Individualized treatment of epilepsy based on real-world risks of treatment failure and side effects

Epilepsy is treated with a large number of antiseizure medications (ASMs).

However, the knowledge regarding which patient groups are more vulnerable to treatment failure and/or specific long-term side effects of certain ASMs is limited. Suboptimal treatment can lead to seizures, reduced quality of life, injury, and death.

Clinical trials only assess tolerability, but not long-term side effects like osteoporosis, cardiovascular events, or cognitive decline.Here we will use big data and machine learning to determine side effect vulnerability and risk-benefits, with the goal to provide individualized ASM selection and improved treatment.

We will create a database of >100 000 epilepsy patients, track ASM use, and analyze risk factors of 1) discontinuation; 2) epilepsy outcomes (seizures/trauma/death); and 3) side effects.

We will model how ASM choice affects risks of different outcomes in relation to patient characteristics (age, sex, number of ASMs, epilepsy severity etc).

The project builds on our combined epilepsy and big data competence; collaborators include data scientists and international epilepsy experts.My group has published some of the largest studies to date on risk factors and prognosis of many forms of epilepsy, demonstrating a high potential for Swedish registers to capture real-world outcomes in relation to ASM exposure.

Thus, the project is likely to lead to new guidelines on individualized ASM treatment.