Estimating Health Risks Associated With Flooding Following Hurricane Harvey Using Earth Observations and the CDC Social Vulnerability Index

Increases in cases of diarrheal disease, respiratory infections and pregnancy complications have been reported in the literature following floods caused by heavy rainfall. Analyzing the association between health records of outcomes related to flooding demarcated by satellite observations will be helpful to evaluate the use of satellite observed products in the mitigation of health risks for future flood events. The Centers for Disease Control and Prevention Social Vulnerability Index (CDC SVI), a relative index assigned to census tracts, measures sociodemographic factors that may affect the ability of communities to prepare for, respond to, and recover from extreme weather events. This index, which quantifies social vulnerability is expected to have a positive relationship with health outcomes associated with flooding. This study uses an inundation map created using observations from active remote sensing satellites to classify census tracts that were flooded after the historic rainfall caused by Hurricane Harvey in Texas in 2017. The duration or period of the inundation was determined using United States Geological Survey (USGS) stream gauge data. A controlled before and after study design was used, and the relative risk (RR) of 11 cause-specific emergency department (ED) visits among the flooded census tracts compared to non-flooded tracts during and after the flood period was modelled using modified Poisson regression while adjusting for a baseline period and the age, ethnicity, race and sex of the patient. Further modification of this relationship by social vulnerability, as measured by CDC SVI quartiles, was examined. The results of this study show that flooding was associated with an increase in ED visits related to carbon monoxide poisoning, insect bites, dehydration, hypothermia, intestinal infectious diseases, and pregnancy complications during the flood period. The average rate of ED visits related to pregnancy complications and insect bites were greater among the flooded tracts compared to the non-flooded tracts in the month following the inundation. Modification of this association by CDC SVI was observed in some cases, such that ED visits were higher or lower in census tracts within higher vulnerable quartiles compared to the least vulnerable quartile. Evaluating the usefulness of earth observations and the CDC SVI in estimating the health risk associated with floods due to Hurricane Harvey has provided understanding the use of these products for future flooding events in identifying specific communities with increased health risks during and following flooding events.