Basket, umbrella, and platform trials are three forms of a master protocol. A basket design tests one treatment across multiple diseases or subtypes sharing a common feature; an umbrella design assigns individuals within one disease to biomarker-related treatments. A platform trial allows trial arms to be added, modified, or closed during its course.
The Master Protocol as Common Framework
The FDA describes a master protocol as a protocol with multiple sub-studies that may pursue different objectives and coordinate the evaluation of one or more investigational products in one or more disease subgroups. It may include parallel comparisons, separate or shared controls, and fixed or adaptive elements. The master protocol therefore establishes the overarching organisational and statistical framework.
The three terms do not automatically denote the same concrete trial. Their practical design depends on how disease, biomarkers, interventions, control groups, and temporal modifications are organised. A trial programme may exhibit features of several structures, but must clearly describe the respective assignment in the protocol and analysis plans.
Basket and Umbrella Designs
In a basket trial, a targeted treatment is typically investigated across multiple tumour types or disease subtypes that share a common predictive biomarker feature. The cohorts may be analysable separately by tumour type; a shared molecular mechanism does not replace the justification for why the diseases should be treated comparably.
An umbrella trial, by contrast, remains within a specific disease or tumour type and assigns biomarker-defined subgroups to different targeted treatment approaches. The central structural feature is therefore the division of a disease population. For each subgroup, diagnostic procedures, assignment criteria, and control strategy must be aligned.
Distinction from Platform Trials
The platform trial is characterised by the ability to continuously add, adapt, or close arms according to the rules of the master protocol. It is therefore temporally dynamic. A basket or umbrella design may have fixed cohorts; the terms do not become a platform simply because multiple arms are conducted in parallel.
Basket, umbrella, and platform are therefore not synonyms, but different forms under the umbrella term master protocol. Basket aligns the structure around one feature across multiple diseases, umbrella around multiple approaches within one disease, and platform around the continuous management of trial arms. The designation should not serve as a substitute for the concrete statistical plan.
The FDA notes that a master protocol may coordinate sub-studies with different objectives. This creates particular requirements for decision governance: which hypotheses each sub-study tests, which data may be shared with other parts, and when a modification takes effect must be transparent to all parties. A shared control group may be efficient, but requires a convincing justification of comparability across recruitment periods and cohorts. In a platform, rules for adding or closing arms are added. The master protocol therefore does not replace the precise description of each sub-study, but is the place where common principles and dependencies are defined.
For the terms, the orientation of the research question is more important than the number of participants. A small trial may have a basket structure if it investigates an approach across biomarker-defined diseases. A large trial remains an umbrella design if it tests different therapies within one tumour type according to biomarker groups. A platform does not arise from size, but from the regulated continuous modification of arms. The designation must therefore match the decision matrix actually described.
Common terminology must also not obscure the fact that each cohort may have its own target population, target variable, and decision rule. The transparency of the master protocol arises from the fact that these differences remain visible despite shared infrastructure.
Relevance for clinical trials
Master protocols require particularly careful governance for versions, sub-studies, and biomarker processes. Changes to one platform arm must not unintentionally alter the data collection or control group of other arms. In basket and umbrella trials, the quality of biomarker-based screening is critical, because it determines cohort assignment and analysability.
Full-service CROs such as Mediconomics support the structuring of master protocol, sub-studies, and change management. They coordinate submission strategy, site training, screening data flows, and monitoring of arm-specific procedures; biostatistics and medical writing ensure that populations, controls, interim decisions, and results of each cohort are reported unambiguously.
Frequently Asked Questions (FAQ)
Is every trial with multiple treatment arms a platform trial?
No. Multiple parallel arms are not sufficient. Characteristic of a platform trial is the pre-defined possibility to add, modify, or discontinue arms during the course of the trial.
What connects the cohorts of a basket trial?
They typically share a predictive biomarker or biological approach, even though they may belong to different tumour types or disease subtypes.
Why does an umbrella trial require reliable biomarker tests?
The test determines to which subgroup and therefore which treatment arm an individual is assigned. Errors in this assignment compromise the validity of arm-specific analyses.
Regulatory References
- FDA Master Protocols for Drug and Biological Product Development – addresses design, analysis, and submission of master protocols.
- FDA Master Protocols: Efficient Clinical Trial Design Strategies – explains the structure of multiple coordinated sub-studies.
- ICH E9 Statistical Principles for Clinical Trials – contains statistical principles for pre-specified comparisons.