Increasingly Complex Epilepsy Therapy Decisions

Our machine learning algorithm is a collaboration project with the Pestian Lab and Cincinnati Children’s divisions of biostatistics, emergency medicine, informatics, neurology and neurosurgery.

AI and machine learning are tools that allow us to link databases and mine them for information we can use. The field of biomedical informatics takes the field from data to wisdom. Our algorithm fuses multiple data modalities.

We use natural language processing to optimize structured and unstructured data from patient EHRs. This includes: 

• Anti-seizure medications
• Demographics
• Electroencephalogram (EEG) reports (free text)
• Magnetic resonance imaging (MRI) reports (free text)
• Neurology visit notes (free text)
• Procedure and lab orders
• Visit patterns

The AI notices subtle changes in a physician’s written language that indicate a patient is a candidate for surgery conference earlier in the disease course. The tool sends alerts to physicians before patient appointments to call attention to a possible surgical referral.

Future Applications for Our Machine Learning Algorithm

Our research team is now deploying the epilepsy algorithm at other health centers. The goal is to test the model for use in adult patients, at other pediatric sites and at community hospitals.