The EU joint clinical assessment dossier, PICO by PICO, with six agents doing the mapping, drafting and checking
Every EU joint assessment scope becomes a cited evidence plan and a full draft dossier by day 39 of 100.
See one case, screen by screen ↓Day 0 is the day the assessment scope arrives. From then, the dossier team has the 90–100-day window to answer every PICO question the Member States asked for — each population, each comparator, each outcome — and upload a complete dossier to the HTA IT platform. A PICO with no evidence still needs an answer: an indirect comparison, or a justified gap.
Most of those days do not go on judgement. They go on reading the scope into a spreadsheet, hunting the right number for each PICO × outcome cell across clinical study reports and statistical outputs, working out which comparisons can be anchored, and then writing and re-checking sections in the Annex I template. A number copied one decimal wrong, or an indication worded differently from the SmPC draft, is found late — if it is found at all.
Estimated split from the scope to a full draft dossier, by hand and with the solution.
Six specialist agents read the scope, map the evidence, plan comparisons, draft and check; the statistician and dossier lead approve and sign.
Corventa (corventamab vedotin) in previously treated non-squamous NSCLC, JCA-26-014: assessor Netherlands, co-assessor Spain, six PICO questions, day 37 of 100. Here is one morning, screen by screen, in the working solution.
Lena opens the solution on day 37: 63 days left, submit by Wed Dec 9, and the next milestone — comparison plans signed off — on day 44, Oct 14. A map shows the requests of all 27 Member States; Germany asked for 4 PICOs. Only 14 of 48 PICO cells have comparative evidence so far, and 4 decisions wait for her: 2 comparison plans, 1 approach, 1 gap.
Needs you: “Approve the comparison plan · PICO 2 vs docetaxel + angiokinase inhibitor — Sam Patel recommends approval with a sensitivity analysis.”
The PICO evidence matrix puts a value in each cell — overall survival, PFS, response, quality of life and four safety outcomes — and marks it direct, indirect, planned, high uncertainty or gap. PICO 1 vs docetaxel is direct: OS HR 0.78 (0.65–0.94), median 13.1 vs 10.2 months, opened at Table 14.2.1.1 on page 612 of the CRV-301 study report. PICO 4 values come from subgroups outside the analysis plan and are labelled post hoc.
“Every value cites its table or publication.”
Seven states asked for Corventa against docetaxel plus an angiokinase inhibitor in adenocarcinoma. No trial compares them, so the analyst proposes a Bucher comparison anchored on docetaxel: the CRV-301 adenocarcinoma subgroup (n=548) against a 2014 published trial (n=658). It runs seven feasibility checks — same docetaxel dose, same histology and line, ECOG 0–1 in both, Schoenfeld p = 0.21 — fails one and flags follow-up (median 18.2 vs 31.7 months) as a caution.
Prior immunotherapy: “100% in CRV-301 vs 0% in the comparator trial — a likely effect modifier.”
Sam recommends approving the plan with a sensitivity analysis restricted to patients with at least 6 months of prior anti-PD-(L)1 therapy. Lena has three choices — approve the plan, approve with the sensitivity analysis, or declare an evidence gap so the dossier explains why — and a note for the record. She approves with the sensitivity analysis.
“Signing as Dr. Lena Ortiz · dossier lead · the run starts on approval.”
The plan is locked as v1.0 with Sam’s recommendation attached. The analyst pulls the inputs, the validated statistics engine runs Bucher on OS, PFS, response and four safety outcomes in 38 seconds, then the sensitivity analysis. A second program reproduces every estimate to three decimals, and six PICO 2 cells update in the matrix.
The analyst proposes only, and runs only plans that were approved.
Overall survival HR 0.92 (0.70–1.20); restricted to ≥6 months of prior anti-PD-(L)1, 0.95 (0.71–1.27); PFS 0.82 (0.61–1.09); grade ≥3 adverse events RR 0.80 (0.66–0.97). The decision is recorded on the audit trail: recommended by Sam Patel, decided by Dr. Lena Ortiz.
“All overall survival intervals cross 1: no statistically significant difference. Certainty low — no patient in the comparator trial had received immunotherapy.”
Lena presses Draft section 6.2. The writer reads PICO 2 from the scope and the approved results, writes six subsections — evidence base, overall survival, PFS and response, safety, quality of life, certainty — with 10 citations, and the consistency checker compares every number with the statistics engine output. Lena and Priya Raman, the medical writer, review it next.
Writer’s rule: never claim superiority where the interval crosses 1 — and describe uncertain results as uncertain.
The check compares every number with its cited table or output — 1,846 of 1,846 match — and the indication wording with SmPC draft v3.2. Section 3 says “advanced”; the SmPC draft says “locally advanced or metastatic”. Lena uses the SmPC wording in one click. Post hoc labels, a results section for every PICO, outcome definitions and abbreviations all pass.
The checker reports differences — it never edits silently.
With a clean check, Lena signs dossier v0.6 with her password. The signature records its meaning — “approved for internal review” — with her name and the time. Dr. Maya Chen reviews the results next; Omar Haddad, head of market access, signs the final dossier on day 96, Dec 5.
Five assessments with a clock running, sections drafted by day against plan, evidence coverage by PICO, Member State requests by PICO, and where every cited number comes from — statistical outputs, study reports, published trials and review, protocols, analysis plans and the SmPC.
The same assessment, seen by the four people who carry it — what their weeks looked like, and what they look like now.
Six specialist agents read the scope, map the evidence, plan comparisons, draft and check; the statistician and dossier lead approve and sign.
Reads the assessment scope from the assessors and turns it into structured PICO questions — population, comparators, outcomes — with the Member States behind each one. If the scope is revised, it shows what changed.
Reads clinical study reports, protocols, analysis plans and statistical outputs and puts the right number into every PICO × outcome cell, with the table it came from — head-to-head results first, then subgroups, then single-arm data.
Links the studies included in the systematic literature review to the comparators in the scope, builds the evidence network and finds which trials share a common comparator.
For every gap, proposes the indirect comparison method and checks feasibility — common comparator, effect modifiers, proportional hazards, sample size — following the HTA Coordination Group guidance. Prefers anchored methods; reports the effective sample size when only a matching-adjusted comparison is possible.
Drafts each template section in plain, neutral language, in the order of the JCA dossier template, citing the source of every number — including where there is no evidence.
Checks every number in the dossier against its source to the last decimal, the indication wording against the SmPC draft, that every PICO has a results section and that every post hoc analysis is labelled.
Approve plans, sign the dossier. The agents propose; a named person decides.
Everyone on the dossier team can ask in plain words — or say what to change. Answers point to the source they rest on.
What is still open on the Corventa dossier?
Corventa is on day 37 of 100; the dossier is due Wed Dec 9. Still open: the PICO 2 comparison plan, waiting for your approval (Oct 14); the PICO 6 matching-adjusted comparison, effective sample 23 of 61 (Oct 14); the PICO 5 evidence gap, because CRV-301 excluded ECOG 2 (Oct 21); section 3 wording against the SmPC draft (Oct 9); and section 7, certainty and limitations (Nov 14). The PICO 2 plan unblocks section 6.2.
Show only the PICOs Germany asked for
Germany asked for 4 of the 6 PICO questions: PICO 1 vs docetaxel (direct, CRV-301), PICO 2 vs docetaxel + angiokinase inhibitor (indirect), PICO 5 vs best supportive care (gap) and PICO 6, platinum doublet in EGFR / ALK (unanchored). The matrix is now filtered to Germany.
Why is the comparison with docetaxel + angiokinase inhibitor uncertain?
Two reasons. Both trials used docetaxel, so a Bucher comparison is feasible — but every CRV-301 patient had received anti-PD-(L)1 therapy and no patient in the 2014 comparator trial had, so prior immunotherapy is a likely effect modifier. And the estimate itself, OS HR 0.92 (0.70–1.20), has an interval that crosses 1. Sam Patel recommends a sensitivity analysis restricted to patients with at least 6 months of prior immunotherapy.
Remind Sam Patel two days before each comparison milestone
Done. Sam Patel gets a reminder two days before every comparison milestone — the next one on Mon Oct 12 for the Oct 14 sign-off. The rule is live in Settings and on the audit trail.
13 screens from the working solution, on its sample data. Pick one to see it large.
The window, the milestones, who signs what and how strict the evidence rules are — all settings.
| Setting | Default | Choose from |
|---|---|---|
| Dossier window | 100 days | 100 · 90 days |
| Comparison plans signed off | Day 44 | Day 40 · 44 · 50 |
| Full draft dossier | Day 75 | Day 70 · 75 · 80 |
| Final sign-off | Day 96 | Day 95 · 96 · 98 |
| Indirect comparison plans | Sam Patel → Dr. Lena Ortiz | Sam Patel → Dr. Lena Ortiz · Sam Patel only |
| Warn when the matched sample shrinks | Below 40% of patients | Below 30 · 40 · 50% |
| Prefer anchored comparisons | On | Unanchored only when no common comparator exists |
| Numbers without a source block signing | On | Every number must link to a table, output or publication |
Every figure is labelled: a target the solution is built to, an estimate, a typical published result, or a proven one.
“demo” = seen in the working solution, on its sample assessment data · “target” = the design goal, measured in the live solution · “estimated” = our estimate, including ≈ 70 days by hand · “typical” = the 90–100-day window set by EU HTA Regulation 2021/2282 and Implementing Regulation 2024/1381, as described in industry first-year lessons (2026). People named on this page are characters in the working solution; product names are fictional sample data.
Since 12 January 2025, the EU HTA Regulation (EU) 2021/2282 requires one joint clinical assessment for new cancer medicines and advanced therapies, instead of separate national ones. Member States state what they need as PICO questions — population, intervention, comparator, outcomes — and an assessor and co-assessor consolidate them into the assessment scope. The dossier answers every PICO and is due by day 100, and no later than 45 days before the CHMP opinion.
One row per PICO question and one column per outcome, so every cell the assessors asked for is visible at once. In the JCA Dossier each cell holds the value, its evidence type — direct, indirect, planned, high uncertainty or gap — its certainty and the table or publication it came from, with the Member States behind each PICO shown under it.
The comparison feasibility analyst proposes a method for every gap — anchored Bucher or network meta-analysis first, a matching-adjusted comparison only when no common comparator exists — and checks comparator, histology, line of therapy, prior treatment, performance status, proportional hazards and follow-up. Where no comparison is possible, the gap is justified in the dossier.
No. The evidence mapper never computes a comparison, and the feasibility analyst only proposes. A biostatistician recommends the plan, the dossier lead approves it, and only then does the validated statistics engine run it — with a second program reproducing every estimate to three decimals.
Every number links to the table, output or publication it came from. The consistency checker compares each one with its source to the last decimal, compares indication wording with the latest SmPC draft, and confirms post hoc analyses are labelled. It reports differences and never edits silently; a number without a source blocks signing.
The revised scope is compared with the previous version. A changed PICO reopens its sections, which move back to draft, and the owners are told.
Yes. Comparison plans are recommended by the biostatistician and approved by the dossier lead; results sections go to the clinical reviewer, wording to regulatory affairs, and the final dossier is signed by the head of market access. Every decision is recorded on the audit trail.
The Agentic Solution Engine builds and deploys it from your requirements and documents — your dossier template, study reports, statistical outputs and review library — and it goes live once every quality gate has passed. We will walk you through it on one of your own scopes first.
A trial synopsis becomes a checked first-draft protocol in days, in your own approved language.
1-min filmTMF CheckEvery trial document is read, filed in the right place and checked, so the file is inspection-ready every day.
1-min filmCSR WriterResults sections and patient narratives are drafted within days of database lock, every number linked to its table.
1-min filmWe’ll run the JCA Dossier on one of your own assessment scopes and its evidence.