PS41: Overcoming Obstacles in Using Hybrid Controls in Phase 2/3 Clinical Trials as the Basis for Regulatory Approval

Design and implementation of hybrid clinical trials that incorporate historical data, real world evidence and data from randomized controlled trials pose several challenges arising from both traditional and pragmatic clinical trial elements. Ensuring the compatibility and comparability of the external data with the clinical trial data is one of the most significant challenges, as differences in patient populations, treatments, and data collection methods can introduce biases and confounding. Additionally, there is a need for robust statistical methods to appropriately integrate and analyze the hybrid control data. Acceptance of hybrid controls by regulatory agencies is also presented with challenges, as agencies from different parts of the world have varying guidelines and demands for evidence of the validity and reliability of the external data sources. Finally, operational complexities, such as ensuring data privacy, audit and management of diverse data sources, can further complicate the utilization of hybrid controls in clinical trials. Clinical Research Data Sharing Alliance (CRDSA) is a multi-stakeholder consortium that serves the clinical data sharing ecosystem by creating a forum that defines common data standards and best practices, develops innovative approaches to data sharing principles, and thus ensures the integrity, quality, and usability of contributed data. DAHSHU IDSWG Oncology Team Working Group focuses on statistical methods in oncology clinical trials, with the objectives to develop, explore, promote, and implement innovative clinical trial designs in cancer drug development. We are proposing a panel of experts from industry, academia and regulatory agencies to discuss what we can collectively do to overcome the obstacles in using hybrid controls used as a basis for evaluating the effectiveness of investigational treatments in clinical trials. First, each panelist will talk about what they have experienced in their respective fields as challenges in dealing with hybrid controls, what technical and operational methods they have used to overcome those challenges and what they have learned from their experiences. Second, to look into the future, we would like to discuss how AI can be utilized with the high-dimensional data sets that have been created through the use of hybrid controls can help overcome these challenges. Finally, we would open it up to the audience to share their experiences, reflections and suggestions.