Modeling recurrent epileptic seizures in patients with focal epilepsy
Conference
64th ISI World Statistics Congress - Ottawa, Canada
Format: IPS Abstract
Keywords: "children, longitudinal
Session: IPS 161 - Modeling complex correlated data: new directions and innovations
Thursday 20 July 10 a.m. - noon (Canada/Eastern)
Abstract
Accurate seizure tracking is critically important for clinical management of people with epilepsy and for clinical research. Seizure frequency is typically the primary outcome for treatment trials and is an important metric in relation to other disease-related outcomes such as comorbidities, quality of life, and indirect consequences of seizures. For these reasons, gathering precise data on seizure frequency is important. However, there are barriers to seizure tracking. Retrospective questionnaires and clinical interviews are subject to recall bias, which is compounded by the fact that many seizures cause loss of awareness and can be difficult for patients to remember. Motivated by the Human Epilepsy Project (HEP), a dataset of epileptic seizures in patients with focal epilepsy, we develop a longitudinal model of epileptic seizures that addresses tracking patterns. The HEP dataset comes from a six-year international observational study whose primary goal was to identify clinical characteristics predictive of disease outcome, progression, and treatment response in 450 participants with newly treated focal epilepsy. We have identified three subgroups of tracking patterns: good, medium and poor seizure trackers. Seizure occurrence was explored across the three subgroups.