Leveraging earth observations for estimating health risks associated
with flooding precipitated by heavy rains
Abstract
Purpose:Flooding following heavy rains precipitated by
hurricanes/tropical storms has previously been shown to increase
fecal-oral diseases, vector-borne disease transmission and pregnancy
complications during or following inundation. Remote sensing can be used
to spatio-temporally resolve inundation extents for subsequent analysis
of risks associated with flooding at a finer scale. Here we combined
earth observations of the flooding caused by Hurricane Harvey in 2017
with Emergency Department (ED) visit data to evaluate health outcomes
associated with flooding.Methods:Our study area included 1073 flooded
and 1809 non-flooded census tracts in Texas which were categorized using
the inundation maps from Dartmouth Flood Observatory. These maps were
created using Sentinel and MODIS satellite imagery captured between 28th
Aug - 4th Sep 2017 following the landfall of the hurricane. ED visits in
the study area were obtained from Texas Department of State Health
Services. A controlled interrupted time series design was employed using
ED visits from non-flooded tracts as the control series and ED visits
before a week of the landfall and through 2018 as control period.
Poisson regression using generalized estimating equation with census
tracts as the group variable was used to estimate the relative risk of
the health outcomes associated with flooding during and following the
flooded days, adjusting for the age, ethnicity, race, sex of the
patient, day of week, month and year trends. Results:Flooding was
associated with a 35% (95% CI: 22%-48%) increase in risk for insect
stings and 24% (17%-31%) increase in risk of pregnancy complications
during the flood period. Similarly, relative risks were also elevated
(>1) for drowning, hypothermia, and intestinal infectious
diseases in the flooded tracts. Also, in the months following the flood
period, the relative risk was still elevated (>1) for
pregnancy complications and insect stings while asthma and acute
respiratory infections showed decreased risks. Conclusion: Earth
observations have helped in understanding the health risks that are
related to flooding. These earth observations can in turn be used to
identify specific communities with increased health risks during and
following flooding events.