医学
布里氏评分
接收机工作特性
心脏外科
队列
危险分层
体外循环
回顾性队列研究
儿科
外科
急诊医学
内科学
计算机科学
人工智能
作者
Claudia Cattapan,Alvise Guariento,Jeffrey P. Jacobs,Mark S. Bleiweis,Zdzisław Tobota,Bohdan Maruszewski,Steven J. Staffa,David Zurakowski,George E. Sarris,Vladimiro L. Vida
标识
DOI:10.1093/ejcts/ezaf178
摘要
Abstract OBJECTIVES Current preoperative counselling in neonatal cardiac surgery is mainly focused on the primary procedure. However, other factors must be considered when evaluating the surgical risk of a neonate. We aimed to develop a risk adjustment model to personalize preoperative counselling using data from the European Congenital Heart Surgeons Association Congenital Database (ECHSA-CD). METHODS A retrospective, multicentre analysis of the ECHSA-CD dataset was conducted, including 20 687 neonates undergoing cardiac surgery between 2013 and 2022. A risk adjustment model was developed on a training set (70%) and validated on a separate cohort (30%). RESULTS A model incorporating age, weight, STAT mortality category and need for cardiopulmonary bypass (CPB) demonstrated good predictive performance. Lower age (≤10 days), lower weight (<2.5 kg), higher STAT category and need for CPB were associated with increased risk of operative mortality. The model’s area under the receiver operating characteristic curve was 0.701 in the training set and 0.700 in the validation set, indicating good discrimination. Additionally, the Brier quadratic probability score was 0.08 in both datasets, indicating good calibration. CONCLUSIONS This study underscores the importance of patient characteristics in predicting outcomes in neonatal cardiac surgery. The developed risk adjustment model can be used as a tool in preoperative counselling, decision-making and risk stratification for neonates undergoing cardiac surgery. By providing a more accurate estimate of operative mortality, this model can help families make more informed decisions about their child’s care and improve the overall quality of care for neonates with congenital heart defects.
科研通智能强力驱动
Strongly Powered by AbleSci AI