EEC 2024 | Developing a clinical prediction model to predict seizure risk in patients with epilepsy

Thomas Spain, University of Liverpool, Liverpool, UK, discusses the development of a clinical prediction model to predict seizure risk in patients with epilepsy. Unlike traditional clinical models that predict single events, epilepsy requires the use of recurrent event prediction models as seizures occur regularly. This model uses two Cox proportional hazards model extensions, the Andersen-Gill and Prentice-Williams-Peterson extensions, to allow for precise timing predictions between seizures while minimizing risks of overfitting. Validation involved historical and recent datasets from the SANAD trials, providing confidence in the model’s applicability to modern patient populations. This interview took place at the 15th European Epilepsy Congress (EEC) 2024 in Rome, Italy.

These works are owned by Magdalen Medical Publishing (MMP) and are protected by copyright laws and treaties around the world. All rights are reserved.