Ethical clinical trial design and differences in treatment effects.

Many global clinical trials primarily estimate a single overall treatment effect. However, when treatment effects are likely to differ between populations, for example due to differences in the disease, population characteristics or health-care systems, this approach can lead to misleading conclusions and raise ethical concerns. Justice is compromised when research conducted in low-resourced countries benefits primarily or exclusively populations of wealthier countries. A clinical trial design and analysis that focuses on estimating a single treatment effect, assumed to apply to all participating populations, goes against the ethical principle of justice and the positions of the World Health Assembly. To address this issue, we suggest a methodological strategy based on hierarchical modelling. This approach enables researchers to estimate treatment effects that are valid for each participating population, while potentially retaining efficiency comparable to traditional pooled analysis, as we demonstrate in an example. When substantial between-population differences exist, it produces valid, region-specific results. Strategies such as this one, if adopted into the standards for global trials, would allow regulators, funders and other stakeholders to ensure that trials are designed to help preserve justice for all participant populations.