Link to the University of Pittsburgh Homepage
Link to the University Library System Homepage Link to the Contact Us Form

Incident prepregnancy obesity and perinatal mortality: application of nonparametric doubly robust methods

Yu, Ya-Hui (2018) Incident prepregnancy obesity and perinatal mortality: application of nonparametric doubly robust methods. Doctoral Dissertation, University of Pittsburgh. (Unpublished)

Submitted Version

Download (3MB) | Preview


This dissertation’s objective is to address methodological challenges in estimating the association of prepregnancy obesity with stillbirth and infant mortality. Our goal was to advance understanding of this topic of high public health importance to decrease the gap of high perinatal mortality between United States and other counties. Our focus was on the association of incident prepregnancy obesity with stillbirth and infant mortality. We constructed our analytic cohort of multiparous women who were non-obese in their first pregnancy from a population-based cohort study in Pennsylvania from 2003-2013 (n=1,551,919 singleton pregnancies).
A visualization tool was developed for variable selection considering bias-variance tradeoff for inverse probability weights. Applying this tool to our study, we identified two high influential confounders which informed us to adjust for them carefully.
Next, we examined the impact of parametric modeling assumptions on these associations by analytic methods with different reliance on parametric assumptions. Consistent, increased risk of stillbirth among women who became obese compared to those staying non-obese was observed with all methods. However, discrepancies in magnitude between methods were found: risk differences from semiparametric inverse probability weighting and nonparametric targeted minimum loss-based estimation (TMLE) were larger than those from parametric methods.
Then, the analysis was constrained to those at normal first pregnancy weight, and we examined the effect of becoming overweight and obese on stillbirth and infant mortality in the second pregnancy. Nonparametric TMLE estimates showed that becoming overweight increased the risk of stillbirth (RD=1.4, 95% CI: 0.6, 2.2) while those becoming obese had increased risks of both stillbirth (RD=4.0, 95% CI: 1.4, 6.6) and neonatal mortality (RD=2.3, 95% CI: 0.1, 4.5) in the second pregnancy. A dose-response relationship was observed where increasing interpregnancy BMI change was associated with increased risk of stillbirth and infant mortality.
Our findings provided the evidence that transitioning from normal weight to overweight or obese increases risk of stillbirth and infant mortality. Health care providers can communicate the importance of weight maintenance. These are important public health opportunities to prevent the onset of obesity and minimize fetal and infant risks, particularly since pregnancy is an important period to optimize maternal health.


Social Networking:
Share |


Item Type: University of Pittsburgh ETD
Status: Unpublished
CreatorsEmailPitt UsernameORCID
Yu, Ya-Huiyay55@pitt.eduyay55
ETD Committee:
TitleMemberEmail AddressPitt UsernameORCID
Committee ChairNaimi,
Committee MemberBodnar,
Committee MemberBrooks,
Committee MemberHimes,
Date: 18 September 2018
Date Type: Publication
Defense Date: 8 June 2018
Approval Date: 18 September 2018
Submission Date: 4 June 2018
Access Restriction: 5 year -- Restrict access to University of Pittsburgh for a period of 5 years.
Number of Pages: 129
Institution: University of Pittsburgh
Schools and Programs: School of Public Health > Epidemiology
Degree: PhD - Doctor of Philosophy
Thesis Type: Doctoral Dissertation
Refereed: Yes
Uncontrolled Keywords: Maternal obesity, stillbirth, infant mortality, TMLE
Date Deposited: 18 Sep 2018 15:10
Last Modified: 01 Jul 2023 05:15


Monthly Views for the past 3 years

Plum Analytics

Actions (login required)

View Item View Item