Roche Innovative Approach to Data-Driven Site Selection in Clinical Trials

In a groundbreaking presentation at SCOPE Europe 2024, Susannah Finney, Clinical Operations Lead and Product Owner at Roche, UK, introduced a novel approach to site selection in clinical trials. The presentation, “Embedding Digital Tools and Data-Driven Approaches for Site Selection,” highlighted integrating digital tools and data analytics to enhance diversity and inclusion (D&I) and predict site performance. This initiative aims to address the traditional challenges in site selection by shifting from intuition-based decisions to data-driven insights, ensuring that trial populations accurately reflect real-world demographics.

Identifying the Core Problem

Susannah Finney began by addressing the traditional challenges in site selection, emphasizing the reliance on intuition and past experiences rather than data-driven insights. She noted that clinical operations professionals often chose sites based on previous success without considering their suitability for new trials. This approach led to a lack of diversity in trial populations, as assumptions were made that large cities would naturally provide diverse participant pools. However, diversity encompasses more than geographic location; it includes race, ethnicity, socioeconomic status, and environmental factors, all critical risk factors for disease. The core problem identified was the failure to use data to inform site selection, resulting in trial populations that did not accurately reflect real-world demographics. This oversight could skew trial results, as the sample population might not represent the broader patient population affected by the disease being studied. By not incorporating data on diversity and inclusion, trials risked missing out on crucial insights that could affect the efficacy and safety of the tested treatments.

Defining the Minimum Viable Product (MVP)

The next step in Roche’s strategy was to define the Minimum Viable Product (MVP), which would serve as the foundation for their data-driven site selection tool. Finney explained that the MVP’s primary goal was to enable users to leverage data for optimal site selection. The scope of the product was twofold: to provide data for trial performance and to enhance diversity and inclusion. The product needed to educate users and change existing processes to achieve this. The MVP included key features such as a site score/ranking system, allowing users to evaluate potential sites based on various performance metrics. Customizability was crucial, enabling users to adjust weightings and tierings according to their needs. Machine learning models were also incorporated to predict site performance, providing a more accurate and data-driven approach to site selection. For diversity and inclusion, the product offered an interactive map with demographic data at a granular level. This feature allowed users to gain meaningful insights into potential trial sites, ensuring that the selected sites would reflect the diverse populations affected by the studied diseases. By defining these key features, Roche aimed to create a “good enough” product to address the problem of not using data in site selection.

Implementation Strategy

Finney outlined a comprehensive implementation strategy, emphasizing the importance of addressing potential risks early in the process. She noted that a “white-glove” model was chosen over a “self-service” model to ensure the product’s successful integration into clinical operations. In a self-service model, users would receive training and then use the product independently. However, this approach could lead to infrequent use and forgotten training, resulting in user frustration. In contrast, the white-glove model involved a consultative process where subject matter experts would guide users through the product. This approach significantly reduced training needs and change management challenges, as there was no need to onboard or train every end-user. The service was ready anytime, greatly reducing the risk of knowledge loss after training. The implementation process included several key steps. First, subject matter experts were recruited in each country, focusing on individuals passionate about technology, data and championing its use in clinical operations. These experts formed a community led by the product owner, who provided insights to the software team. The network underwent extensive training to empower them to provide white-glove service. A template was provided to ensure consistency in the consultation process. Finally, an organization-wide mandate and key performance indicators (KPIs) were set to embed the product into clinical operations.

Training and Change Management

Training was critical to Roche’s strategy, focusing on onboarding a “SuperUser Community” of subject matter experts. These experts were trained on various aspects of the product, including the rationale behind its development, data inputs, and outputs, the overall process, and how to provide white-glove consultation. The training also included a deep dive into the data, ensuring the experts were well-equipped to guide users through the product. Change management involves a staggered communication approach to prepare leadership for the transition. Leadership buy-in was essential to ensure accountability and oversight at the country level. Success stories were published to demonstrate the product’s effectiveness and encourage adoption. By embedding the product formally into the process and holding country leadership accountable, Roche aimed to facilitate global adoption and ensure that the product became an integral part of clinical operations. This comprehensive approach to training and change management was designed to overcome resistance and ensure that the new data-driven processes were embraced across the organization.

Conclusion

Roche’s innovative approach to site selection in clinical trials represents a significant shift towards data-driven decision-making. By addressing the core problem of underutilized data, defining a clear MVP, and implementing a robust strategy, Roche aims to improve trial outcomes and ensure that they reflect the diverse populations affected by diseases. This initiative underscores the importance of integrating digital tools and data analytics in clinical research, paving the way for more inclusive and effective trials. As Roche continues to refine and expand its data-driven approach, the potential for more accurate and representative clinical trials becomes increasingly attainable, promising better outcomes for patients worldwide.