Electronic phenotyping over the years has been evolving from simple to complex rule-based definitions, and more recently entering the machine learning age with probabilistic phenotype models. With the added complexity comes the additional need to have consistent and reproducible phenotype definitions for maintenance, replicability and community sharing. In this work we introduce how to construct probabilistic phenotype definitions with Automated PHenotype Routine for Observational Definition, Identification, Training and Evaluation (APHRODITE) that follow the FAIR principles to improve their reproducibility and quality. By using a centralized repository and creating a standard list of meta-data elements, we aim to guide probabilistic phenotype definition developers with a FAIR-compatible standard. By developing this standard within the Observational Health Data Sciences (OHDSI) initiative, we aim to ensure community wide compatibility and maximum reproducibility.
Special issue (if applicable):
FAIR Data, Systems and Analysis
Data repository URLs: