Status of Presurgical Evaluation Among Patients with Drug-Resistant Epilepsy Identified with a Computable Electronic Health Record Phenotype

Introduction: Interdisciplinary practice parameters recommend that all patients with drug-resistant epilepsy (DRE) undergo comprehensive presurgical evaluation at an accredited epilepsy center. Reporting suggests long delays to referral and underutilization of surgery, however, there are no established methods for ascertaining the care gap among all patients who interact with the wider medical systems in which epilepsy centers reside.

Methods: Using our previously validated computable phenotypic algorithm, we identified patients with probable DRE at or preceding any encounter in our medical system’s electronic health record between 4/1/2020 and 6/1/2022. We randomly sampled 200 patients for chart abstraction by two independent reviewers. Patients with confirmed DRE were classified by their furthest progress in the presurgical evaluation care pathway. We then tested for association between demographic and clinical variables and whether patients had begun Phase 1 presurgical diagnostic testing.

Results: Our algorithm identified 3,028 patients with probable DRE. Among 200 randomly sampled patients, 42.5% (n=85) were confirmed to have DRE, 34% (n=68) were drug-responsive, 14.5% (n=29) had indeterminate drug responsiveness, and 9% (n=18) were not confirmed to have epilepsy. Among those with DRE, 56% (n=48) had no pending presurgical referrals or diagnostic studies, 12% (n=10) were pending presurgical (Phase I) diagnostic studies, 7% (n=6) had completed Phase 1 interdisciplinary case conference presentation with surgery pending or not recommended, 5% (n=4) had completed Phase 2 intracranial monitoring with therapeutic surgery pending, and 20% (n=17) had completed at least one therapeutic surgery. Patients already identified and tracked by our epilepsy surgery patient navigator at the time of chart review were more likely to have begun phase 1 diagnostic testing (OR 42.5, 95% CI 8.8-203.8, p< 0.001).

Conclusion : A computable phenotypic algorithm can be used to identify patients with DRE who have not begun presurgical diagnostic evaluation and may benefit from referral to a patient navigator or other interventions.