Ongoing Pediatric Epilepsy Research Consortium Initiatives and the Future of Pediatric Epilepsy Care: Anup Patel, MD
AnUP PATEL, MD
TRANSCRIPT
Interviewer: There are two grants currently underway: the Lennox-Gastaut Syndrome Foundation grant and the PERC grant on status epilepticus. Can you talk a little bit about those two projects and what they're working on?
Anup Patel, MD: Absolutely. I think this is where we're really going to learn how large language models and augmented intelligence can help inform care and care trajectories, with the ultimate goal of improving outcomes.
I'll start with the Lennox-Gastaut Syndrome Foundation grant. We were fortunate to receive a small pilot grant to explore whether a large language model can read clinical notes and identify children who may have Lennox-Gastaut syndrome but have not yet received the diagnosis. We also have the advantage of being able to compare its performance against a gold-standard registry that's part of another research grant and cohort.
We can use patients with verified Lennox-Gastaut syndrome to teach the model what to look for in the medical record. Then we can apply that model outside of the original cohort to identify additional patients who may be missing the diagnosis.
The hope is that, working through the PERC LGS Special Interest Group, we'll be able to develop a larger federal grant opportunity that allows centers across the country to do the same work, eventually expanding nationally and potentially even internationally.
That effort led to a second project through PERC. One of the exciting developments at PERC is that we're now able to offer pilot research grants. We've never been able to do that before, and it's been a tremendous growth opportunity for the organization.
Krista Eschbach, who received one of these grants, is studying pediatric status epilepticus. The idea is similar: Can we train a large language model to identify children who present with status epilepticus by analyzing clinical notes and other available data? Can it also determine what medications were administered, when they were given, and how those interventions influenced outcomes?
We're piloting that work at Cornell, Boston Children's Hospital, Nationwide Children's Hospital, and Children's Hospital Colorado, with Krista serving as principal investigator. It's exactly the kind of project that would not have been possible without the grant funding we've been able to provide.
Interviewer: That's fantastic. These are obviously major institutions with significant resources behind them, so it's really exciting to see this work moving forward. It also seems like you're adapting to many of the broader trends around AI in neurology, which brings me to my next question.
A lot is happening with AI across medicine. What excites you most about AI in epilepsy specifically, and what excites you more broadly across neurology?
Patel: I think, at a global level, what excites me most is how we're going to incorporate augmented intelligence and other technological solutions to address the challenges we face, both in delivering care and in improving the systems that support that care.
We're dealing with two realities. One is how we care for patients with the disease in front of us. The other is how we improve efficiency within our healthcare systems so clinicians can spend more time doing what matters most: caring for patients and families.
Throughout this meeting, you're seeing presentations that touch on all aspects of that. How can AI help document clinical information? How can it support research? How can it improve workflows? The possibilities are extensive.
We're also seeing tremendous interest in wearable technologies. After this interview, I'll be giving a presentation on a paper we recently published that was featured by NPR this morning. The paper explores how patients are increasingly coming to neurology providers with data collected from smart devices such as watches, rings, and other wearable technologies, asking clinicians to help interpret that information.
You're starting to see that trend emerge across neurology and across many different disease states. It's becoming an increasingly important part of how we think about monitoring, data collection, and ultimately patient care.
Transcript edited for clarity.