Ensuring phenotyping algorithms using national electronic health records are FAIR: Meeting the needs of the cardiometabolic research community

Phenotyping algorithms enable the extraction of clinically-relevant information (such as diagnoses, prescription information, or a blood pressure measurement) from electronic health records for use in research. They have enormous potential and wide-ranging utility in research to improve disease understanding, health, and healthcare provision. While great progress has been achieved over the past years in standardising how genomic data are represented and curated (e.g. VCF files for variants), phenotypic data are significantly more fragmented and lack a common representation approach. This lack of standards creates challenges, including a lack of comparability, transparency and reproducibility, and limiting the subsequent use of phenotyping algorithms in other research studies. The FAIR guiding principles for scientific data management and stewardship state that digital assets should be findable, accessible, interoperable and reusable, yet the current lack of phenotyping algorithm standards means that phenotyping algorithms are not FAIR. We have therefore engaged with the community to address these challenges, including defining standards for the reporting and sharing of phenotyping algorithms. Here we present the results of our engagement with the community to identify and explore their requirements and outline our recommendations to ensure FAIR phenotyping algorithms are available to meet the needs of the cardiometabolic research community.