Platform Trial

Platform trials are one of the clearest examples of a clinical trial design innovation that changes not just the statistics but the administrative and regulatory machinery underneath a trial. Because a single master protocol governs multiple arms over an extended, sometimes open-ended timeline, the usual per-trial administrative cycle — one protocol, one IRB review, one IND, one statistical analysis plan, one set of case report forms per study — gets restructured around a platform that persists while its constituent arms change.

How a platform trial is defined

FDA’s guidance on master protocols (originally issued as draft guidance in 2018 and finalized in 2022 for oncology master protocols, with a broader draft guidance, Master Protocols for Drug and Biological Product Development, addressing master protocols across drug and biological products more generally) defines a master protocol as a single protocol designed to answer multiple questions, potentially about multiple interventions, in one overall trial infrastructure. Within that master-protocol category, FDA and the trial-methodology literature distinguish three related but distinct structures:

• Platform trial — multiple interventions evaluated for a single disease or condition on an ongoing basis, with arms entering and leaving the trial over time against a shared, concurrently randomized control.
• Basket trial — a single intervention (or small fixed set) tested across multiple diseases, conditions, or molecularly defined subtypes.
• Umbrella trial — multiple interventions tested for a single disease, but with a fixed, pre-specified roster of arms rather than an open-ended one that can add or drop arms mid-trial.

The distinguishing feature of a platform trial specifically is perpetuity combined with a shared control: the trial infrastructure (governance, data systems, statistical framework) is built to outlast any individual arm, and every experimental arm is compared to the same concurrently enrolled control population rather than each new intervention triggering an entirely separate trial with its own control group.

Regulatory and administrative implications

Single IRB / central ethics review

Because a platform trial typically enrolls across many sites and adds arms over time, FDA’s master-protocol guidance recommends use of a central or single Institutional Review Board to review the protocol, informed consent materials, and amendments across all sites and arms, rather than requiring separate local IRB review each time an arm changes. This mirrors — and for federally funded multi-site research is reinforced by — the NIH single IRB (sIRB) policy for cooperative multi-site research generally, but for platform trials it is close to a practical necessity: re-running full local IRB review at every site every time an arm is added or dropped would make the perpetual-platform model unworkable. See Commercial IRB and Independent Ethics Committee (IEC) for how centralized ethics review is structured in practice.

Informed consent across evolving arms

Consent documents for a platform trial have to be written and administered so that participants understand they may be randomized to one of several current or future arms, and that the roster of arms may change while they remain in the trial — a materially different consent problem than a fixed two-arm study, and one FDA guidance addresses directly.

Regulatory submissions

FDA generally recommends that a master protocol be submitted as a single Investigational New Drug (IND) application (or, for a platform intended to support marketing applications for more than one product, coordinated across the relevant INDs), with sponsors encouraged to request a pre-IND meeting to align on design, arm-addition criteria, and analysis plans before the platform opens. Amendments to add or drop an arm are then handled as protocol amendments to the existing IND rather than new IND filings each time.

Adaptive statistical design

Platform trials are almost always paired with adaptive trial design elements: interim analyses that can drop an arm for futility, respond to accumulating safety signals, or (in Bayesian response-adaptive designs) shift randomization probability toward arms performing better in the shared comparison. Because these adaptations happen mid-trial and can affect the interpretation of the overall Type I error rate across the whole platform, the statistical analysis plan and the pre-specified decision rules for adding, dropping, or re-weighting arms are reviewed with particular scrutiny by both the IRB/ethics committee and the regulator — they are what keeps an evolving, multi-arm platform statistically interpretable rather than an ad hoc collection of comparisons.

Data and safety monitoring

A single Independent Data Monitoring Committee (IDMC) typically oversees the whole platform rather than a separate committee per arm, since arm-addition and arm-dropping decisions depend on comparisons across the shared control and need to be made consistently across the trial. Documentation of these platform-level governance structures — the master protocol, statistical analysis plan, IRB approvals, and IDMC charters — is maintained centrally in the trial’s Trial Master File (TMF), alongside arm-specific documentation as each intervention enters or exits.

Why the design matters administratively, not just statistically

Platform trials are frequently framed as a statistical innovation, but for research administrators, IRB staff, and regulatory-affairs teams the more immediate change is operational: a single, ongoing governance and documentation structure has to support an evolving set of interventions, sites, and consent versions over a multi-year timeline, instead of the trial’s administrative lifecycle being coextensive with a single, fixed protocol. That is what the FDA master-protocol guidance, and the single-IRB and centralized-monitoring practices built around it, are designed to accommodate.