新生儿癫痫
发作类型
医学
科恩卡帕
卡帕
脑电图
癫痫
儿科
听力学
精神科
计算机科学
语言学
机器学习
哲学
作者
Elissa Yozawitz,Maria Roberta Cilio,Eli M. Mizrahi,Jee‐Young Moon,Solomon L. Moshé,Magda Lahorgue Nunes,Perrine Plouin,Sampsa Vanhatalo,Sameer M. Zuberi,Ronit Pressler
摘要
The International League Against Epilepsy (ILAE) Neonatal Seizure Framework was tested by medical personnel.Attendees at the 2016 ILAE European Congress on Epileptology in Prague, the International Video-EEG Course in Pediatric Epilepsies in Madrid 2017, and a local meeting in Utrecht 2018, were introduced to the proposed ILAE neonatal classification system with teaching videos covering the seven types of clinical seizures in the proposed neonatal classification system. Five test digital video recordings of electroencephalography (EEG)-confirmed motor neonatal seizures were then shown and classified by the rater based on their knowledge of the proposed ILAE Neonatal Seizure Framework. A multi-rater Kappa statistic was used to assess the agreement between observers and the true diagnosis.The responses of 194 raters were obtained. There was no single predominant classification system that was currently used by the raters. Using the ILAE framework, 78%-93% of raters correctly identified the clinical seizure type for each neonate; the overall inter-rater agreement (Kappa statistic) was 0.67. The clonic motor seizure type was most frequently accurately identified (93% of the time; κ = 0.870). EEG technicians correctly identified all presented motor seizure types more frequently than any other group (accuracy = 0.9).The ILAE Neonatal Seizure Framework was judged by most raters to be better than other systems for the classification of clinical seizures. Among all seizure types presented, clonic seizures appeared to be the easiest to accurately identify. Average accuracy across the five seizure types was 84.5%. These data suggest that the ILAE neonatal seizure classification may be used by all healthcare professionals to correctly identify the predominant clinical seizure type.
科研通智能强力驱动
Strongly Powered by AbleSci AI