Online Program Home
  My Program

Abstract Details

Activity Number: 407 - Data Science Applications in Epidemiology
Type: Contributed
Date/Time: Tuesday, August 1, 2017 : 2:00 PM to 3:50 PM
Sponsor: Section on Statistics in Epidemiology
Abstract #324298 View Presentation
Title: A Comparison of Statistical Methods to Evaluate the Relationship Between Gestational Weight Gain and Gestational Age at Birth
Author(s): Lucia Petito* and Nicholas P. Jewell
Companies: University of California, Berkeley and University of California, Berkeley
Keywords: perinatal ; prediction ; Cox proportional hazards ; z-score ; SuperLearner
Abstract:

Studying the relationship between gestational weight gain (GWG) and gestational duration is difficult due to their inherent dependence. Serial GWG measurements provide ideal data, but are rarely available in population health datasets. To address this challenge, Hutcheon et al. (2013) developed GWG-for-gestational age z-scores. However, this approach was challenged by Mitchell et al. (2016), who claimed that patterns of weight gain through pregnancy are the gold standard for predicting gestational outcomes. They proposed a proportional hazards (Cox) model with GWG as a time-varying covariate to assess the relationship between GWG and gestational duration. Here, we test both methods on electronic medical record data (of 175,522 pregnant women in Sweden) that contains serial weight measurements. We use SuperLearner, a data-adaptive machine-learning algorithm, to create two models to assess the relationship between GWG and gestational duration, one using cumulative GWG as z-scores, and the other using serial weight measurements as a time-dependent covariate in a Cox model. We assess both models' goodness of fit, and the agreement of predicted risks of preterm birth across models.


Authors who are presenting talks have a * after their name.

Back to the full JSM 2017 program

 
 
Copyright © American Statistical Association