ICH M11 Is Final — but the Agency That Makes It Matter Is Unsteady

On May 22, 2026, FDA published the final ICH M11 CeSHarP guidance package (Docket FDA-2022-D-3054) — three documents: a guidance with rationale and recommendations, a standardized protocol template with defined headers and common text elements, and a technical specification carrying harmonized terminologies and structured data fields for electronic protocol exchange. ICH formally adopted CeSHarP at Step 4 in November 2025; the May 22 publication brings FDA into conformance and starts the implementation clock across all ICH member jurisdictions — FDA, EMA, PMDA, Health Canada, MHRA, and others.

This is not a guidance to skim and file. For biostatisticians and statistical programmers, M11’s structured template will directly affect how estimands, endpoints, analysis populations, and statistical methods sections are authored — and how they are machine-readably encoded. The technical specification’s harmonized terminologies will need to map onto existing CDISC and HL7 FHIR ecosystems, so data management and standards teams face a genuine gap assessment before any conformant submission is possible. Protocol authoring systems, eTMF platforms, and regulatory submission pipelines all sit in scope. The operational work for most sponsors is just opening.

The practical ROI has been quantified, at least anecdotally. A former FDA CDER statistical reviewer — eight years as primary reviewer in Division III — wrote in the Spring 2026 BIOP Report that applications implementing ICH E9(R1)-aligned protocols and SAPs are “observably more reviewable and generate fewer information requests.” That applies directly to the M11 question: structured, machine-readable protocols reduce friction on both sides of the submission desk. She also flags treating deaths as missing data in long-term fatal-disease trials as the canonical estimands failure mode — precisely what a rigorous M11-compliant protocol section should prevent.

The paradox sitting next to the press release

The same week M11 was finalized, FDA was navigating the aftermath of CBER’s Vinay Prasad personally signing a refusal-to-file against Moderna’s mRNA flu vaccine BLA — unprecedented for a vaccine — citing inadequate comparator choice and overruling internal reviewers who were prepared to accept the application. FDA had previously signed off on the trial design. One week later, FDA reversed course on Moderna’s mRNA RSV/COVID vaccine, agreeing to a standard review for ages 50–64 and accelerated approval for 65+. The sequence reads as institutional dysfunction rather than principled recalibration.

For biometrics teams, the operational implication is blunt: pre-submission design agreements — the mechanism M11 is partly designed to make more reliable — are currently worth less than they were. When a comparator choice accepted during trial design becomes grounds for refusal-to-file at the BLA stage, no amount of structured protocol harmonization closes that gap. The risk is not M11’s fault. But sponsors who treat M11 conformance as a regulatory predictability guarantee are optimistic in a way the current environment does not support.

Also on May 22, FDA published a NEJM article formalizing the shift to a single pivotal trial as the default for drug approval — a change that concentrates evidentiary risk in a way that makes estimand precision more consequential, not less. A concurrent systematic review in Journal of Clinical Epidemiology found evidence of estimand changes between Trial 1 and Trial 2 under the two-trial paradigm — a reminder that replication was doing more methodological work than it was getting credit for.

MHRA, RISW, and what to watch

While FDA generates uncertainty through leadership action, MHRA is generating structured optionality through transparent process. The agency has launched a public consultation (closing July 30, 2026) on a proposed Rare Disease Therapies Framework centered on a new Investigational Marketing Authorisation — a single instrument merging clinical trial approval with a progressive route to marketing authorization for diseases with UK prevalence ≤1 in 50,000. The IMA explicitly endorses adaptive designs, RWE integration, and digital twins, with NICE oversight built in. For sponsors in ATMPs or ultra-rare indications, this consultation is a direct opportunity to shape eligibility criteria and evidentiary standards before the framework is finalized.

At the RISW 2025 plenary, career FDA statisticians from CDER, CBER, and CDRH articulated methodologically coherent positions — RWE as complement not replacement, rare disease methodology beyond classical hypothesis testing, AI/ML under unchanged safety standards. These are reasonable positions from credible scientists. Whether political appointees sustain them in high-visibility cases is the structural question that the Prasad RTF leaves open. Biometrics teams should treat the RISW positions as accurate signals of career FDA statistical thinking while building in scenario planning for overrides. That is not a comfortable operating posture. It is, however, the accurate one.

What to watch: FDA’s VRBPAC review of Moderna’s mRNA flu vaccine is scheduled June 18 — the briefing documents will be the first public look at what statistical package FDA is willing to put before an advisory committee after the RTF-and-reversal sequence. And the M11 implementation question the guidance does not yet answer — what transition timeline for conformant submissions is acceptable — remains open. Watch the docket.

Protocol read: M11 is the predictability tool the field has needed for a decade — but FDA’s political layer can override career statistical judgment in any given week, and pre-submission design agreements no longer carry the assurance value sponsors used to assign them. Implement M11 because it improves the work, not because it buys regulatory certainty.

What to do now:

• Begin the M11 gap assessment now — protocol templates, eTMF, CDISC/FHIR mapping — without waiting for an FDA-imposed transition date.
• Treat pre-submission design agreements as informative but no longer assurance-grade in politically exposed areas; build review-reversal scenarios into late-phase risk plans.
• Read the VRBPAC June 18 Moderna briefing package as the canonical signal of what statistical evidence FDA is willing to put before an advisory committee after the RTF-and-reversal sequence.