Risk prediction for postpartum depression based on random forest.
10.11817/j.issn.1672-7347.2020.190655
- Author:
Meili XIAO
1
;
Chunli YAN
2
;
Bing FU
3
;
Shuping YANG
4
;
Shujuan ZHU
3
;
Dongqi YANG
5
;
Beimei LEI
5
;
Ruirui HUANG
5
;
Jun LEI
6
Author Information
1. Xiangya Nursing School, Central South University, Changsha 410013. 786001878@qq.com.
2. Department of Oncology, Third Xiangya Hospital, Central South University, Changsha 410013.
3. Department of Obstetrics and Gynaecology, Third Xiangya Hospital, Central South University, Changsha 410013.
4. School of Mathematics and Statistics, Central South University, Changsha 410013.
5. Xiangya Nursing School, Central South University, Changsha 410013.
6. Xiangya Nursing School, Central South University, Changsha 410013. leijunbao@126.com.
- Publication Type:Journal Article
- Keywords:
influencing factors;
postpartum depression;
random forest;
risk prediction
- MeSH:
Depression, Postpartum/epidemiology*;
Female;
Humans;
Postpartum Period;
Pregnancy;
Pregnancy Trimester, Third;
Psychiatric Status Rating Scales;
Risk Factors;
Sensitivity and Specificity
- From:
Journal of Central South University(Medical Sciences)
2020;45(10):1215-1222
- CountryChina
- Language:English
-
Abstract:
OBJECTIVES:To explore the application of random forest algorithm in screening the risk factors and predictive values for postpartum depression.
METHODS:We recruited the participants from a tertiary hospital between June 2017 and June 2018 in Changsha City, and followed up from pregnancy up to 4-6 weeks postpartum.Demographic economics, psychosocial, biological, obstetric, and other factors were assessed at first trimesters with self-designed obstetric information questionnaire and the Chinese version of Edinburgh Postnatal Depression Scale (EPDS). During 4-6 weeks after delivery, the Chinese version of EPDS was used to score depression and self-designed questionnaire to collect data of delivery and postpartum. The data of subjects were randomly divided into the training data set and the verification data set according to the ratio of 3꞉1. The training data set was used to establish the random forest model of postpartum depression, and the verification data set was used to verify the predictive effects via the accuracy, sensitivity, specificity, positive predictive value, negative predictive value, and AUC index.
RESULTS:A total of 406 participants were in final analysis. Among them, 150 of whom had EPDS score ≥9, and the incidence of postpartum depression was 36.9%. The predictive effects of random forest model in the verification data set were at accuracy of 80.10%, sensitivity of 61.40%, specificity of 89.10%, positive predictive value of 73.00%, negative predictive value of 82.80%, and AUC index of 0.833. The top 10 predictive influential factors that screening by the variable importance measure in random forest model was antenatal depression, economic worries after delivery, work worries after delivery, free triiodothyronine in first trimesters, high-density lipoprotein in third trimester, venting temper to infants, total serum cholesterol and serum triglyceride in first trimester, hematocrit and serum triglyceride in third trimester.
CONCLUSIONS:Random forest has a great advantage in risk prediction for postpartum depression. Through comprehensive evaluation mechanism, it can identify the important influential factors for postpartum depression from complex multi-factors and conduct quantitative analysis, which is of great significance to identify the key factors for postpartum depression and carry out timely and effective intervention.
