Multi-arm trials that test several treatments at once against a single control group are becoming common, but there is no agreed standard for when and how to adjust the statistics to account for the multiple comparisons being made. This matters because without clear guidance, statisticians designing these trials do not know what approach regulators and funders will accept, and reviewers cannot consistently judge whether a proposed design is appropriate. The result is inconsistency, inefficiency, and potential unreliability in how treatments are tested in publicly funded research. If this project succeeds, it will produce a practical guidance document that researchers can use to decide whether to adjust for multiplicity, how to do it, and how to explain their reasoning. This would make the design and analysis of multi-arm trials more consistent and transparent, ensuring that treatments are tested reliably without wasting patient time or public money. The guidance will be illustrated with real case studies and validated through a consensus process involving statisticians, clinicians, and patients. The impact is methodological rather than directly clinical—it improves the machinery of clinical research itself.
View original technical description
Background Multi-arm trials, such as platform trials and parallel-group trials comparing multiple treatments to control, are increasingly used to efficiently evaluate treatments. Patient numbers, staff time, and cost can all be lower compared to conducting separate trials. Additionally, having more experimental arms increases the likelihood of finding at least one successful treatment within a single trial. However, opinions and practices vary regarding the statistical adjustment for multiplicity that arises from conducting multiple treatment comparisons within a multi-arm trial. Statisticians are unsure what is acceptable and Molloy et al. [2022] recently called for clearer guidance from stakeholders on the appropriate setting for multiplicity adjustment. It is understood that a universal approach overlooks the diversity of publicly funded multi-arm trials [Juszczak et al., 2019] and adjustment should be thought through carefully for each trial context, but there is insufficient guidance on how this should be done, and what factors should influence the choice. Varying practice and lack of guidance means testing treatments in multi-arm trials is not approached consistently. Without guidance, reviewing proposed multi-arm designs can be challenging with no consensus on what is appropriate in what context. Aims This project will explore the factors influencing the choice of statistical multiplicity approaches in confirmatory publicly funded multi-arm trials. It will establish consensus and produce guidance for researchers on the factors to consider when designing a multi-arm trial. It will also recommend how to communicate the rationale effectively, so research consumers can make informed judgements about the suitability of the chosen methods. Methods The project will comprise four stages. I will conduct two scoping reviews. The first will examine methodological literature on multiplicity adjustment approaches in multi-arm trials. The second will review ongoing and recently published multi-arm trials to identify and describe contemporary multi-arm trials and corresponding approaches to handling multiplicity. Qualitative interviews and focus groups will be used to investigate the opinions, ideas and experiences of key stakeholders (e.g. statisticians, clinicians, patients) on the design, multiplicity adjustment methods, and interpretation of results in multi-arm trials. The findings from the reviews and qualitative research will be used to develop a guidance document to advise researchers on the principles to use when designing multi-arm trials and deciding if and how to adjust for multiplicity. The guidance will also recommend what to include to justify the chosen approach. A modified Delphi process will engage a large group of statistical experts and stakeholders in order to achieve consensus on the items to include in the guidance, and their relative importance. Case studies of contemporary multi-arm trials with different designs will be used to illustrate the use of the guidance. Impact and Dissemination Results from this project will guide researchers on how to choose and justify a multiplicity adjustment approach in multi-arm trials. This will improve the design and analysis of publicly funded multi-arm trials and increase the consistency of approach, which in turn ensures treatments are being tested reliably and efficiently. Dissemination will include journal articles, conference presentations, webinars and feedback to interest groups.
Plain English summaries and category classifications on this site are generated by AI and may not perfectly reflect the original research.
Is something wrong? Let us know