Authors :
Presenting Author: Alicia Milam, PhD – MEDVAMC
Erin Sullivan-Baca, PhD, ABPP – Baylor College of Medicine
Marissa Kellogg, MD, PhD – VA Portland Healthcare System
Molly Horstman, MD, MS – Baylor College of Medicine
Maria Raquel Lopez, MD, FAES – Miami VAMC
Stephan Eisenschenk, MD – VA National Tele-EEG and Epilepsy Program
Rizwana Rehman, PhD – Durham VAMC
Rebekah Kaska, MSN, AGACNP, CNRN – Richmond VAMC
Andrea Hildebrand, MS – VA Portland Healthcare System
Amtul Farheen, MD – VHA
Jacob Pellinen, MD – ECHCS
Sarah Durica, MD – Oklahoma City VAMC
Nina Garga, MD, FAES – San Francisco VAMC
Steven Tobochnik, MD – VA Boston Health Care System
Zulfi Haneef, MBBS MD – Baylor College of Medicine
Rationale:
Drug-resistant epilepsy (DRE) remains underrecognized and undertreated within the Veterans Health Administration (VHA) due to barriers such as inconsistent referral patterns, variable provider awareness regarding DRE and referral options, and care expertise and capacity differences across sites. Recently, a validated algorithm was developed to identify DRE cases using VHA administrative data. This project aimed to develop a scalable, data-informed framework to integrate this algorithm into clinical workflows and enhance the timely identification and referral of veterans with DRE.Methods:
A multidisciplinary panel of experts including epileptologists, general neurologists, nurse practitioners, implementation scientists, and statisticians employed a Nominal Group Technique (NGT), to systematically identify, evaluate, and prioritize implementation strategies for the DRE algorithm. The NGT included idea generation, round robin sharing, clarification and discussion, scoring of ideas, and subsequent discussion of barriers and common elements among integration ideas. The process facilitated structured discussions to achieve consensus on key elements of algorithm integration (Table 1). Discussion points were then used to synthesize practical recommendations for algorithm integration into routine clinical practice across the VHA.Results:
The 15-member NGT panel developed consensus recommendations to help identify and refer Veterans with DRE for specialized care (Figure 1). Using existing VA dashboards (EpiData, Neurology Cube) and electronic health records to automatically flag patients who may need specialty care, the following priority system was suggested to help focus outreach: (1) Veterans with no neurology visits, (2) Veterans seen by general neurology but not epilepsy specialists, and (3) Veterans lost to specialty follow-up. To avoid overwhelming providers, simple tools like dashboards, referral templates, and email reminders were recommended. The model was designed to fit into current workflows and can be scaled across VA sites. Periodic re-evaluation and re-adjustment of the model to iteratively improve performance was recommended.
Conclusions:
This structured, consensus-based model offers a practical, scalable approach for integrating of a DRE detection algorithm into routine care by addressing existing gaps in DRE management across the VHA. These recommendations can enhance care coordination, reduce treatment delays, and improve outcomes for veterans living with DRE. These recommendations could also inform strategies to address similar challenges in managing other chronic, under-referred conditions in complex health systems.
Funding: This project did not receive funding support.