The State of Data FAIRification in Clinical Research

In 2016, a diverse group of researchers came together to publish the FAIR Data Principles. Their goal? To push science further, faster by improving the infrastructure supporting the reuse of scholarly data. The FAIR principles call for proper data management to speed up knowledge gathering and propel innovation forward. Published in 2016, the guidelines provide key requirements to make scientific data FAIR—findable, accessible, interoperable and reusable.

As clinical trials increase in data volume and complexity, the FAIR principles and guidelines have been widely adopted since publication. FAIR is a core element and driver for the digital transformation of drug development and life sciences R&D. Here’s what to know about the current adoption of FAIR principles in clinical research today.

FAIR is not a standard and the principles offer no explicit guidance on how researchers should use the procedures for FAIRification. Many different approaches exist regarding how to FAIRify data. In order to guide researchers during the process, several workflows have been developed and published. These include: A generic workflow for the data FAIRification process and From raw data to FAIR data: The FAIRification workflow for health research, both published in 2020. Most of the steps in these workflows, however, require specific types of expertise. To ensure that the steps are properly executed, they should be carried out by a multidisciplinary team under the guidance of FAIR data stewards as opposed to individual researchers.

To ‘become more FAIR,’ one needs to make sure that collected data and associated metadata are both readable for humans (the study team or other researchers that might want to reuse the data) and machines (algorithms that process and analyze the data). Therefore, when incorporating the FAIR principles into a new trial’s setup, specific emphasis should be placed on amplifying the ability of machines to find and use data automatically—while supporting data's ability to be reused by individuals.

Last year, a team of researchers at Castor and Amsterdam UMC set out to get a reading on the industry’s perception and acceptance of FAIR. A questionnaire was distributed to researchers and support staff across six Dutch University Medical Centers and users of Castor EDC. The questionnaire assessed the individual’s perceptions and behavior concerning the FAIR Data Principles.

In May 2022, the findings were published, which showed that 62.8% of researchers and 81.0% of support staff are currently putting forth some effort to achieve any aspect of FAIR (findability, accessibility, interoperability, or reusability). The research found that less than half of the researchers add metadata to their datasets. Less than 40% add metadata to data elements, and just over a third deposit their data in a repository. Only 11% of the researchers and less than 25% of the support staff address all aspects of FAIR.

Of those surveyed, just over 1 in 3 researchers said that they can make their data more FAIR on their own. But if they received help to follow the principles more closely, more than 4 in 5 researchers reported success. This statistic confirmed researchers are willing to make their data FAIR if they received the right resources—but more help is needed to further implement the FAIRification approach.

One of the study's most interesting statistics revealed that almost 95% of researchers are aware of the usefulness of their data being FAIR for others. Even more interesting was that almost 90% are willing to FAIRify their data—given the right resources and support. The critical resources needed? Help from experts, budget, and time.

The study showed that when researchers are required to spend their own money and time on the FAIRification process, their intention and motivation significantly decrease. The industry needs to increase awareness of the short-term and long-term value of FAIR data and metadata and take the principles into account when planning (and budgeting) future studies. More help from institutions and funders is required to develop FAIRification training and tools. They also can financially support researchers and staff throughout the process. They could also appoint support staff and semantic data experts to help researchers with the FAIRification process.

If study teams think of FAIRification before they start a trial—and then receive the support they need to make their data FAIR—future research will be fueled by human- and machine-readable data and metadata. That data can be re-used by others and combined with other data, leading to exciting new insights. FAIRification therefore holds the promise of revolutionizing research science and greatly benefiting the entire research community and patients worldwide.