The Pistoia Alliance, a global, not-for-profit alliance that advocates for greater collaboration in life sciences R&D launched its freely accessible FAIR4Clin guide. The guide will help clinical regulators, biopharmaceutical and healthcare organizations to implement the FAIR principles (Findable, Accessible, Interoperable, Reusable) in clinical datasets. The diversity of data types, standards, and regulatory and privacy requirements, combined with the huge volume of clinical datasets, has presented a particular problem in life sciences R&D. The guide emphasizes the value of FAIR to the clinical space, including making data machine-readable to support AI, innovating in clinical trial design, and enabling the transfer of data between sponsors, CROs, and regulatory agencies. FAIR4Clin was co-authored by the Pistoia Alliance and leading experts in the field, including from AstraZeneca, Bayer, Roche, Genentech, Galapagos, The Hyve, Elixir, The University of Manchester and Oxford University.
“While the importance of FAIR data is well recognized in pre-clinical research and academia, implementation is not yet at a mature phase in the clinical domain. This guide will accelerate the adoption of FAIR principles and the move towards improved data standards in the clinical sector, ultimately creating a more collaborative research ecosystem that will bring new therapies to patients faster,” commented Giovanni Nisato, PhD, Project Lead at the Pistoia Alliance. “We’ve had huge interest in the FAIR project since its launch, even more so after the Covid-19 pandemic reiterated the importance of being able to reuse and share clinical trial and healthcare data to accelerate drug development timelines. We are now calling on even more organizations to get involved in the next phase of the project, which will examine the business value of FAIR and develop further tangible resources, enabling best practices across industries, increasing more interoperability and effective collaborations.”
The way clinical studies are conducted often leads to a fragmentation of datasets because of unharmonized metadata, which compounds existing issues with data variety and veracity. Silos emerge because clinical data is typically only generated for its primary purpose – often in an unFAIR format for regulators, rather than with interoperability and reuse in mind. For example, clinical studies must comply with the Clinical Data Interchange Standards Consortium (CDISC) Study Data Tabulation Model (SDTM) or Analysis Data Model (ADaM), but the focus of these standards is not on FAIR metadata elements. Consistent FAIR metadata at the study level could reduce fragmentation, for example by “tagging” terms from an appropriate ontology to enable access to synonyms and cross-references, resulting in semantically stronger metadata. This means data become interoperable and machine-readable, and are more easily integrated and interpreted including by external parties when value can be added.
“FAIRification at scale has become a main pillar in the R&D data strategy at Roche. The Pistoia Alliance is a strong partner to tackle foundational issues pre-competitively as an industry. In the second phase of the FAIR implementation project, we supported the FAIR4Clin Guide. The guide provides a comprehensive overview on FAIR resources in the clinical space to further support our FAIR journey,” commented Martin Romacker, Senior Principal Scientist in Data and Analytics at Roche.
“The FAIR4Clin guide has been comprehensively developed by domain experts coordinating with multiple standards bodies and data integration alliances to provide a foundation for clinical data interoperability and reuse. We believe this to be the perfect starting point for improved clinical data stewardship that can galvanize these efforts in the research and medical community alike,” added Tom Plasterer PhD, Semantic Data Integration Lead, Oncology Data Science, AstraZeneca
The guide is aimed at data practitioners in the clinical field, such as clinical data managers or analysts, as well as data practitioners in the research domain (registry data curators, terminology managers or bioinformaticians), and those in the post authorization domain (medical affair or market access). The first iteration of the guide focuses on implementing the FAIR principles related to data and its metadata for clinical trials and Real-World Evidence. This includes:
- Interventional data – clinical trial data from phases 1, 2 and 3 of clinical trials, which are conducted prior to product launch and phase 4 data collected after the product launch.
- Results summary data and the associated metadata such as the study description and protocol.
- Real-World Data, for example routinely collected healthcare data from hospitals and primary care sites, data from wearable sensors, or patient reported outcomes.