Senior Principal Data Scientist F. Hoffmann-La Roche Ltd, Switzerland
Background: Analysis of observational real world data (RWD) can inform assumptions and design elements of interventional clinical trials. In turn, better designed clinical trials can have a greater probability to increase scientific understanding, operational and regulatory success, and have a greater impact for patients and society, for example, by enabling earlier access to new medicines.
Objectives: We have developed a playbook aimed towards multidisciplinary teams involved in clinical trial design. The playbook suggests a variety of research questions which can typically be addressed using RWD, and which may inform various important aspects of clinical trials. The guide is intended for use early in the clinical trial design process (before protocol finalization), and includes links to additional relevant and publicly available guidance.
Methods: A cross-discipline industry team came together in a series of workshops to co-develop the playbook. Input was sought from various team members including epidemiology/RWD, biostatistics, clinical science, personalized healthcare, and data-sciences. Typical sections (i.e. chapters) of a clinical trial protocol were used to frame the discussion.
Results: The resultant playbook consists of 8 main chapters, describing different aspects of clinical trial design which can be informed by RWD: Disease background, Control arm identification and characterization, Inclusion/exclusion criteria, Choice of endpoint, Statistical Assumptions and Intercurrent Events, Safety considerations, Biomarkers, and Recruitment Strategy. We also found that analysis of other secondary data (for example from historical clinical trials) can complement RWD analyses for some questions. Each chapter highlights a number of research questions, along with published examples and further reading.
Conclusions: Our playbook, “Informing clinical trials with RWD”, can be used as a guide by industry and/or academic teams involved in clinical trial protocol development (or review), to systematically consider opportunities where analysis of RWD and other secondary data can inform trial characteristics and assumptions and ultimately improve the trial’s relevance and scientific accuracy.
