Development and validation of a predictive model for febrile seizures

Area of Science:

• Pediatrics
• Clinical Prediction Modeling
• Hematology

Background:

• Febrile seizures (FS) are common in children, but current predictive models lack accuracy.
• Accurate prediction of FS is crucial for timely intervention and management.

Purpose of the Study:

• To develop and validate a high-performing predictive model for febrile seizures (FS) in children.
• To identify key predictors of FS using demographic data and complete blood counts (CBCs).

Main Methods:

• Retrospective analysis of febrile children from July 2020 to March 2021.
• Utilized the least absolute shrinkage and selection operator (LASSO) for predictor selection.
• Developed a predictive model using multivariate logistic regression and constructed a nomogram.

Main Results:

• The predictive model demonstrated strong performance with an Area Under the Curve (AUC) of 0.884 in the training set.
• Internal and external validation showed robust predictive capabilities with AUCs of 0.883 and 0.858, respectively.
• The nomogram also exhibited good predictive accuracy with a C-statistic of 0.884.

Conclusions:

• A novel predictive model for FS based on CBCs exhibits significant clinical utility.
• The developed model offers improved predictive performance compared to existing methods.
• This tool can aid clinicians in assessing the risk of febrile seizures in pediatric patients.