摘要
目的 系统评价低危初产妇阴道试产失败风险预测模型,为临床实践及未来研究提供参考.方法 聚焦低危初产妇阴道试产失败风险预测模型,检索中国知网、万方数据库、中国生物医学文献数据库、PubMed、Embase、Cochrane Library、CINAHL、Web of Science等数据库中与低危初产妇阴道试产失败风险预测模型有关的文献,检索时限均为建库至2023年9月20日,由2名研究员独立筛选文献、提取资料并评价纳入研究的偏倚风险.结果 共纳入10篇文献,涉及14个模型,其中12个模型采用Logistic回归分析建模,模型的受试者工作特征曲线下面积为0.680-0.847.模型常见预测因素包括孕妇年龄、孕妇身高、孕前体重指数、孕期增重、孕周、胎儿出生体重、引产方式、胎儿生物学测量值、临产前宫颈Bishop评分等.结论 低危初产妇阴道试产失败风险预测模型的预测性能较好,但模型构建方式较为单一,整体偏倚风险较高,需要借助机器学习、人工智能等手段进一步优化模型,并进行广泛的外部验证.
Abstract
Objective To systematically review the risk prediction model for vaginal delivery failure in low-risk primiparas,so as to provide reference for clinical practice and future research.Methods Literature on the risk prediction model for vaginal delivery failure in low-risk primiparas was retrieved from China National Knowledge Infrastructure,WanFang Data,China Biology Medicine disc,PubMed,Embase,Cochrane Library,CINAHL,Web of Science and so on.The search period was from the establishment of the database to September 20,2023.Two researchers independently screened literature,extracted data,and evaluated the risk of bias in the included studies.Results A total of 10 articles were included,involving 14 models.Twelve models were constructed using Logistic regression,and the area under the receiver operating characteristic curve of the models ranged from 0.680 to 0.847.Common predictive factors of the model involved age of pregnant woman,height of pregnant woman,pre-pregnancy body mass index,gestational weight gain,gestational week,fetal birth weight,induction method,fetal biological measurements,and pre-delivery cervical Bishop score.Conclusions The predictive performance of the vaginal delivery failure risk prediction model for low-risk primiparas is good,but the model construction method is relatively simple,and the overall bias risk is high.The model needs to be further optimized through machine learning,artificial intelligence,and other means,and extensively validated externally.
维普
万方数据