产后2个月内非纯母乳喂养风险预测评分模型的构建和验证

Construction and validation of risk prediction scoring model for non exclusive breastfeeding two months after delivery

目的:构建产后2个月内非纯母乳喂养风险预测评分模型,进行风险分层后,对模型预测效能进行验证。方法:对2019年7月至2021年8月间中国人民解放军联勤保障部队第九○九医院分娩的产妇961例、福建省妇幼保健院分娩产妇426例进行问卷调查,将中国人民解放军联勤保障部队第九○九医院分娩产妇随机分为建模组(752例)和本院验证组(209例),通过二元Logistic回归分析筛选产后2个月内非纯母乳喂养影响因素,构建风险预测模型,并从3个层次对模型的预测能力进行验证。结果:年龄≥30岁、配偶母乳喂养支持量表评分(PBIS)<60、多胎、未参加孕期健康宣教、母乳喂养自我效能量表评分(BSES)<105、分娩后1 h内无皮肤接触和哺乳、乳头异常、乳头皲裂、乳腺炎、新生儿黄疸均为产后2个月内非纯母乳喂养的独立影响因素,分别赋值1、2、7、2、2、2、13、3、2、2。建模组模型受试者工作特征曲线的曲线下面积(AUC)为0.755(0.717~0.793),Hosmer-Lemeshow(H-L)检验结果为χ^(2)=3.854,P=0.216;本院验证组模型AUC为0.764(0.693~0.834),H-L检验结果为χ^(2)=4.302,P=0.266;外院验证组模型AUC为0.720(0.670~0.769),H-L检验结果为χ^(2)=5.302,P=0.102。对产后2个月内非纯母乳喂养进行风险分层,0~4分为低非纯母乳喂养风险,5~13分为中非纯母乳喂养风险,≥14分为高非纯母乳喂养风险。结论:构建的产后2个月内非纯母乳喂养风险预测评分模型有较好的区分度和校准度,方便临床对产妇喂养方式进行预测,并及时提供相关指导和干预。

Objective:To construct a risk prediction scoring model for non exclusive breastfeeding in two months after delivery,and verify the prediction efficiency of the model.Methods:961 parturients in the 909 Hospital of Joint Logistics Support Force and 426 parturients in Fujian Maternal and Child Health Hospital from July 2019 to June 2021 were investigated by questionnaire and followed up.The parturients in the 909 Hospital of Joint Logistics Support Force were randomly divided into modeling group(752 cases)and validation group(209 cases).The influencing factors of non exclusive breastfeeding in two months after delivery were screened by binary logistic regression analysis,the risk prediction model was constructed,and the prediction ability of the model was verified in three levels.Results:Age≥30 years,partner support breastfeeding influence scale(PBIS)<60,multiple births,not participating in pregnancy health education,breastfeeding self-efficacy scale(BSES)score<105,no skin contact and breastfeeding within 1 hour after delivery,abnormal nipple,chapped nipple,mastitis and neonatal jaundice were independent influencing factors of non exclusive breastfeeding in two months after delivery,with values of 1,2,7,2,2,2,13,3,2 and 2,respectively.The area under the curve(AUC)of the model group was 0.755(0.717-0.793),and the Hosmer-Lemeshow(H-L)test result wasχ^(2)=3.854,P=0.216.The AUC of the model in validation group of our hospital was 0.764(0.693-0.834),and the H-L test result wasχ^(2)=4.302,P=0.266.The AUC of the model in external validation group was 0.720(0.670-0.769),and the H-L test result wasχ^(2)=5.302,P=0.102.The risk of non exclusive breastfeeding within 2 months postpartum was stratified.0-4 points were low risk of non exclusive breastfeeding,5-13 points were medium risk of non exclusive breastfeeding,and≥14 points were high risk of non exclusive breastfeeding.Conclusion:The risk prediction score model of postpartum non exclusive breastfeeding in two months after delivery has good discrimination and calibration,which is convenient for clinical prediction of maternal feeding mode,and provided relevant guidance and intervention in time.