Health Care Utilization Among Texas Veterans Health Administration Enrollees Before and After Hurricane Harvey, 2016-2018

Gosia S Clore,Eric Tate,Michihiko Goto,Margaret Carrel,Seungwon Kim,Eli N Perencevich,Mary Vaughan Sarrazin

doi:10.1001/jamanetworkopen.2021.38535

Abstract

Hurricanes and flooding can interrupt health care utilization. Understanding the magnitude and duration of interruptions, as well as how they vary according to hazard exposure, race, and income, are important for identifying populations in need of greater retention in care. To determine how the differential exposure to Hurricane Harvey in August 2017 is associated with changes in utilization of Veterans Health Administration health care. This is a retrospective cohort analysis of primary care practitioner (PCP) visits, emergency department visits, and inpatient admissions in the Veterans Health Administration among Texas veterans residing in counties impacted by Hurricane Harvey from 2016 to 2018. Data analysis was performed from September 2020 to May 2021. Residential flooding after Hurricane Harvey. Interrupted time series analysis measured changes in health care utilization over time, stratified by residential flood exposure, race, and income. Of the 99 858 patients in the cohort, 89 931 (90.06%) were male, and their median (range) age was 58 (21 to 102) years. Compared with veterans in nonflooded areas, veterans living in flooded areas were more likely to be Black (24 715 veterans [33.80%] vs 4237 veterans [15.85%]) and low-income (14 895 veterans [20.37%] vs 4853 veterans [18.15%]). Rates of PCP visits decreased by 49.78% (95% CI, -64.52% to -35.15%) for veterans in flooded areas and by 45.89% (95% CI, -61.93% to -29.91%) for veterans in nonflooded areas and did not rebound until more than 8 weeks after the hurricane. Rates of PCP visits in flooded areas remained lower than expected for 11 weeks among White veterans (-6.99%; 95% CI, -14.36% to 0.81%) and for 13 weeks among racial minority veterans (-7.22%; 95% CI, -14.11% to 0.30%). Low-income veterans, regardless of flood status, experienced greater suppression of PCP visits in the 8 weeks following the hurricane (-13.72%; 95% CI, -20.51% to -6.68%) compared with their wealthier counterparts (-9.63%; 95% CI, -16.74% to -2.26%). These findings suggest that flood disasters such as Hurricane Harvey may be associated with declines in health care utilization that differ according to flood status, race, and income strata. Patients most exposed to the disaster also had the greatest delay or nonreceipt of care.

Highlights

In late August 2017, a 4-day rain event triggered by Hurricane Harvey deposited more than 270 trillion gallons of water across southeastern Texas, with some areas receiving more than 50 inches of precipitation
Rates of primary care practitioner (PCP) visits decreased by 49.78% for veterans in flooded areas and by 45.89% for veterans in nonflooded areas and did not rebound until more than 8 weeks after the hurricane
Rates of PCP visits in flooded areas remained lower than expected for 11 weeks among White veterans (−6.99%; 95% CI, −14.36% to 0.81%) and for 13 weeks among racial minority veterans (−7.22%; 95% CI, −14.11% to 0.30%)

Summary

Introduction

In late August 2017, a 4-day rain event triggered by Hurricane Harvey deposited more than 270 trillion gallons of water across southeastern Texas, with some areas receiving more than 50 inches of precipitation. The presence, magnitude, and duration of negative outcomes following disasters are often highly conditioned on race and ethnicity and socioeconomic status. Differential social vulnerabilities before flooding or other disaster exposures frequently exacerbate and accelerate differential trajectories following disasters.[15,20,21] Differences in exposure to flooding after hurricanes break down along race and ethnicity and socioeconomic status: low-income and racial and ethnic minority populations often experience more flooding than do White and wealthy populations, as has been observed following Hurricanes Katrina, Sandy, and Harvey, because of historical and current structural inequities.[22,23,24,25,26,27]

Methods

Results

Discussion

Conclusion