Our Lessons Learned In Implementing AI In Clinical Development
By Christopher Zergebel and Ed Chapman, Taiho Oncology
At Taiho Oncology, we work with multiple CROs across our spectrum of clinical trials – a common scenario for many pharma companies. As a result, in 2019 we also had multiple sources of clinical data, including electronic data capture (EDC), clinical trial management system (CTMS), lab data, etc., that were siloed, difficult to access, and not being leveraged to their fullest potential. This scenario was unsustainable and suboptimal for our company, customers, and patients.
Aggregating that complex and voluminous data into a single source of truth at the right frequency to better inform business decision-making and collaboration was a clear priority for us. We tried multiple single-point solutions, none of which yielded the results we required. We also had to address internal interest in identifying and implementing risk-based monitoring (RBM) tools as a potential solution. Another path considered and discarded was developing a reporting solution using standard reporting tools, which would have provided only a point solution that, in the long run, would prove inflexible as our business changed. We wanted to ensure the ability to be nimble and adapt rapidly and accordingly via whatever solution we implemented.
We needed a foundational data management and analytics system to aggregate data into standardized models, ensure comprehensive trial oversight, enable real-time access to quality data, and facilitate faster reporting to our end users.
In July of 2020, we implemented Saama Technologies’ AI powered data analytics platform, Life Science Analytics Cloud. Eight weeks later, our first batch of studies with EDC went live, after which additional data sources, such as CTMS, interactive response technology (IRT), lab data, study milestones, and planning data, were incrementally added along with the onboarding of additional studies. This effort involved users from our clinical operations, clinical development, and data management teams and required users from each of these groups to participate in reviewing/testing the functionality of the platform that was being implemented.
We now have a single source of truth for clinical data analysis and reporting. Data was centralized, integrated, and automated for processing and analytics functions/purposes, ensuring consistency and organization alignment with the Clinical Data Interchange Standards Consortium (CDISC). For example, all identified studies are now mapped into the system and all quality and integrity checks are done as data is processed through the system.
We can now access and run various analytical reports for key internal stakeholders, including medical monitors, data management, and other functions spanning the clinical trial process. The result is greater ownership of our own data, available to us 24x7, as opposed to traditional approaches relying on our CRO for insights into our studies. We also achieved significant reduction in the manual effort required to create analysis of oncology-specific reports, such as Tumor Assessment and RECIST. These insights are now automated and available as and when our medical monitors require them. Our clinical operations and clinical development teams can now focus on proactively addressing risks related to their respective studies, as opposed to reacting to events after they happen.
Key Takeaways
Taiho learned a number of valuable process- and people-oriented lessons from this experience.
Go All-In
Taiho found that point technology solutions are only Band-Aid fixes to clinical process problems. They are not scalable, are expensive to maintain, and they still encounter problems with analyzing and correlating data. Though Taiho had to make an initial financial investment in a new system, the return was significant in terms of improving our overall insight into specific studies.
Internal and External Alignment Is Critical
When implementing and leveraging a new platform for operational and clinical development needs, it's important to ensure alignment throughout the organization. From the time this initiative was conceived, leaders from Taiho’s IT and business units were identified as stakeholders and were involved in the decision-making on a regular basis. Likewise, existing data partnerships with our CROs proved very important, and getting them to align with our strategic vision for a foundational, AI-powered data analytics system was a big contributor to making this initiative a success. If data access from CRO, vendor, and lab partners needs to be addressed, it is worth considering updating future partner contracts to establish timelines and cadence for frequent data sharing.
Leadership Support Matters
Executive buy-in for this transformation was critical to its success. Top-down endorsement, and awareness throughout the organization of the support for this shift, helped motivate and align key internal constituents on the need for and benefit of our new approach. Leadership’s clear understanding of the benefits of AI helped instill confidence throughout the organization that this move was in the best interests of the business.
Empower Employees Through Training
To ensure employees’ ability to optimize the new platform, Taiho instituted comprehensive training both before and during the migration, working to troubleshoot any issues associated with adoption as needed.
Predictive Insights Across Therapeutic Areas Is a Gateway
Arriving at and leveraging a single source of data truth was only the first step for Taiho. Future layering of additional smart applications to further inform our clinical and operational insights will enable us to analyze a specific compound across different therapeutic areas for deep insights or predictive outcomes we could not derive before the data was centralized. The goal is additional and earlier market success.
A Single Source of Data Truth Leads to Clinical Optimization and Success
A reliable single source of data truth across the entire clinical development life cycle results in faster database lock times and completion of clinical studies. When data is organized into a central system and is traceable and auditable, the business and IT teams can spend their valuable time analyzing reports from an integrated solution – not creating or working from spreadsheets. Medical reviewers can access a graphical longitudinal profile view of study subjects via a single report, and they have the ability to easily drill down for detailed analysis.
Particularly in a COVID-19 and post-pandemic world, it becomes extremely critical to have one place from which to get an end-to-end oversight into clinical trials, especially when on-site activities are limited. Taiho found that having a NASA-like command center from which we can monitor all the events that are happening within our clinical trial portfolio is invaluable.
Ensuring Data Privacy is Crucial
Top of mind for Taiho throughout this process was guaranteeing the privacy of patient data. We prioritized accountability for stewarding our sensitive clinical trial data anywhere on our platform, a robust GDPR compliance program, and Privacy Shield certification. We adhered to the most stringent standards in the market, not just reaching the minimum requirements by law.
Conclusion
Implementing an AI-powered data analytics system was a game-changer for us. We aligned key clinical trial stakeholders, both internally and externally, which resulted in improved risk management and study performance. Taiho not only changed the course of our current clinical journey, but we also positioned our company for continued and enhanced clinical and business success.
About The Authors:
Ed Chapman is the Chief Information Officer for Taiho Oncology, Inc. He and his team are responsible for anticipating the future technology needs of the company and building and executing the technology strategy.
Chris Zergebel is the vice president of R&D services for Taiho Oncology, Inc. He and his team are responsible for execution and delivery of the company’s development programs.