these gaps, the reimbursement decision is delayed, restricted to a narrower population than the trial enrolled, or made conditional on post-market data generation.
This is the RWE evidence gap. It is not a gap in clinical science. It is a gap in evidence architecture, and it is one that HTA bodies across NICE, G-BA, HAS, and ICER are now actively demanding be filled. Pienomial's Knolens platform, as a purpose-built life sciences AI platform, enables HEOR teams to identify, synthesise, and present real-world evidence in a structured, traceable, HTA-ready format, closing the gap before payers ask for it rather than after. [8]
1. Why HTA Bodies Are Demanding RWE in 2026
Real-world evidence is ranked number two on ISPOR's 2026 to 2027 Top 10 HEOR Trends report, with the focus now firmly on data quality, transparency, and reproducibility rather than simply access to data.[1] The EMA's third annual RWE framework report, covering February 2024 to February 2025, documents 59 RWE studies conducted across the reporting period, representing a 47.5 percent increase from the previous year, spanning drug utilisation, safety, and disease epidemiology. [3] This expansion is not incidental. It reflects a structural shift in how regulators and HTA bodies view the relationship between trial evidence and real-world practice.
Three specific drivers are accelerating RWE demand from HTA bodies. First, accelerated approval pathways such as FDA Accelerated Approval and EMA Conditional Marketing Authorisation are producing approvals based on surrogate endpoints with limited long-term outcome data. HTA bodies require RWE to confirm or challenge the long-term benefit that the trial surrogate predicts. Second, trial populations are narrow. Biomarker-selected, protocol-defined, and performance-status-restricted trial populations do not reflect the full range of patients who will receive the product in clinical practice. Payers want outcome data for their specific population. Third, the surrogate-to-outcome correlation problem. PFS, biomarker response, and similar surrogates are not sufficient for many HTA bodies in isolation. RWE that links trial surrogates to patient-relevant outcomes including OS, quality of life, and hospitalisation is increasingly required.[2]
2. The Four Most Common RWE Evidence Gaps in HTA Submissions
Understanding which RWE gaps most commonly drive negative or restricted HTA outcomes enables HEOR teams to build targeted evidence generation strategies rather than conducting broad, unfocused RWE programmes.[4]
Gap 1, Long-term survival beyond trial follow-up: The most frequent single cause of HTA uncertainty is limited overall survival data. Survival extrapolation models are required but are routinely challenged by HTA bodies as speculative when the trial follow-up is short relative to the expected disease course. RWE from linked registries and claims databases that tracks long-term survival in the treated population outside trial conditions provides the empirical anchor that extrapolation models require.
Gap 2, Broader population effectiveness: Trials enrol selected populations. Real-world prescribing is broader, covering patients who were excluded by performance status, prior therapy requirements, or comorbidity thresholds. Payers want to know what outcomes look like in the full prescribing population, including the subgroups underrepresented or excluded in trials.
Gap 3, Comparative effectiveness against the real-world standard of care: The trial comparator may not reflect current clinical practice. In a rapidly evolving therapeutic area, the active comparator in a trial completed three years ago may already have been superseded by newer options in clinical practice. RWE-based propensity-score matched or adjusted comparative analyses against the actual real-world standard of care address this gap.
Gap 4, Treatment patterns, adherence, and discontinuation: Payers and health systems want to understand how the product is used in practice, covering dose modifications, treatment duration, discontinuation rates, and re-treatment patterns. Claims-based treatment pattern analyses provide this evidence in a format that payers can use directly in formulary modelling.[5]
3. How AI-Powered RWE Synthesis Works in Practice
The core challenge of RWE for HTA submissions is not the absence of data. It is the heterogeneity of available data sources and the analytical and documentation requirements that make RWE credible to HTA reviewers. Claims databases, electronic health records, disease registries, observational studies, and patient-reported outcome datasets each have different structures, different quality profiles, different bias patterns, and different levels of acceptance by different HTA bodies.
Manual synthesis across these sources is slow, inconsistent, and rarely produces the level of methodology documentation that NICE, G-BA, and HAS now require. KnolAI, the research module of the Knolens life sciences AI platform, ingests and structures RWE across all relevant source types. It conducts automated bias assessment using validated instruments including ROBINS-I for non-randomised studies of interventions, documents confounding adjustment methodology, and generates PICOTS-structured RWE summaries aligned to each HTA body's evidence standards. Every claim in the RWE synthesis is attributed to a specific data source, analysis, and publication. There is no unsourced assertion in the output.[8]
The output format is HTA-body-specific. For a NICE submission, KnolAI generates RWE summary tables in the format expected by the NICE technical team, with explicit uncertainty characterisation and limitation documentation. For a G-BA dossier, RWE is structured around patient-relevant endpoint categories with IQWiG methodology documentation. For ICER, US-specific population data and cost implications are incorporated.
4. Connecting RWE to the Payer Value Story
RWE is most persuasive when it directly addresses the specific uncertainty driving a payer's reimbursement hesitation, not when it provides a broad portrait of product use in the real world. The most effective RWE evidence strategies are built backwards from the payer question, not forwards from available data.
KnolAI identifies the specific uncertainties that drove negative or conditional outcomes in prior HTA decisions for analogous products in the same indication. This analysis, drawn from structured knowledge of HTA precedent decisions, identifies the RWE questions that matter most for the target payer population. The output is a prioritised RWE evidence agenda mapped to specific HTA uncertainty categories, which becomes the brief for the RWE programme rather than a general data landscape review.[4]
As a competitive intelligence in healthcare layer, the Knolens knowledge base also identifies competitor RWE programmes in the same indication, including data sources used, study designs registered, and any published outcomes from competitor real-world analyses. For a HEOR team planning an RWE strategy, this competitive intelligence is as important as understanding what the HTA body requires, because the RWE evidence bar is often set by what the leading product in the indication has already demonstrated in real-world practice.
5. Propensity Score Matching and Adjusted Indirect Comparisons with RWE
When head-to-head trial data against the real-world standard of care is unavailable, adjusted indirect comparisons using RWE are the primary mechanism for demonstrating comparative effectiveness to payers. The methodological requirements are specific, and failures in documentation are the most common reason RWE comparative analyses are rejected or downgraded by HTA reviewers.
Acceptable methods include propensity score matching, requiring covariate balance documentation and standardised mean difference reporting across all matched covariates; inverse probability of treatment weighting, requiring weight distribution documentation and trimming methodology; and model-adjusted analyses, requiring covariate selection rationale, missing data handling documentation, and sensitivity analysis results. ISPOR's RWE Transparency Initiative has published harmonised protocol templates specifically to support the pre-specification and transparent reporting of RWE comparative analyses, and HTA bodies are increasingly requiring registration of RWE study protocols before results are submitted.[7]
KnolAI structures RWE comparative analysis inputs by identifying available databases for the indication, assessing covariate availability for propensity score construction, and generating a structured methodology document for the chosen approach. The output is a reproducible, pre-specified, documented analysis that satisfies the transparency requirements of NICE, G-BA, and ICER reviewers.
6. Audit-Ready RWE: The Documentation Standard HTA Bodies Now Require
The shift in HTA body attention from RWE access to RWE quality and transparency, identified by ISPOR as the defining theme of the number two trend in 2026 to 2027, [1] has concrete implications for the documentation standard that RWE submissions must meet.
Four documentation requirements now define audit-ready RWE for HTA submissions. First, data source documentation: which database, covering which years, which patient population definition, which inclusion and exclusion criteria, and how many patients at each stage. Second, analysis methodology documentation: the statistical approach, covariate selection rationale, sensitivity analyses conducted, and limitation acknowledgement. Third, AI tool documentation: where AI was used in RWE synthesis or analysis, NICE requires disclosure of the tool, method, and human oversight, consistent with its 2024 position statement on AI in evidence generation. [6] Fourth, traceability: every RWE result cited in dossier text must link to a specific analysis output, which links to a specific data source and methodology document.
Knolens generates a complete audit trail for all KnolAI RWE synthesis actions, with source attribution at the dataset, study, and result level. Methodology documentation is auto-generated and linked directly to the dossier text sections where RWE results appear, satisfying the documentation chain that NICE, G-BA, and HAS reviewers now routinely check.[8]
7. Building a Prospective RWE Data Generation Strategy
The most credible RWE for HTA submissions comes from prospectively planned studies, not post-hoc analyses of whichever databases happen to be available at submission time. A retrospective search for conveniently available data is one of the patterns that HTA reviewers recognise and penalise in their uncertainty assessments. Pre-specified, registered RWE studies with documented protocol, defined endpoints, and pre-committed analysis methodology carry substantially higher evidential weight.
KnolAI supports prospective RWE planning by identifying the specific evidence gaps that will be scrutinised by each target HTA body, based on precedent analyses of prior decisions for analogous products. It then assesses which available RWE data sources can address each gap with existing data and which require prospective data collection. The output is a ranked RWE data generation plan with estimated feasibility timelines, data source quality assessments, and a prioritisation framework that reflects the relative importance of each gap to the target payer's reimbursement decision.
The timing requirement is critical. Prospective RWE studies commissioned at Phase III initiation can provide two to three years of real-world follow-up data by the time of HTA submission. Studies commissioned post-approval, in response to HTA body requests, arrive too late to influence the initial reimbursement decision and are typically positioned as managed access agreement conditions rather than primary evidence.
8. RWE for Outcomes-Based Contracts
Outcomes-based contracts are expanding beyond their traditional stronghold in Italy and moving into Germany, England, and increasingly the US commercial payer market. These contracts tie reimbursement levels to demonstrated real-world outcomes, creating a direct link between the quality of a sponsor's RWE infrastructure and the commercial terms they can negotiate with payers.
Effective outcomes-based contract design requires pre-specified, measurable real-world outcome endpoints; agreed data sources and analysis methodology; a continuous monitoring infrastructure that tracks outcomes over the contract period; and a pre-agreed adjustment mechanism when outcomes differ from trial predictions. Each of these requirements is an RWE infrastructure requirement, not just a contracting question.
Knolens enables outcomes-based contract design and performance monitoring through its enterprise knowledge & AI memory platform architecture. KnolAI identifies which real-world outcomes are measurable in available data sources for the indication and patient population, models the likely outcome distribution based on available RWE, and generates the evidence specification document that forms the basis of the OBC negotiation. The Knolens monitoring layer then tracks real-world outcomes data continuously against pre-specified OBC metrics, alerting the market access team when performance data requires reporting.
9. How Fast Can You Build an RWE Intelligence Layer with Knolens?
Building a credible RWE intelligence layer does not require months of custom data engineering. Knolens ships with pre-built RWE synthesis pipelines, pre-configured bias assessment frameworks including ROBINS-I, and ready-to-run PICOTS-structured output templates for NICE, G-BA, HAS, and ICER. Your team is not building from scratch. You are configuring a product that is already built.
Most HEOR teams are running their first live RWE evidence output within six weeks of onboarding. Here is what that looks like in practice.
Sprint 1, Weeks 1 to 2, RWE landscape live: Knolens is connected to your indication and patient population. KnolAI runs the first RWE source mapping across available claims databases, registries, and observational studies. Your team receives a structured RWE landscape report showing which sources exist, what they cover, and which of the four common HTA evidence gaps each source can address. No data engineering required from your side.
Sprint 2, Weeks 3 to 4, Gap prioritisation and first synthesis: KnolAI generates a prioritised RWE evidence agenda for each target HTA body, identifying which gaps are addressable with existing data now and which require prospective study commissioning. Where existing data is available, the first AI-powered RWE synthesis runs with full methodology documentation and source attribution. [8]
Sprint 3, Weeks 5 to 6, HTA-ready output and monitoring configured: RWE summary tables are formatted to NICE, G-BA, and HAS submission requirements. Audit trail documentation is generated automatically. The Knolens monitoring layer is configured to track new RWE publications in the indication, competitor RWE study registrations, and registry data updates on an ongoing basis.
From Sprint 3 onward, Knolens operates as a living RWE intelligence layer. For prospective studies identified in Sprint 2, Knolens generates the protocol specification document and tracks study registration on ISPOR's RWE Transparency Initiative registry or EU-PAS. As new RWE evidence becomes available, the evidence layer updates automatically and flags any changes relevant to existing HTA submissions or managed access agreement reviews. [7]
Final Thought
Real-world evidence is no longer a supplementary element of HTA submissions. It is a core requirement that payers use to assess whether the value demonstrated in a controlled trial translates to the populations and settings they are responsible for. The HEOR teams achieving the best reimbursement outcomes in 2026 are those with a proactive, structured RWE evidence strategy built before the HTA body requests it, not those scrambling to identify available data sources after submission.[1]
Pienomial's Knolens platform provides the life sciences AI platform and market access analytics platform infrastructure to identify, synthesise, and present RWE in a traceable, audit-ready, HTA-body-specific format. From evidence gap mapping through OBC performance monitoring, every RWE output is structured for payer scrutiny and built on a governed knowledge layer that updates continuously as the evidence landscape evolves.CTA: Explore Pienomial's HEOR and Market Access RWE Solutions.
