Mapping ICD-10 and ICD-10-CM Codes to Phecodes: Workflow Development and Initial Evaluation | JoVE Visualize

Area of Science:

Genetics and Genomics
Health Informatics
Biomedical Data Science

Background:

The phecode system, crucial for phenome-wide association studies (PheWAS) in electronic health records (EHR), was initially based on the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM).
The transition to newer coding systems necessitates updated mapping strategies to maintain research continuity and expand analytical capabilities.

Purpose of the Study:

To develop and evaluate mapping algorithms connecting International Classification of Diseases, 10th Revision (ICD-10) and ICD-10-CM codes to the established phecode system.
To enable the utilization of extensive ICD-10 and ICD-10-CM data for PheWAS research within EHR systems.

Main Methods:

Multiple resources, including concept relationships and explicit mappings from governmental and biomedical organizations, were employed to map ICD-10 and ICD-10-CM codes to phecodes.
Map coverage was assessed using two large databases: Vanderbilt University Medical Center (VUMC) for ICD-10-CM and UK Biobank (UKBB) for ICD-10.
Map fidelity was evaluated by comparing phenotype reproducibility and conducting a PheWAS against the gold-standard ICD-9-CM phecode map.

Main Results:

Over 75% of all ICD-10 and ICD-10-CM codes were successfully mapped to phecodes, with over 90% of unique codes in the UKBB and VUMC cohorts being mapped.
Phenotype reproducibility reached 70-75% for chronic diseases but was less than 10% for acute diseases when using the ICD-10-CM phecode map.
A PheWAS using both ICD-9-CM and ICD-10-CM maps with a Lipoprotein(a) genetic variant (rs10455872) replicated known genotype-phenotype associations for coronary atherosclerosis and chronic ischemic heart disease with comparable effect sizes.

Conclusions:

Beta versions of ICD-10 and ICD-10-CM to phecode maps have been developed, facilitating research.
These new maps empower researchers to effectively leverage existing and future ICD-10 and ICD-10-CM data for PheWAS analyses in EHRs.