Development of a standard for epilepsy semiology and semiological description dataset: a modified Delphi study

Epilepsy semiology is a major component of epilepsy diagnosis and plays a crucial role in its clinical management. This study aimed to develop the "Standard for Epilepsy Semiology and Semiological Description Dataset" to provide a reference for clinical practice and research, and to promote the standardization of relevant data. This work lays the foundation for the further development and construction of specialized epilepsy databases. A modified Delphi method was employed to reach a consensus on the framework and content of the dataset.

Two rounds of the Delphi process were conducted. Participants included experts from the fields of neurosurgery, neurology, pediatrics, and clinical neurophysiology, all specializing in epilepsy. The development was primarily based on the 2025 International League Against Epilepsy (ILAE) "Updated Classification of Seizures by the International League Against Epilepsy: Position Statement". The draft was also informed by previously published guidelines and consensus statements, case reports, and semiological reports extracted from clinical examination records of epilepsy patients. Data analysis included the expert positive coefficient, expert authority coefficient (Cr), expert coordination coefficient (Kendall's W), and coefficient of variation (CV) to reflect reliability and consensus.

Consensus was reached on 285 key indicators for the dataset standard. To organize the data systematically, the standard was designed into two sections: the "Standard for Epilepsy Semiology and Semiological Description Dataset - Main Table"and the "Standard for Epilepsy Semiology and Semiological Description Dataset - Semiological Description Table". The first section includes 10 Level-1 indicators and 93 Level-2 indicators. The second section includes 6 Level-1 indicators, 88 Level-2 indicators, and 88 Level-3 indicators (encompassing 832 free-text fields). The expert authority coefficient was 0.83 for both rounds. Kendall's coefficient of concordance (W) was 0.222 (p < 0.001) in the first round and 0.175 (p < 0.001) in the second round, indicating a statistically significant degree of consistency among the experts.

This study employed a modified Delphi method to develop the "Standard for Epilepsy Semiology and Semiological Description Dataset". By standardizing and integrating relevant terminology, this standard incorporates Chinese-specific synonym fields, aiming to provide a reliable data foundation for clinical practice and scientific research. The outcomes not only offer standardized references for specialists but also serve as practical tools for grassroots physicians in epilepsy semiological data collection, thereby effectively promoting the standardization of data in this field and laying a solid foundation for the in-depth utilization, sharing of epilepsy data, and the construction of specialized disease databases.