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

Area of Science:

• Neurology
• Clinical Medicine
• Medical Informatics

Background:

• Accurate seizure type classification is crucial for effective therapeutic decision-making in epilepsy management.
• Existing classification methods can be complex, posing challenges for timely and accurate diagnosis, particularly in non-specialist settings.

Purpose of the Study:

• To develop and validate a pragmatic, web-based algorithm for classifying seizure types.
• To facilitate therapeutic decision-making by providing a simple tool for healthcare professionals.

Main Methods:

• A modified Delphi method was used by five experts to develop a classification of nine seizure types and a seven-question algorithm.
• The algorithm was prospectively validated in a multicenter study involving junior physicians, nurses, and physician assistants.
• Agreement was assessed against expert epileptologist classification, and feasibility was evaluated by healthcare professionals globally.

Main Results:

• The algorithm achieved 83.2% agreement with expert classification (AC1 = .82), indicating almost perfect agreement.
• The web-based tool was found to be applicable and useful by 32 healthcare professionals from 14 countries.
• The study included 262 patients with various seizure types and nonepileptic events.

Conclusions:

• The developed web-based algorithm offers an accurate and pragmatic method for classifying seizure types in adolescents and adults.
• This tool can significantly aid in the selection of appropriate antiseizure medications, improving epilepsy treatment.
• The algorithm's ease of use and validated accuracy support its integration into clinical practice, especially in underserved areas.