Navigating Evidence Standards: Stakeholder Perspectives on External Controls and Randomization in Cancer Trials

INTRODUCTION: Randomized controlled trials (RCTs) have historically been considered the 'gold standard' for generating evidence about emerging cancer therapies. However, conducting conventional RCTs has become increasingly challenging due to the increasing subdivision of cancer into smaller subtypes, the cost of conducting large-scale trials, and various ethical concerns. The use of 'external controls' has received increased attention of late as a potential mechanism for generating evidence in circumstances when obtaining RCT-level evidence is deemed infeasible (or otherwise challenging). External controls encompass a variety of statistical techniques whereby data are sourced from 'real world' or other settings and form the basis of comparative analyses. However, there are concerns about the potential for bias and other challenges to the veracity of non-randomized data. Official guidance regarding the appropriate use of external controls also varies widely across jurisdictions and agencies. METHODS: In this qualitative study, 52 semi-structured interviews were conducted with key stakeholders across the cancer clinical trials landscape to explore perspectives regarding external controls in oncology. This included various types of 'trial experts' (trialists, statisticians, ethicists, ethics review board members, industry, regulatory, and reimbursement body representatives). RESULTS: Data arising from interviews converged on three main themes: (a) support for randomization and the methodological limitations of non-randomized controls, (b) consideration as to possible utility of external controls, and (c) the relationship of external controls to public decision-making. CONCLUSIONS: Findings of this study highlight the 'value laden' nature of apparently technical questions about evidence generation and the need for further studies to systematically recognize, articulate and balance values of trustworthiness, efficiency, and fairness.