摘要
目的 探讨基于机器学习的糖尿病母亲婴儿(infant of diabetic mother,IDM)低阿氏评分的影响因素。方法选取2019年1月1日至2021年12月31日首都医科大学附属北京妇产医院孕期诊断为妊娠期IDM 6 187例,根据新生儿出生后1 min的阿氏评分,分为低评分组(70例)及正常评分组(6 117例),以新生儿体重和身长为混淆因素,采用倾向性得分1∶1的比例进行匹配,匹配成功后,采用过4种机器学习方法筛选低阿氏评分的影响因素,同时对筛选出的危险因素采用多因素logistic回归方程进行验证。结果 倾向性得分成功匹配65例。产妇年龄、分娩次数和胎儿宫内窘迫(胎心型)是4种机器学习方法中共有的影响因素。多因素logistic回归分析结果显示,产妇年龄≥35岁(OR=0.456,95%CI:0.252~0.824,P=0.009),分娩多次(OR=0.225,95%CI:0.057~0.898,P=0.035)是IDM低阿氏评分保护因素,有胎儿宫内窘迫(胎心型)(OR=4.840,95%CI:1.770~13.232,P=0.002)是IDM低阿氏评分的危险因素。结论 结合产妇年龄、分娩次数和胎儿宫内窘迫(胎心型)等因素综合判断,可以为IDM预后提供参考。
Objective To explore the influencing factors of low Appar score in infant of diabetic mother(IDM) based on machine learning.Methods A total of 6 187 infants of diabetic mothers during pregnancy in Beijing Obstetrics and Gynecology Hospital affiliated to Capital Medical University from January 1,2019 to December 31 were selected,and were divided into lower Apgar score group(n = 70) and normal Apgar score group(n = 6 117) according to the Apgar score at1 minute after birth.Taking the length and weight of the fetus as the confounding factors,the two groups were matched by propensity score of 1∶1.Risk factors were screened by four machine learning methods,and the selected risk factors were evaluated by multivariate logistic regression analysis.Results After propensity score matching,65 cases were successful matched.Maternal age,delivery times and fetal distress(fetal heart-type) were selected as the prenatal risk factors for low Apgar score of IDM by the four maching learning methods.These risk factors were evaluated by multivariate logistic regression.The results showed that maternal age ≥ 35 years old(OR = 0.456,95%Cl:0.252-0.824,P = 0.009) and multiple deliveries(OR = 0.225,95% Cl:0.057-0.898,P = 0.035) were protective factors for low Apgar scores in IDM and fetal distress(fetal heart-type)(OR = 4.840,95% Cl:1.770-13.232,P = 0.002) were risk factor for low Apgar scores in IDM.Conclusions Combined with the comprehensive judgment of maternal age,delivery times and fetal distress(fetal hearttype),it can help preliminarily judge the prognosis of IDM.
作者
高正平
程序
寇晨
Gao Zhengping;Cheng Xu;Kou Chen(Department of Neonatology,Beijing Obstetrics and Gynecology Hospital,Capital Medical University,Beijing Maternal and Child Health Care Hospital,Beijing 100026,China)
出处
《北京医学》
CAS
2023年第4期325-330,共6页
Beijing Medical Journal
关键词
机器学习
糖尿病母亲婴儿
阿氏评分
影响因素
倾向性得分匹配
machine learning
infant of diabetic mother(IDM)
Apgar score
influencing factors
propensity score matching(PSM)
作者简介
通信作者:寇晨,Email:kcviva@ccmu.edu.cn。
