Subgroup Identification of Early Preterm Birth (EPTB): Informing a Future Prospective Enrichment Clinical Trial Design

• Activity Number: 466 - Statistical Models for Complex Biomedial Data
• Type: Contributed
• Date/Time: Wednesday, August 2, 2017 : 8:30 AM to 10:20 AM
• Sponsor: Biometrics Section
• Author(s): Chuanwu Zhang*
• Companies: University of Kansas Medical Center
• Keywords: Early preterm birth ; Risk factor ; Interaction ; Classification and regression tree ; Logistic regression ; Enrichment trial design

Abstract

Background: For little knowing of the risk factors of early preterm birth (A baby born less than 34 weeks of gestation, ePTB) and a future clinical trial aim to identify if supplementing pregnant women with DHA will decrease the ePTB rate of the risk subgroup population, a study to identify risk factors and the risk subgroups was executed. Methods: The data were from 2014 CDC and NCHS. The sample was split into training and validation cohorts for model fitting and assessment. Logistic regression and CART model were used to examine the risk factors and their interactions. The risk factors are 14 maternal characteristic variables, including mothers'race, preterm birth history, etc. Results: Logistic regression with 10 risk factors produced the C-index of 0.646 based on the training cohort, and the C-index is 0.645 for the validation one. The CART model revealed the subgroup with a preterm birth history and Black race had the highest risk for ePTB. The c-index and misclassification rate were 0.579 and 0.034 for the training cohort, and 0.578 and 0.034 for the validation one, respectively. Conclusions: Both models identify the ePTB risk factors and risk subgroups for further trial.