Law report No. GLW-7228 · filed November 11, 2026
Regulation & EnforcementReported case
EU Law Does Not Require Clinical Trial Simulations — Yet
EU law requires no trial design simulations: neither the CTR nor the Pharma Package obliges sponsors, but soft law is filling the gap for complex trials.
By Amara Osei4 min read887 words
Holding
- Regulation (EU) No 536/2014 does not require sponsors to simulate trial designs before starting a study.
- The Pharma Package mentions in silico methods only as an alternative to animal testing, not as a trial-design obligation.
- The EC/EMA/HMA Complex Clinical Trials Q&A requires methodology validation 'including via simulation' for Bayesian settings.
- A Nature Reviews Drug Discovery paper proposes five good-simulation-practice principles: validity, transparency and reproducibility, thoroughness, efficiency, comparability.
- Simulation algorithms and code qualify as trade secrets under Directive (EU) 2016/943.

European law does not require sponsors to simulate a clinical trial's design before enrolling a single patient. Neither Regulation (EU) No 536/2014, the Clinical Trials Regulation (CTR), nor the forthcoming Pharma Package imposes such an obligation. That is the plain legal position set out in an analysis by Marco Blei, partner at McDermott Will & Schulte, published through the MedCity Influencers program — and it may not remain plain for long.
The question matters because simulations are cheap insurance. A recent paper, "Design, simulate, refine: simulation-guided clinical trials for accelerated drug development," published in Nature Reviews Drug Discovery, describes running thousands of virtual versions of a study on a computer before the first real volunteer consents. The authors call this a "test flight": teams learn how a design behaves, how likely it is to reach a correct answer, how long it takes, and how many subjects end up on ineffective arms.
What does EU law actually say?
Article 25 of the CTR requires a study to be scientifically sound. Article 3 protects subjects' rights, safety, and well-being. Annex I, Section D, calls for a description of the statistical methodology in the protocol. None of these expressly requires design simulations.
Simulation nevertheless enters, as Blei puts it, "through the window" — via soft law and regulatory practice. Three instruments carry the weight:
- The joint EC/EMA/HMA document Complex Clinical Trials – Questions and Answers, which requires validation of design methodology and, for Bayesian settings, expressly calls for validation "including via simulation."
- The EMA Reflection Paper on adaptive designs.
- The draft ICH E20 guideline on adaptive clinical trials, which devotes a section to the simulation plan and report.
The Pharma Package, Blei notes, concerns marketing authorization, post-market obligations, the EMA's role, and the incentive system. Where it refers to in silico methods, it does so as an alternative to animal testing at the preclinical stage — not as an obligation to simulate trial design. Simulating a design remains an optional methodological tool. Not a general legal duty.
Why "simulation" means three different things
The analysis draws a distinction practitioners would do well to memorize. Trial design simulation keeps patients real and simulates the design's behavior, measuring "operating characteristics" — statistical power, false-positive risk, expected duration, exposure to ineffective doses. In silico trials use virtual patients to replace or supplement real data. New Approach Methodologies apply computational models to replace animal testing. Obligations, evidentiary standards, and ethical implications differ radically across the three. Whether the law should mandate pre-trial simulation concerns mainly the first.
Blei offers one crisp ethical framing: if a preventive simulation avoids needlessly exposing dozens of patients to an ineffective arm, it is not merely good statistics — it is consistent with legal and ethical obligations already in force in the EU.
Would a mandatory simulation requirement make sense?
The arguments in favor are serious. Ethicality: fewer patients exposed to futile arms through well-calibrated early-stopping rules. Efficiency: faster, cheaper studies with fewer protocol amendments. Reliability: robust designs tested under many scenarios before launch. Systemic consistency: better assessment of the adaptive, Bayesian, and platform designs agencies are moving toward.
The counterarguments are equally real. A simulation is only as good as its assumptions — "garbage in, garbage out" — and "biased comparisons" can build a seemingly rigorous analysis to favor the preferred design. A general obligation would weigh above all on academic, nonprofit, and SME sponsors, favoring large groups and conflicting with the CTR's goal of facilitating noncommercial research. Assessors often cannot re-run all simulations. And for a simple fixed design, classical statistical formulae suffice.
Blei's synthesis: a risk- and complexity-based obligation, not a blanket one. Simulation reports would be mandatory only for adaptive, Bayesian, seamless, and platform designs, and strongly encouraged for others. This should come with a codified good-simulation-practice standard built on the paper's five principles — validity; transparency and reproducibility; thoroughness; efficiency; comparability — plus a strengthened EMA qualification pathway.
One boundary stays firm: requiring a design to be simulated is not the same as accepting in silico evidence that replaces patients.
What about the intellectual property?
The algorithms, assumptions, calibration parameters, and source code behind simulation models are, in Blei's words, paradigmatic examples of trade secrets under Directive (EU) 2016/943. A tension may arise between transparency and reproducibility on one side and the holder's interest in keeping proprietary algorithms undisclosed on the other.
His proposed accommodation is familiar to the sector: regulators could require submission of simulation reports and results while allowing underlying code to remain confidential, subject to inspection under secure conditions — as already happens with manufacturing process data. Looking ahead, patent applicants, trade-secret holders, data generators, and regulators need clarity on how the EPC, the Trade Secrets Directive, data exclusivity rules, and the Data Act interact. The goal: an ecosystem that incentivizes innovation and openness in equal measure.
The takeaway
Simulations do not replace human judgment; they inform it, making transparent what would otherwise remain a "black box." EU law does not require them today, but a proportionate embrace — in the right legislative vehicle, with clear methodological standards, and without crushing noncommercial research — could make trials safer, faster, cheaper, and more reliable. Europe's regulatory framework is entering a phase of profound renewal. This is a direction worth discussing now.
via doi.org (Original)
More from Amara Osei
Show full bio
Senior reporter covering people, launches and day-to-day practice in International law and legal practice at Global Law Wire.
327 articles
Also before the court
- EU Judges Give Cautial Green Light to Musical Sampling
- Halozyme Blocks Subcutaneous Keytruda Launch in Parts of Europe
- Anthropic Rolls Out Claude for Legal Artifacts With 12 Practice Plug-Ins
- Court Order Blocks Merck's Injected Keytruda in Parts of Europe
- UK Government Issues Response to Standard Essential Patents Consultation