Dataset Information

Electronic Health Record Driven Prediction for Gestational Diabetes Mellitus in Early Pregnancy.

ABSTRACT: Gestational diabetes mellitus (GDM) is conventionally confirmed with oral glucose tolerance test (OGTT) in 24 to 28 weeks of gestation, but it is still uncertain whether it can be predicted with secondary use of electronic health records (EHRs) in early pregnancy. To this purpose, the cost-sensitive hybrid model (CSHM) and five conventional machine learning methods are used to construct the predictive models, capturing the future risks of GDM in the temporally aggregated EHRs. The experimental data sources from a nested case-control study cohort, containing 33,935 gestational women in West China Second Hospital. After data cleaning, 4,378 cases and 50 attributes are stored and collected for the data set. Through selecting the most feasible method, the cost parameter of CSHM is adapted to deal with imbalance of the dataset. In the experiment, 3940 samples are used for training and the rest 438 samples for testing. Although the accuracy of positive samples is barely acceptable (62.16%), the results suggest that the vast majority (98.4%) of those predicted positive instances are real positives. To our knowledge, this is the first study to apply machine learning models with EHRs to predict GDM, which will facilitate personalized medicine in maternal health management in the future.

SUBMITTER: Qiu H

PROVIDER: S-EPMC5703904 | biostudies-other | 2017 Nov

REPOSITORIES: biostudies-other

ACCESS DATA

Publications

Electronic Health Record Driven Prediction for Gestational Diabetes Mellitus in Early Pregnancy.

Qiu Hang H Yu Hai-Yan HY Wang Li-Ya LY Yao Qiang Q Wu Si-Nan SN Yin Can C Fu Bo B Zhu Xiao-Juan XJ Zhang Yan-Long YL Xing Yong Y Deng Jun J Yang Hao H Lei Shun-Dong SD

Scientific reports 20171127 1

Gestational diabetes mellitus (GDM) is conventionally confirmed with oral glucose tolerance test (OGTT) in 24 to 28 weeks of gestation, but it is still uncertain whether it can be predicted with secondary use of electronic health records (EHRs) in early pregnancy. To this purpose, the cost-sensitive hybrid model (CSHM) and five conventional machine learning methods are used to construct the predictive models, capturing the future risks of GDM in the temporally aggregated EHRs. The experimental d ...[more]

PMID: 29180800

Similar Datasets

Project description:BackgroundEpidemiological findings are inconsistent regarding the associations between air pollution exposure during pregnancy and gestational diabetes mellitus (GDM). Several limitations exist in previous studies, including potential outcome and exposure misclassification, unassessed confounding, and lack of simultaneous consideration of air pollution mixtures and particulate matter (PM) constituents.ObjectivesTo assess the association between GDM and maternal residential exposure to air pollution, and the joint effect of the mixture of air pollutants and PM constituents.MethodsDetailed clinical data were obtained for 395,927 pregnancies in southern California (2008-2018) from Kaiser Permanente Southern California (KPSC) electronic health records. GDM diagnosis was based on KPSC laboratory tests. Monthly average concentrations of fine particulate matter < 2.5 μm (PM2.5), <10 μm (PM10), nitrogen dioxide (NO2), and ozone (O3) were estimated using kriging interpolation of Environmental Protection Agency's routine monitoring station data, while PM2.5 constituents (i.e., sulfate, nitrate, ammonium, organic matter and black carbon) were estimated using a fine-resolution geoscience-derived model. A multilevel logistic regression was used to fit single-pollutant models; quantile g-computation approach was applied to estimate the joint effect of air pollution and PM component mixtures. Main analyses adjusted for maternal age, race/ethnicity, education, median family household income, pre-pregnancy BMI, smoking during pregnancy, insurance type, season of conception and year of delivery.ResultsThe incidence of GDM was 10.9% in the study population. In single-pollutant models, we observed an increased odds for GDM associated with exposures to PM2.5, PM10, NO2 and PM2.5 constituents. The association was strongest for NO2 [adjusted odds ratio (OR) per interquartile range: 1.176, 95% confidence interval (CI): 1.147-1.205)]. In multi-pollutant models, increased ORs for GDM in association with one quartile increase in air pollution mixtures were found for both kriging-based regional air pollutants (NO2, PM2.5, and PM10, OR = 1.095, 95% CI: 1.082-1.108) and PM2.5 constituents (i.e., sulfate, nitrate, ammonium, organic matter and black carbon, OR = 1.258, 95% CI: 1.206-1.314); NO2 (78%) and black carbon (48%) contributed the most to the overall mixture effects among all krigged air pollutants and all PM2.5 constituents, respectively. The risk of GDM associated with air pollution exposure were significantly higher among Hispanic mothers, and overweight/obese mothers.ConclusionThis study found that exposure to a mixture of ambient PM2.5, PM10, NO2, and PM2.5 chemical constituents was associated with an increased risk of GDM. NO2 and black carbon PM2.5 contributed most to GDM risk.

Dataset Information

Electronic Health Record Driven Prediction for Gestational Diabetes Mellitus in Early Pregnancy.

Publications

Electronic Health Record Driven Prediction for Gestational Diabetes Mellitus in Early Pregnancy.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets