Trajectory of breastfeeding among Chinese women and risk prediction models based on machine learning: a cohort study
Abstract Background Breastfeeding is the optimal source of nutrition for infants and young children, essential for their healthy growth and development. However, a gap in cohort studies tracking breastfeeding up to six months postpartum may lead caregivers to miss critical intervention opportunities...
Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
BMC
2024-12-01
|
| Series: | BMC Pregnancy and Childbirth |
| Subjects: | |
| Online Access: | https://doi.org/10.1186/s12884-024-07010-z |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1846100836261822464 |
|---|---|
| author | Yi Liu Jie Xiang Ping Yan Yuanqiong Liu Peng Chen Yujia Song Jianhua Ren |
| author_facet | Yi Liu Jie Xiang Ping Yan Yuanqiong Liu Peng Chen Yujia Song Jianhua Ren |
| author_sort | Yi Liu |
| collection | DOAJ |
| description | Abstract Background Breastfeeding is the optimal source of nutrition for infants and young children, essential for their healthy growth and development. However, a gap in cohort studies tracking breastfeeding up to six months postpartum may lead caregivers to miss critical intervention opportunities. Methods This study conducted a three-wave prospective cohort analysis to examine maternal breastfeeding trajectories within the first six months postpartum and to develop risk prediction models for each period using advanced machine learning algorithms. Conducted at a leading Maternal and Children's hospital in China from October 2021 to June 2022, data were gathered via self-administered surveys and electronic health records. Results Of the 3307 women recruited, 3175 completed the surveys, yielding a 96% effective response rate. Breastfeeding(BF) rates were observed at 100%, 96%,93% and 83% at discharge, 42 day, 3 month and 6 month postpartum, respectively. Exclusively breastfeeding(EBF) rates were recorded at 91%, 64%,72% and 58% for the same intervals. Among the five machine learning methods employed, Random Forest (RF) demonstrated superior accuracy in predicting breastfeeding patterns, with classification accuracies of 0.629 and an area under the receiver operating characteristic curve (AUC) of 0.8122 at 42 days, 0.925 and an AUC of 0.9800 at 3 months, and 0.836 and an AUC of 0.9463 at 6 months postpartum, respectively. Key predictive factors for breastfeeding at 42 days postpartum included the newborn’s birth weight and the mother’s pre-delivery and prenatal weights. Predictors for feeding type at 3 months and 6 months postpartum included early feeding types and the scores from the Breastfeeding Self-Efficacy Scale-short Form (BSES-SF) at 6 months. The predictive model based on follow-up data showed strong performance. Conclusion Breastfeeding rates slightly declined from discharge to 6 months postpartum. The breastfeeding context in this region is comparatively optimistic both within China and internationally. Factors such as newborn’s birth weight, gestational age, maternal weight management before and during pregnancy, early support and breastfeeding success, breastfeeding knowledge and self-efficacy are intricately linked to long-term breastfeeding outcomes. This study highlights critical, modifiable risk factors for early breastfeeding stages, providing valuable insights for enhancing breastfeeding intervention programs and informed decision-making. |
| format | Article |
| id | doaj-art-74829046f05545b3b651ffcb11553f82 |
| institution | Kabale University |
| issn | 1471-2393 |
| language | English |
| publishDate | 2024-12-01 |
| publisher | BMC |
| record_format | Article |
| series | BMC Pregnancy and Childbirth |
| spelling | doaj-art-74829046f05545b3b651ffcb11553f822024-12-29T12:51:43ZengBMCBMC Pregnancy and Childbirth1471-23932024-12-0124111710.1186/s12884-024-07010-zTrajectory of breastfeeding among Chinese women and risk prediction models based on machine learning: a cohort studyYi Liu0Jie Xiang1Ping Yan2Yuanqiong Liu3Peng Chen4Yujia Song5Jianhua Ren6Department of Obstetric Nursing, West China Second University Hospital, Sichuan UniversityDepartment of Obstetric Nursing, West China Second University Hospital, Sichuan UniversityDepartment of Obstetric Nursing, West China Second University Hospital, Sichuan UniversityDepartment of Obstetric Nursing, West China Second University Hospital, Sichuan UniversitySchool of Computer and Software Engineering, Xihua UniversitySchool of Computer and Software Engineering, Xihua UniversityDepartment of Obstetric Nursing, West China Second University Hospital, Sichuan UniversityAbstract Background Breastfeeding is the optimal source of nutrition for infants and young children, essential for their healthy growth and development. However, a gap in cohort studies tracking breastfeeding up to six months postpartum may lead caregivers to miss critical intervention opportunities. Methods This study conducted a three-wave prospective cohort analysis to examine maternal breastfeeding trajectories within the first six months postpartum and to develop risk prediction models for each period using advanced machine learning algorithms. Conducted at a leading Maternal and Children's hospital in China from October 2021 to June 2022, data were gathered via self-administered surveys and electronic health records. Results Of the 3307 women recruited, 3175 completed the surveys, yielding a 96% effective response rate. Breastfeeding(BF) rates were observed at 100%, 96%,93% and 83% at discharge, 42 day, 3 month and 6 month postpartum, respectively. Exclusively breastfeeding(EBF) rates were recorded at 91%, 64%,72% and 58% for the same intervals. Among the five machine learning methods employed, Random Forest (RF) demonstrated superior accuracy in predicting breastfeeding patterns, with classification accuracies of 0.629 and an area under the receiver operating characteristic curve (AUC) of 0.8122 at 42 days, 0.925 and an AUC of 0.9800 at 3 months, and 0.836 and an AUC of 0.9463 at 6 months postpartum, respectively. Key predictive factors for breastfeeding at 42 days postpartum included the newborn’s birth weight and the mother’s pre-delivery and prenatal weights. Predictors for feeding type at 3 months and 6 months postpartum included early feeding types and the scores from the Breastfeeding Self-Efficacy Scale-short Form (BSES-SF) at 6 months. The predictive model based on follow-up data showed strong performance. Conclusion Breastfeeding rates slightly declined from discharge to 6 months postpartum. The breastfeeding context in this region is comparatively optimistic both within China and internationally. Factors such as newborn’s birth weight, gestational age, maternal weight management before and during pregnancy, early support and breastfeeding success, breastfeeding knowledge and self-efficacy are intricately linked to long-term breastfeeding outcomes. This study highlights critical, modifiable risk factors for early breastfeeding stages, providing valuable insights for enhancing breastfeeding intervention programs and informed decision-making.https://doi.org/10.1186/s12884-024-07010-zBreastfeedingMachine-learningPrediction modelsFollow-up |
| spellingShingle | Yi Liu Jie Xiang Ping Yan Yuanqiong Liu Peng Chen Yujia Song Jianhua Ren Trajectory of breastfeeding among Chinese women and risk prediction models based on machine learning: a cohort study BMC Pregnancy and Childbirth Breastfeeding Machine-learning Prediction models Follow-up |
| title | Trajectory of breastfeeding among Chinese women and risk prediction models based on machine learning: a cohort study |
| title_full | Trajectory of breastfeeding among Chinese women and risk prediction models based on machine learning: a cohort study |
| title_fullStr | Trajectory of breastfeeding among Chinese women and risk prediction models based on machine learning: a cohort study |
| title_full_unstemmed | Trajectory of breastfeeding among Chinese women and risk prediction models based on machine learning: a cohort study |
| title_short | Trajectory of breastfeeding among Chinese women and risk prediction models based on machine learning: a cohort study |
| title_sort | trajectory of breastfeeding among chinese women and risk prediction models based on machine learning a cohort study |
| topic | Breastfeeding Machine-learning Prediction models Follow-up |
| url | https://doi.org/10.1186/s12884-024-07010-z |
| work_keys_str_mv | AT yiliu trajectoryofbreastfeedingamongchinesewomenandriskpredictionmodelsbasedonmachinelearningacohortstudy AT jiexiang trajectoryofbreastfeedingamongchinesewomenandriskpredictionmodelsbasedonmachinelearningacohortstudy AT pingyan trajectoryofbreastfeedingamongchinesewomenandriskpredictionmodelsbasedonmachinelearningacohortstudy AT yuanqiongliu trajectoryofbreastfeedingamongchinesewomenandriskpredictionmodelsbasedonmachinelearningacohortstudy AT pengchen trajectoryofbreastfeedingamongchinesewomenandriskpredictionmodelsbasedonmachinelearningacohortstudy AT yujiasong trajectoryofbreastfeedingamongchinesewomenandriskpredictionmodelsbasedonmachinelearningacohortstudy AT jianhuaren trajectoryofbreastfeedingamongchinesewomenandriskpredictionmodelsbasedonmachinelearningacohortstudy |