A Comprehensive Phenotype Repository
for Real-World Studies
ComPLy offers computable phenotype definitions tailored for real-world clinical studies, enabling the identification of patients based on disease status, drug exposures, treatment responses, and other critical health factors. By standardizing phenotype definitions, ComPLy enhances research reproducibility, facilitates multicenter collaboration, and ensures consistent cohort identification across diverse electronic health record systems, driving robust and scalable clinical insights.
Phenotypes
Phenotype definitions in ComPLy have been successfully standardized and represented in a unified format, enabling seamless integration and accessibility. To compile phenotype definitions from diverse sources, we developed a data ingestion pipeline that extracts data using a combination of automated and manual approaches. This process involved curating data element mappings and implementing multiple data conversion scripts to ensure accuracy and consistency. All phenotype data and metadata were then aggregated into a harmonized dataset, facilitating uniform representation. To enhance usability, we provide an open data portal and API services, supporting efficient phenotype discovery, reuse, and integration into real-world studies.
Coding Systems
ComPLy incorporates 46 coding systems with standardized codes to support precise phenotype definitions. These standardized definitions enable consistent cohort identification based on diverse population characteristics, ensuring consistency in real-world clinical studies. By integrating standardized codes and queries within EHR systems, ComPLy strengthens the reliability of observational studies while ensuring consistency in data analysis and reporting across diverse healthcare data sources.
Data Sources
The phenotype library ComPly integrated five publicly available data sources of phenotype definitions, including the OHDSI Phenotype Library, PheKB, PheMap, the HDRUK Phenotype Library, and PubMed. ComPly includes 247 phenotype definitions from the OHDSI phenotype library, 50 eligible phenotype definitions that were collected based on public availability and phenotyping algorithms from the PheKB, 1,866 phenotype definitions from the PheMap, 1,375 phenotype definitions from HDRUK Phenotype Library and *** from the PubMed.
