Associations Between Hurricane Weather Exposure and Pregnancy Outcome: Demonstration of Ecological and Individual-Level Methods Public Deposited

Downloadable Content

Download PDF
Last Modified
  • March 19, 2019
  • Grabich, Shannon
    • Affiliation: Gillings School of Global Public Health, Department of Epidemiology
  • The impacts related to natural disasters are influenced by population growth, increasing coastal settlement, and global climate change. Socially vulnerable groups, including pregnant women, may be disproportionately affected by hurricane-related health effects. The limited literature on this topic is comprised primarily of pre-post ecological studies that may have biases due to confounding. The results of current studies, although mixed, have suggested that adverse fetal outcomes, including preterm delivery, may increase with hurricanes exposure. We assessed hurricane exposure in 2004 with pregnancy outcomes using a cohort of Florida pregnancies constructed from 2003-2005 using vital statistics records. We first utilized a difference-in-differences modeling technique to assess the county-level association between hurricane exposure and birth rates. We then conducted an individual-level analysis using time-to-event modeling to investigate hazard rates of single and multiple hurricane exposures on preterm delivery stratified by race/ethnicity. In county-level analysis the difference-in-differences method consistently produced estimates that suggested no association between hurricane and live birth rate, while the results from generalized linear model sensitivity analysis were inconsistent across exposure methods. The consistency of the difference-in-differences method suggests potential ability to control for time-invariant confounders. In individual-level analysis we found evidence of association between hurricane exposure and increased hazard of extremely preterm delivery (<32 weeks gestation) but no association with overall preterm delivery (<37 weeks gestation). Suggested associations appeared driven by white Hispanic and black Hispanic subgroups although we found limited evidence of statistical interaction. We did not find evidence of increasing hazards of preterm delivery with exposure to multiple hurricanes. This work provides examples for confounding control for future research to reduce biases. Although we found no association between hurricane exposure and live birth rate, we demonstrated a potential increase in hazard of extremely preterm delivery with hurricane exposure consistent with some of the current literature. As coastal populations and hurricane severity increases, future research is needed to understand the impacts of hurricane exposure on reproductive health, with a potential focus on the heterogeneity among race/ethnicity subgroups for targeting interventions, such as messages about evacuation or education about the importance of prenatal care during and after a disaster.
Date of publication
Resource type
Rights statement
  • In Copyright
  • Konrad, Charles
  • Robinson, Whitney
  • Engel, Stephanie
  • Horney, Jennifer
  • Richardson, David
  • Doctor of Philosophy
Degree granting institution
  • University of North Carolina at Chapel Hill Graduate School
Graduation year
  • 2015
Place of publication
  • Chapel Hill, NC
  • There are no restrictions to this item.

This work has no parents.