产后早期非母乳喂养影响因素分析及列线图预测模型建立

Influence factors of the early postpartum non breastfeeding and the establishment of nomograph predictive model for the verification

目的:探索影响产后早期非母乳喂养的相关因素并构建列线图预测模型验证。方法:调查本院2020年6月-2021年9月分娩的产妇541例,根据是否进行早期母乳喂养分为非母乳喂养组(n=169)及母乳喂养组(n=372),采用logistic回归分析影响母乳喂养的相关因素并构建预测模型,选取同期103例孕妇作为验证组,对列线图进行验证。结果:母乳喂养组与非母乳喂养组年龄、新生儿是否进NICU、即刻母婴接触、产次(≥2次)、配偶母乳喂养支持量表(PBIS)<60分、母乳喂养喂养自我效能量表(BSES)评分<105分、乳头异常及乳腺炎均有差异(P<0.05)。行logistic分析显示,产妇年龄≥30岁、即刻母婴接触、产次(≥2次)、PBIS<60分、乳头异常以及乳腺炎均为影响产后早期非母乳喂养因素。基于logistic回归分析结果生成了产后早期非母乳喂养风险列线图预测模型,其曲线下面积(AUC)为0.858(95%CI 0.729~0.903),验证组模型的AUC为0.784(0.684~0.841),模型在验证组有良好的区分度和校准度。结论:基于产妇年龄、母婴接触、产次、PBIS得分、乳头异常以及乳腺炎构建的产后早期非母乳喂养列线图预测模型有较好的校准度和区分度,具有临床指导价值。

Objective:To explore the relevant factors affecting the early postpartum non breastfeeding,and to construct a nomograph predictive model for the verification.Methods:A retrospective analysis was conducted on 541puerperas who were delivery in hospital from June 2020to September 2021.These puerperas were divided into group A(169puerperas with early non breastfeeding)and group B(372puerperas with early breastfeeding)according to whether early breastfeeding or not.Logistic regression analysis was used to analyze the relevant factors affecting breast-feeding,and the nomogram predictive model was established.In addition,103puerperas were selected as the validation group to verify the nomogram predictive model during the same period.Results:There were significant differences in the age,whether the neonates transfered to NICU or not,the immediate mother-infant contact,the number of births(≥2 times),the score<60points of the spousal breastfeeding Support Scale(PBIS),the score<105points of the breast-feeding Feeding Self-efficacy Scale(BSES)of the puerperas between group B and group A(P<0.05).Logistic analysis showed that the age≥30years old,the immediate maternal contact,the number of delivery(≥2times),the PBIS score<60points,the abnormal nipples,and the mastitis of the puerperas were all the factors affecting their early postpartum non breastfeeding.Based on the results of Logistic regression analysis,a nomograph predictive model for the risk of early postpartum non breastfeeding was established,and the area under the curve(AUC)of which was 0.858(95%CI:0.729-0.903).The AUC of the model in the validation group was 0.784(95%CI:0.684-0.841),and the model had good differentiation and calibration in the validation group.Conclusion:The nomogram predictive model of early postpartum non breast feeding based on the age,the mother-infant contact,the number of delivery,the PBIS score,the nipple abnormalities,and the mastitis of the puerperas has good calibration and differentiation,and which has clinical guiding value.