A web‐based algorithm to rapidly classify seizures for the purpose of drug selection

Abstract

Objective: To develop and validate a pragmatic algorithm that classifies seizure types, to facilitate therapeutic decision-making. Methods: Using a modified Delphi method, five experts developed a pragmatic classification of nine types of epileptic seizures or combinations of seizures that influence choice of medication, and constructed a simple algorithm, freely available on the internet. The algorithm consists of seven ques-K E Y W O R D S algorithm, classification, epilepsy, seizure, web-based application Key Points • We have developed an algorithm that, based on seven questions, classifies seizures for the purpose of drug selection • We have validated the algorithm, implemented in a web-based application (Epipi ck.org), in a prospective, multicenter study on 262 patients • Agreement between the algorithm and the expert classification was 83.2%, with an agreement coefficient .82 (almost perfect agreement)

NLP-based tools for localization of the Epileptogenic Zone in patients with drug-resistant focal epilepsy

“…A great motivation for the research in this sector derives from the will to speed up the process that leads the patient to be considered a candidate for surgery and also to find alternative types of information that healthcare professionals can collect more easily without necessarily requiring highly specialized diagnostic machinery. Recently, many research projects aimed at developing decision support systems that could provide a more accurate classification of seizures relying on questionnaires [19]. Studies based on self-reported questionnaires are highly dependent on patients’ willingness to personally contribute to the study.…”

“…As regards the epilepsy field, literature reports different tools developed as answers to specific needs. For instance, the authors of [19] present a decision support system that exploits 7 questions addressed to the patients providing an accurate classification of seizure types. Epifinder [20] is a clinical decision support system for epilepsy diagnosis which, given some keywords extracted from the semiological descriptions of the seizures, returns the probability to being affected by epilepsy.…”

Choose Your Own Adventure: Expert Advice at Your Fingertips

“…2,3 Beniczky and colleagues have tackled this problem with a rapid algorithmic web tool (EpiPick.org) intended to provide clinicians, especially in underserved regions, instant expert advice. Using iterative Delphi methods, they had 2 goals: 1) To develop a simple algorithm classifying seizure types, 4 and 2) To determine optimal antiseizure medication (ASM) choices for each seizure type. 5,6 For the first goal (classifying seizure types), 5 expert epileptologists met to agree upon the minimum number of distinct seizure types that would influence treatment decisions.…”

The EpiPick algorithm to select appropriate antiseizure medications in patients with epilepsy: Validation studies and updates

“…Following the publication of the algorithm, we have further validated the app in three studies. In a large, multicenter, prospective study, 2 we validated the first part of the algorithm, which classifies seizure types to facilitate therapeutic decision-making. Agreement between the algorithm and the expert classification was 83.2% [95% confidence interval (CI) 78.6%-87.8%], with an agreement coefficient (AC1) of 0.82 (95% CI 0.77-0.87), indicating almost perfect agreement.…”

“…Agreement between the algorithm and the expert classification was 83.2% [95% confidence interval (CI) 78.6%-87.8%], with an agreement coefficient (AC1) of 0.82 (95% CI 0.77-0.87), indicating almost perfect agreement. 2 In another study 3 we investigated the agreement among experts in selecting an ASM as initial monotherapy and used their choices to validate the app. The percent agreement between the highest ranked selections of the app and the expert selections was 73% (95% CI 64%-82%).…”