Online Program Home
  My Program

All Times EDT

Abstract Details

Activity Number: 53 - New Developments in Survival Analysis
Type: Contributed
Date/Time: Sunday, August 8, 2021 : 3:30 PM to 5:20 PM
Sponsor: Biometrics Section
Abstract #318599
Title: Discrete Frailty Modeling with Selection of Correlated Variables via Elastic Net: Application to Assessing Mixtures of Chemical Toxicants and Human Fecundity
Author(s): Abhisek Saha*
Companies: NICHD, NIH
Keywords: Discrete survival; Frailty; Elastic net; Time-to-pregnancy; Chemical exposure; Limit of detection
Abstract:

Evidence suggests that persistent environmental pollutants may be reproductive toxicants in both males and females. Previous association studies of chemicals have reported delayed time-to-pregnancy (TTP), but none have investigated exposure to chemical mixtures which are more representative of true human exposures. Assessing mixtures of chemicals effects on TTP poses significant statistical challenges, namely (i) TTP being a discrete survival outcome, subject to left truncation and right censoring, (ii) strongly correlated exposures, (iii) accounting for chemicals binding to lipids, (iv) non-linear effects and (v) high percentage below the limit of detection (LOD) of some chemicals. We propose a novel discrete frailty model (named Discnet) that allows selection of correlated exposures while addressing the said issues. Discnet is shown to have better and stable FN and FP rates compared to alternative methods in various simulation settings. In chemical exposure study, we find that older females, female exposure to cotinine (smoking), and DDT conferred a delay in getting pregnant, which was consistent across sensitivity analyses to account for LOD as well as non-linear associations.


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

Back to the full JSM 2021 program