AI-Powered Registry Enhances Epilepsy Care and Surgical Triage

A prospective 90-day pilot program of the Autonomous Registry and Analytics (AURA) platform at West Virginia University (WVU) Medicine demonstrated that generative artificial intelligence (AI) can improve identification of surgical candidates with drug-resistant epilepsy (DRE) while reducing clinician burden. Results from the pilot were presented at the 2025 American Epilepsy Society (AES) Annual Meeting.

The AURA platform analyzes free-text clinical notes, imaging reports, and assessments to address common challenges in multidisciplinary care, including fragmented pathways, incomplete diagnostic workups, and delayed surgical referrals. At WVU Medicine, this includes the autonomous identification of surgical candidates with DRE by locating factors such as multiple failed antiseizure medications (ASMs), seizure burden, and diagnostic results suggestive of surgical candidacy. For incomplete or unclear records, the system queries prior encounters to extract relevant data and flags unresolved gaps.

For the pilot, AURA analyzed data from 820 scheduled visits in WVU Medicine’s EPIC electronic health record (EHR) system. The pilot assessed impact by tracking care gaps, pre-visit readiness, and clinician-reported utility for neurosurgical decision-making.​

AURA flagged the following care gaps in diagnostic workups:

• 54% of cases had outdated or missing MRIs.
• 91% of cases had absent neuropsychological evaluations.​
• 35% of cases had missing electroencephalograms (EEGs). ​
• 88 individuals (11% of cases) exhibited clinical patterns consistent with DRE yet lacked surgical referrals​.

Additional benefits since the deployment of the program include:

• Clinicians reported improved pre-visit readiness, reduced chart review time, and more timely identification of surgical candidates.
• Folic acid deficiencies were identified in 81% of eligible individuals.
• Documentation of post-surgical outcomes (Engel and International League Against Epilepsy [ILAE] scores) increased from 1% to 70.9%​.

The results of the pilot suggest that AURA’s scalable model may be used to support clinicians working in other complex neurological and multidisciplinary care domains by automating time-consuming aspects of chart review.

Source: Adelson PD, D’Haese P, Delaney H, et al. Autonomous clinical registries in epilepsy: generative AI–driven decision support for multidisciplinary care. Presented at: American Epilepsy Society Annual Meeting; December 5–9, 2025; Atlanta, GA.