Preliminary study on the risk of macrosomia using Bayesian discriminant analysis based on prenatal records
10.16462/j.cnki.zhjbkz.2019.11.008
- Author:
Liu-yu WU
1
;
Jing-you LAN
;
Dan-dan HUANG
;
Xiao-qiang QIU
;
Mei-liang LIU
;
Qiu-li LIANG
;
Di ZHANG
;
Xiao-yun ZENG
Author Information
1. Department of Epidemiology and Health Statistics, School of Public Health, Guangxi Medical University, Nanning 530021, China
- Publication Type:Research Article
- Keywords:
Macrosomia;
Influencing factors;
Bayesian discriminant
- From:
Chinese Journal of Disease Control & Prevention
2019;23(11):1338-1341，1347
- CountryChina
- Language:Chinese
-
Abstract:
Objective To explore the clinical effect of Bayesian discriminant analysis in predicting the risk of macrosomia. Methods 169 fetal macrosomia and 169 non-macrosomia were enrolled in a 1:1 matched case-control study. Conditional Logistic regression was used to select the discriminant indexes,and the discriminant indexes were put into the Bayesian discriminant model to obtain the Bayesian discriminant function. The discriminant function was the retrospectively examined and externally tested. Results The results of conditional Logistic regression model indicated that mother's height, early pregnancy body mass index (BMI), gestational diabetes, gestational weeks, the height of uterine and abdominal circumference were associated with the birth of fetal macrosomia. The Bayesian discriminant function were established: Fetal macrosomia:y1=-27.802+8.420×Mother＇s height+8.719×early pregnancy BMI+10.485×gestational weeks+3.375×gestational diabetes+2.862×height of uterine and abdominal circumference; Non-macrosomia y2=-17.477+7.161×Mother＇s height+7.217×early pregnancy BMI+7.862×gestational weeks+2.036×gestational diabetes-0.085×height of uterine and abdominal circumference. Wilks′ Lambda λ=0.489, P<0.001, the Bayesian discriminant function was statistically significant. The internal and external conformity rates of the Bayesian discriminant model were all more than 80%. Conclutions The birth of fetal macrosomia is related to many factors. The Bayesian discriminant model in the present study is valuable to discriminate macrosomia and provide an objective reference for more accurate identification of macrosomia in the future.
