A cluster trial is a clinical trial in which treatment strategies are assigned not to individual participants but to entire groups such as hospitals, wards, or medical practices. The groups are typically randomised; this creates specific requirements for information and consent that differ from individually randomised trials.
Study design and unit of randomisation
In an individually randomised trial, each participating person receives a treatment or comparator strategy according to the planned procedure. In a cluster trial, by contrast, allocation takes place at the level of a unit. For example, entire wards may apply different authorised treatment strategies, or several medical practices may implement different care pathways. Participants therefore belong to a cluster whose strategy is predefined, without each person being randomised individually.
Choosing this design may be appropriate when an intervention can only be implemented at group level, when spillover effects between individuals treated differently are to be expected, or when a care strategy is to be compared under real-world conditions. Scientifically, the protocol and statistical approach must account for the cluster structure: participants within the same unit may be more similar to one another than participants from different units. Randomisation, analysis, sample size, and the presentation of risks therefore require design-appropriate planning.
Specific considerations regarding consent
Article 30 of Regulation (EU) No 536/2014 contains a specific rule for cluster-randomised clinical trials. Under the strict conditions set out there, simplified consent may be permissible. It applies to clinical trials in which groups of participants are randomised as clusters rather than individually, different authorised investigational medicinal products are compared within their authorised indications, and the trial is conducted exclusively in one Member State. The other requirements of the Article must also be met.
This rule is not a general waiver of information or of participant protection. In particular, it requires that the methodology is necessary for conducting the trial, that the trial entails minimal risk and minimal burden compared with normal clinical practice, and that informed consent cannot be obtained because doing so would make conducting the trial impossible. The protocol must set out the reasons for the simplified approach in a comprehensible manner; national law and the assessment by the competent authorities remain decisive.
Distinction from individual randomisation and related entries
The key difference from an individually randomised trial lies in the level of allocation. With individual randomisation, the strategy is assigned to the individual; with cluster randomisation, the unit to which the person belongs is assigned. The legal particularity of a cluster trial lies not only in the design but also in the potential consent-related consequences under Article 30. Simplified consent must therefore not be inferred solely from the label “cluster”.
A cluster trial is also not automatically a multicentre study, a pragmatic trial, or a stepped-wedge design. A multicentre study can be individually randomised. A pragmatic study may, but does not have to, use cluster randomisation. The stepped-wedge design is a specific form in which clusters can switch to the intervention according to a predefined schedule. The related entry randomisation describes the general allocation principle; this entry describes the group-based special case.
Relevance for clinical trials
In cluster trials, feasibility must be linked from the outset with statistical and ethical planning. The unit of randomisation, recruitment of individual persons, information materials, data sources, and the handling of deviations from treatment pathways must be aligned. Simplified consent requires particularly robust justification: the sponsor must be able to demonstrate that the legal requirements are met and that participant protection is maintained despite group-based allocation.
Full-service CROs such as Mediconomics support the design concept, preparation of the protocol and consent documents, statistical planning, and the coordination of sites and data management. They can integrate the rationale for the chosen randomisation and any simplified consent procedure into the submission package, organise training for trial sites, and oversee conduct and data collection across clusters.
Frequently Asked Questions (FAQ)
Is every cluster trial permissible without individual consent?
No. Article 30 sets out strict cumulative requirements. A simplified consent procedure is only an option in the scenario described there and following the corresponding assessment.
What is randomised in a cluster trial?
Groups or units are randomised, such as wards, medical practices, or hospitals. Individual participants receive the strategy of the cluster to which they belong.
Is a stepped-wedge design always a cluster trial?
A stepped-wedge design is typically based on clusters and a staggered switch-over over time. Whether it qualifies as a cluster trial in the legal sense depends on the specific protocol and the applicable requirements.
Regulatory References
- Regulation (EU) No 536/2014, Article 30 – requirements for simplified consent in cluster-randomised clinical trials.
- Regulation (EU) No 536/2014, Articles 28 and 29 – general requirements and procedures for informed consent.
- Regulation (EU) No 536/2014, Article 2(2) – definitions of clinical trial and low-intervention clinical trial.
- European Commission, EudraLex Volume 10 – guidelines and Q&A on the application of the Clinical Trials Regulation.
