Each EU member state contributes to the scope of a Joint Clinical Assessment, and for oncology products this consolidated input frequently produces between 10 and 30 or more distinct PICOs, each specifying a population, intervention, comparator, and outcome combination that the manufacturer's dossier must address. The pivotal trial almost never directly covers this range of comparators and subgroups. Once the final PICO scope is communicated, manufacturers face a short, intense window, typically 100 days, to address every defined PICO with synthesised, methodologically defensible evidence. [1] This is the operational reality that JCA evidence synthesis AI exists to address.
At Pienomial, we built KnolAI as our enterprise AI research platform specifically to help HEOR teams manage this PICO proliferation problem, and ISPOR's US 2026 conference confirmed that the broader industry is moving in the same direction: away from isolated AI experimentation and toward AI-enabled evidence infrastructure as the standard way JCA evidence synthesis gets done. [2] This post explains how AI supports every stage of JCA evidence synthesis, from PICO anticipation before the formal scope is even communicated, through to the final dossier evidence sections, and how KnolAI delivers this capability through the Knolens platform.[9]
1. The PICO Proliferation Problem: Why JCA Evidence Synthesis Is Structurally Harder
The fundamental complexity of JCA evidence synthesis compared to a single national HTA submission is the sheer multiplicity of evidence questions a single dossier must answer. Because the JCA scoping process consolidates input from every contributing EU member state, the resulting PICO list reflects the diverse clinical practice patterns, comparator preferences, and subgroup priorities across the entire EU, not the single national context that a NICE or G-BA submission addresses individually.[1]
This means evidence alignment and gap identification must happen at a scale that manual analysis struggles to keep pace with: existing clinical data must be strategically mapped and synthesised to address every defined PICO, and identifying and robustly addressing the evidence gaps, especially for the indirect comparisons that most PICOs will require, is the central HEOR challenge within the tight JCA timeline. [1] Each of the 10 to 30 PICOs may require a different indirect treatment comparison, a different subgroup analysis, or a different RWE supplement, and the HEOR team must determine the right evidence response for each one within the 100-day submission window.[5]
The most consequential strategic insight in JCA preparation is that formal scope communication is not the starting point for evidence synthesis. It is the deadline. Given the complexity of the JCA and the expected high number of PICOs, health technology developers have to start preparing the JCA submission dossier prior to the communication of the assessment scope, ideally including the evidence syntheses and results for all anticipated PICOs before the formal 100-day clock even begins. [5]
This requires PICO anticipation: predicting, with reasonable confidence, which populations, interventions, comparators, and outcomes the consolidated EU member state input is likely to generate. PICO anticipation can be informed by previous JCA and national HTA assessments in the same indication, national treatment guidelines across the contributing member states, and published information about each member state's HTA system and historical comparator preferences. [5] This is precisely the kind of large-scale precedent analysis that benefits from AI assistance: synthesising patterns across dozens of prior assessments and treatment guidelines across multiple countries is a multi-week manual undertaking that AI can compress substantially while improving the consistency of the analysis.[6]
KnolAI performs PICO anticipation by querying the Knolens knowledge graph for prior JCA assessments, national HTA decisions, and treatment guidelines across the EU member states relevant to the indication, identifying the comparator patterns, subgroup priorities, and outcome expectations that are likely to recur in the formal PICO scope. This anticipated PICO map becomes the evidence synthesis planning document that guides which ITCs and subgroup analyses to begin preparing before the formal scope arrives.[9]
3. The Uncertainty Metric Approach: Structuring Evidence Synthesis Planning
A 2025 ISPOR poster proposed a structured uncertainty metric to guide PICOS assessment and evidence synthesis planning for JCA submissions, organised around five pillars: indication and subpopulations, type of intervention and comparators, societal and patient unmet need, type and quality of evidence source informing evidence synthesis, and methodology of evidence synthesis. [6] This framework provides a systematic way to assess which anticipated PICOs carry the highest uncertainty, and therefore the highest evidence generation priority, rather than treating all anticipated PICOs as equally urgent.
KnolAI operationalises this uncertainty-based prioritisation by scoring each anticipated PICO against the five pillars using the precedent data available in the Knolens knowledge layer: how much uncertainty exists about the likely comparator for this population, how strong is the available evidence source for this outcome, and how methodologically complex would the required evidence synthesis be. The HEOR team receives a prioritised PICO list, not just an anticipated list, enabling resource allocation toward the PICOs that carry the most submission risk if left unaddressed.[9]
4. AI-Assisted Indirect Treatment Comparisons Across Multiple PICOs Simultaneously
Once the formal PICO scope is communicated, the central evidence synthesis task is producing the indirect treatment comparisons that address every PICO where the pivotal trial's comparator does not match the PICO's specified comparator. With potentially dozens of PICOs each requiring a comparison against a different comparator or in a different subpopulation, manually constructing each ITC from scratch is not feasible within the 100-day window.[1]
KnolAI's evidence synthesis workflow constructs ITCs at scale by identifying the shared trial network that underlies multiple related PICOs and generating the comparator-specific analysis for each PICO from that shared network, rather than treating each PICO as an isolated analysis project. For PICOs sharing the same population and outcome but specifying different comparators, the underlying evidence network and much of the data extraction work is common, and KnolAI structures the synthesis to exploit this overlap, generating PICO-specific outputs efficiently rather than duplicating the full evidence synthesis workflow for every individual PICO.[9]
For PICOs where no trial network can support a valid indirect comparison, KnolAI flags the evidence gap explicitly, consistent with the EU HTA dossier template guidance requiring that the methods used in dossier preparation follow international evidence-based medicine standards and that any quantitative synthesis limitation be documented transparently rather than concealed. [4] An honestly documented evidence gap, with a clear explanation of why no valid comparison exists, is methodologically stronger than a forced comparison that does not satisfy transitivity assumptions.[7]
5. RWE Integration for PICOs the Trial Network Cannot Address
A significant proportion of JCA PICOs cannot be addressed through trial-based evidence alone, particularly where the specified comparator was never directly or indirectly tested in any registered trial, or where the population is a rare subgroup that trial enrollment did not adequately capture. For these PICOs, real-world evidence becomes the primary evidence response, and the manufacturer must begin RWE study design and execution well before the formal PICO scope arrives, because RWE generation timelines cannot be compressed into a 100-day window.[7]
KnolAI's PICO anticipation analysis specifically flags the PICOs most likely to require RWE supplementation based on the precedent pattern of comparators and subgroups that prior JCA assessments and national HTA bodies have required without corresponding trial evidence. This early flagging gives the HEOR and clinical development teams the lead time to commission external control arm studies, additional endpoint collection, or extended follow-up studies in broader patient populations before the formal scope locks in the requirement.[9]
6. Building the JCA Dossier Evidence Sections with Full Methodology Documentation
The EU HTA dossier template guidance is explicit that the methods used in evidence synthesis must follow international standards of evidence-based medicine and that the underlying documentation, including clinical study reports, protocols, statistical analysis plans, and software and program code used in any analysis, must be provided in the dossier appendices. [4] This is a methodology transparency requirement that AI-assisted evidence synthesis must satisfy fully, not approximate.
KnolComposer, working from the KnolAI evidence synthesis outputs, generates the JCA dossier clinical effectiveness sections with the full methodology documentation chain required by Appendix D of the template: the search strategy for each evidence synthesis, the statistical methodology applied for each ITC or NMA, the software and code references, and the source attribution for every clinical claim down to the specific study report or publication location. Because every output is generated from the Knolens knowledge graph's sourced entity-relationship triples, this documentation chain exists by construction rather than requiring separate compilation after the analysis is complete.[9]
Joint Scientific Consultation is the formal early advice mechanism under the EU HTA Regulation, enabling manufacturers to seek guidance from EU HTA bodies during clinical study planning on the evidence likely to be needed for future JCAs. Used strategically, JSC helps strengthen the evidence base relied upon for JCA and minimises the risk of needing reactive changes once the formal PICO scope is issued.[3]
The quality of JSC engagement depends directly on the quality of the PICO anticipation and evidence gap analysis the manufacturer brings to the consultation. A manufacturer who arrives with a structured, evidence-grounded anticipated PICO map and a documented evidence gap assessment receives substantive, actionable feedback on specific protocol design choices. A manufacturer who arrives without this preparation receives generic procedural guidance. KnolAI's PICO anticipation and uncertainty scoring directly support JSC preparation, giving the clinical development and HEOR teams a structured, precedent-grounded basis for the specific questions they bring to the consultation.[3]
8. Why AI-Powered PICO Prediction Requires Human Strategic Validation
AI-powered PICO prediction services are an increasingly recognised category in the JCA preparation market, with services explicitly combining AI-driven analysis with therapeutic and payer insight validation. [8] The consistent industry guidance is that effective PICO prediction requires more than automated analog analysis: it requires grounding in real-world treatment patterns, validation against market insight, and integration with the broader submission strategy, rather than relying solely on automated outputs.[8]
KnolAI is built around exactly this principle. The AI-generated PICO anticipation and uncertainty scoring is the structured starting point for the HEOR and market access team's strategic review, not a replacement for it. Our platform surfaces the precedent patterns, the evidence gaps, and the uncertainty priorities systematically and at a scale no manual review could match, and the HEOR team applies their therapeutic area expertise and payer relationship knowledge to validate, refine, and finalise the evidence synthesis strategy. This is the same human-in-the-loop principle that governs every KnolAI workflow: AI accelerates the mechanical and analytical breadth, human expertise applies the final strategic judgement.[9]
9. How Fast Can Your Team Build JCA Evidence Synthesis Capability with KnolAI?
Given that JCA preparation must begin well before the formal PICO scope is communicated, deploying KnolAI's JCA evidence synthesis AI capability early in a product's Phase II to Phase III transition delivers the most value. KnolAI is a pre-built capability within the Knolens platform, configurable to your indication from the first sprint.[9]
Sprint 1, Weeks 1 to 2, PICO anticipation map generated: KnolAI queries the Knolens knowledge graph for prior JCA assessments, national HTA decisions, and treatment guidelines across relevant EU member states for your indication. The anticipated PICO map is generated, identifying the likely comparators, subgroups, and outcomes across the consolidated EU scope before the formal scope is communicated.
Sprint 2, Weeks 3 to 4, Uncertainty scoring and evidence gap prioritisation: Each anticipated PICO is scored against the five-pillar uncertainty framework, producing a prioritised evidence generation plan. Evidence gaps requiring RWE supplementation are flagged with feasibility and timeline assessments, giving the clinical development team the lead time RWE generation requires.[6]
Sprint 3, Weeks 5 to 6, ITC synthesis and dossier section drafting begin: KnolAI begins constructing the shared evidence network and PICO-specific ITC analyses for the highest-priority anticipated PICOs. KnolComposer generates draft dossier clinical effectiveness sections with full methodology documentation, ready for refinement once the formal PICO scope confirms or adjusts the anticipated picture.[4]
Conclusion
The EU Joint Clinical Assessment has fundamentally changed the evidence synthesis challenge facing pharma HEOR teams: not one evidence question to answer, but potentially dozens, consolidated from every contributing EU member state, addressed within a 100-day window that begins only after the formal scope is communicated. The manufacturers who succeed under this model are those who have already done the evidence synthesis work for the most likely PICOs before the clock starts.
At Pienomial, we built KnolAI's JCA evidence synthesis AI capability because we believe that PICO anticipation, uncertainty-prioritised evidence planning, and scaled ITC synthesis are not optional efficiency gains for JCA preparation. They are the only way to genuinely meet the JCA timeline without compromising the evidence quality and methodology transparency that the EU HTA dossier template requires. [9] CTA: See how KnolAI accelerates your JCA evidence synthesis. Book a demo with the Pienomial team.
