Underutilization of syndrome‐specific <scp>ICD</scp> ‐10 codes for genetic epilepsies: Implications for precision medicine

Syndrome-specific International Classification of Diseases, 10th Revision (ICD-10) codes have the potential to improve identification of patients for precision therapies, clinical trials, and research, yet their real-world uptake is not well characterized. We evaluated the utilization of syndrome-specific ICD-10 codes at a large academic medical center among patients with pathogenic or likely pathogenic variants in 10 monogenic epilepsy genes with established codes (CDKL5, EHMT1, KCNQ2, MECP2, MED13L, SCN1A, SHANK3, SLC13A5, SLC2A1, SYNGAP1). Patients were identified from an institutional genetic testing database and were included if they had at least one clinical encounter after code implementation or genetic diagnosis. Variants of uncertain significance were manually curated, and Rett and Dravet phenotypes were reviewed for accuracy. Of 83 patients with qualifying variants, 39 met inclusion criteria. Only 56.4% (22/39) were ever assigned a syndrome-specific ICD-10 code, which appeared in 31.1% of encounters and accounted for 14.5% of all documented codes. Uptake varied by syndrome, provider specialty, and encounter type and increased over time. In the Dravet syndrome subgroup (n = 23), generic epilepsy codes were documented more than twice as often as the Dravet-specific code (G40.83). When G40.83 was documented, other epilepsy codes were used less frequently, suggesting it may be treated as a substitute for broader epilepsy codes. These findings demonstrate inconsistent and limited adoption of syndrome-specific ICD-10 codes, highlighting the need for improved coding support and integration of structured genetic data within the electronic health record.