You are here:
Comparison of hurricane exposure methods and associations with county fetal death rates, adjusting for environmental quality
Citation:
Grabich, S., C. Gray, K. Rappazzo, L. Messer, J. Jagai, AND D. Lobdell. Comparison of hurricane exposure methods and associations with county fetal death rates, adjusting for environmental quality. Presented at International Society for Environmental Epidemiology, Seattle, WA, August 24 - 28, 2014.
Impact/Purpose:
This study utilized the Environmental Quality Index to control for environmental confounding in the primary analysis of examining the effects of hurricanes on fetal death rates.
Description:
Adverse effects of hurricanes are increasing as coastal populations grow and events become more severe. Hurricane exposure during pregnancy can influence fetal death rates through mechanisms related to healthcare, infrastructure disruption, nutrition, and injury. Estimation of hurricane effects is limited by broad exposure definitions. The two current exposure assignment methods are: county of Federal Emergency Management Agency (FEMA) disaster declaration, and kilometer (km) from storm track. Both lack specificity, with potential exposure misclassification of impacted areas. We propose a third method based on mapping of meteorological severity of wind speed. We examined wind speed by quartiles, Saffir-Simpson categories, and explored several binary cut points. We compared methods by examining associations with county-level fetal death rates among Florida women pregnant during the 2004 hurricane season (166,000 births and 800 fetal deaths). To control for county-level environmental factors, we used the Environmental Quality Index (EQI) developed by the U.S. Environmental Protection Agency, which accounts for five environmental domains (air, water, land, built and sociodemographic). Results are reported as rate differences (RD) (95% confidence intervals). Crude analyses using each exposure method for the most severe 2004 hurricane (Charley) were: FEMA declaration: -0.50 (-2.45, 1.45); 60 km of storm track: -0.05(-2.05, 1.96); and binary meteorological severity: 1.40 (-2.20, 5.01). Estimates adjusted for the overall EQI were: FEMA declaration: -0.12 (-2.04, 1.79); 60km of storm track: 0.69(-1.34, 2.71); and binary meteorological severity: 1.65 (-1.77, 5.13). We evaluated adjustment by the five specific domains of the EQI and found similar trends. While comparisons across methods were not statistically different, the proposed method showed consistent, positive associations between wind severity and fetal death. This abstract does not necessarily reflect EPA policy.