Abstract
Objective:
Emergency departments (EDs) are critical sources of care after natural disasters such as hurricanes. Understanding the impact on ED utilization by subpopulation and proximity to the hurricane’s path can inform emergency preparedness planning. This study examines changes in ED utilization for residents in 344 counties after seven U.S. hurricanes between 2005 and 2016.
Methods:
This retrospective observational study used ED data from the Healthcare Cost and Utilization Project State Inpatient Databases and State Emergency Department Databases. ED utilization rates for weeks during and after hurricanes were compared with prehurricane rates, stratified by the proximity of the patient county to the hurricane path, age, and disease category.
Results:
The overall population rate of weekly ED visits changed little posthurricane, but rates by disease categories and age demonstrated varying results. Utilization rates for respiratory disorders exhibited the largest posthurricane increase, particularly 2–3 weeks following the hurricane. The change in population rates by disease categories and age tended to be larger for people residing in counties closer to the hurricane path.
Conclusions:
Changes in ED utilization following hurricanes depend on disease categories, age, and proximity to hurricane path. Emergency managers could incorporate these factors into planning processes.
Introduction
Hospital emergency departments (EDs) are critical sources of health care during and immediately following natural disasters such as hurricanes. In fact, the Emergency Medical Treatment & Labor Act requires EDs to examine and stabilize patients, regardless of their ability to pay, in the time of an emergency.1 Examination of historical ED utilization patterns can contribute important information about community health needs following disasters and help state and local organizations plan future emergency responses.
Previous studies have examined changes in hospital utilization during hurricanes. These analyses were focused on a single storm, a single state, one age group, or specific health conditions, yet all showed an increase in ED utilization in the hurricane evacuation zone.2–8 Two studies found a significant increase in ED utilization for older adults after Hurricane Sandy,2,7 and others noted an increased need for services that are not typically performed in the ED such as dialysis2,5,6,7 and prescription refills.2,3,7 The unique contribution of this study is to examine ED utilization rates for the weeks during and after seven U.S. hurricanes between 2005 and 2016, across five patient age groups, across a range of disease categories, and by proximity to the hurricane’s path.
Methods
Data Sources
This was a retrospective observational study of ED visits in nine states where the Federal Emergency Management Agency (FEMA) declared at least one county a disaster area9 after any of the following seven hurricanes: Dennis, Wilma, Gustav, Irene, Isaac, Sandy, and Matthew. FEMA-designated disaster counties were classified as in, near, or remote to the hurricane path based on National Oceanic and Atmospheric Administration data for hurricane trajectories.10–11 Data on ED utilization were from the Healthcare Cost and Utilization Project (HCUP) State Emergency Department Databases, which capture records for ED visits that do not result in admission to the same hospital, and HCUP State Inpatient Databases, which capture inpatient records that began in the ED (ie, ED visits resulting in admission to the same hospital).12 For each hurricane, encounter-level data for ED visits for the 4 weeks preceding the start date of the hurricane and the following 4 weeks were summarized by age and condition into weekly volume. The HCUP databases include one record per hospital encounter with clinical information such as diagnoses and procedures, in addition to patient demographics.
Analytic Sample and Hurricanes
Two of the storms represented in this analytic sample made landfall in Florida. Hurricane Dennis made landfall in the Florida panhandle in July 2005 as a Category 3 storm. Significant wind gusts occurred in several counties, and storm surge rose above normal tide levels, causing flooding.13 The analysis included information from 27 EDs in FEMA-designated disaster counties in Florida with over 600,000 ED visits a year. Later in 2005, Hurricane Wilma made landfall in southwestern Florida in October as a Category 3 storm. Affected counties experienced high wind, heavy rainfall, storm surge, and tornadoes.13 The analysis included information from 103 EDs in FEMA-designated disaster counties in Florida with over 3.4 million ED visits a year.
Two of the storms represented in this analysis first made landfall in Louisiana but had extensive impacts on Florida. Hurricane Gustav made landfall in southeastern coastal counties of Louisiana in September 2008 as a Category 2 storm. Affected counties were impacted by heavy rainfall, storm surge, freshwater flooding, and tornadoes.13 The analysis included information from 14 EDs in FEMA-designated disaster counties in Florida with over 500,000 ED visits a year. Hurricane Isaac made landfall on the coast of Louisiana in August 2012 as Category 1 storm but days prior had impacted the Florida Keys. Flooding was caused by storm surge and heavy rains.13 The analysis included information from 37 EDs in FEMA-designated disaster counties in Florida with almost 1.3 million ED visits a year.
The analysis also includes three hurricanes that made landfall in other states on the eastern seaboard. Hurricane Irene made landfall in coastal counties of North Carolina in August 2011 as a Category 1 storm. Coastal counties experienced heavy rainfall and tidal surge.13 The analysis included information from 313 EDs in FEMA-designated disaster counties in Maine, Maryland, Massachusetts, New Jersey, New York, North Carolina, Rhode Island, and Vermont with over 14.7 million ED visits a year. In October 2012, Hurricane Sandy made landfall in New Jersey as an extratropical storm and caused significant tidal surge and coastal flooding in coastal counties.13 The analysis included information from 234 EDs in FEMA-designated disaster counties in Maryland, Massachusetts, New Jersey, New York, and Rhode Island with over 13.4 million ED visits a year. Hurricane Matthew made landfall in coastal counties of South Carolina in October 2016 as Category 2 storm. Affected counties experienced hurricane-force winds, heavy rainfall, and storm surge flooding.13 The analysis included information from 176 EDs in FEMA-designated disaster counties in Florida, Georgia, North Carolina, and South Carolina with almost 8.6 million ED visits a year.
Statistical Methods
Weekly ED visit volumes were based on the patient’s county of residence. Population-weighted rates of weekly ED visits were calculated for the 4 weeks preceding the hurricane (averaged to define a rate for the prehurricane period), the hurricane week, and the 3 posthurricane weeks. Rates are reported as ED visits per 10,000 population.
Weekly rates were stratified by age (0–17, 18–44, 45–64, 65–79, and 80 years or older) and disease category. The first-listed diagnosis was grouped into categories using International Classification of Diseases (ICD) coding. We focused on six disease categories likely to be seen after a hurricane (injury, disorders of the respiratory, digestive, circulatory, genitourinary, and nervous systems) using ICD chapters to define most categories (see Table 1). The data for Hurricane Matthew included ICD, Tenth Revision, Clinical Modification (ICD-10-CM) coding, whereas the earlier hurricanes had data coded in the Ninth Revision (ICD-9-CM).
Table 1.
Coding Ranges for Disease Categories
| Disease Category | ICD-9-CM Diagnosis Code Ranges | ICD-10-CM Diagnosis Code Ranges |
|---|---|---|
| Circulatory | 390–459 | I00–I99 |
| Digestive | 520–579 | K00–K95 |
| Genitourinary | 580–629 | N00–N99 |
| Injury | 800–909.2, 909.4, 909.9, 910–994.9, 995.5–995.59, and 995.80–995.85 | All S codes, T07–T34, T51–T76, T79, and T36-T50 with a 6th character of 1, 2, 3, or 4 with the following exceptions: T36.9, T37.9, T39.9, T41.4, T42.7, T43.9, T45.9, T47.9, and T49.9 with a 5 th character of 1, 2, 3, or 4 |
| Nervous system and sense organs | 320–389 | G00–G99 and H00–H95 |
| Respiratory | 460–519 | J00–J99 |
Abbreviations: ICD-9-CM, International Classification of Diseases, Ninth Revision, Clinical Modification; ICD-10-CM, International Classification of Diseases, Tenth Revision, Clinical Modification/Procedure Coding System.
Statistical significance of differences in rates was tested using two-sample t-tests for means and a threshold of P < 0.05.
Limitations
The analysis of the changes in ED utilization after a hurricane was limited to states included in the HCUP ED databases for the applicable data year. Not all states participate in HCUP. In addition, for states included in the analysis, no information was publicly available about EDs that were closed or damaged during the hurricane.
Analysis was complicated by the coding change from ICD-9-CM to ICD-10-CM diagnoses for one hurricane. Within each hurricane, the definition of disease categories was consistent, but the specific diagnoses within a disease category may vary slightly across the two coding systems. We ran a sensitivity analysis comparing posthurricane ED utilization for the six hurricanes with ICD-9-CM data with our results for all hurricanes (one with ICD-10-CM data). Overall and for respiratory and circulatory conditions, the magnitude of change of the posthurricane ED utilization was similar. Differences were noted in the magnitude of change in ED utilization for injury in remote counties. This demonstrates that the use of ICD-10-CM coding for the one later hurricane did not disproportionately change our results.
Results
The HCUP data for the seven hurricanes included ED visits in 904 EDs in 344 counties involving 88 million individuals (Table 2). For in-path counties, the mean maximum sustained wind speed across the seven hurricanes was 90 mph and the mean minimum central pressure was 953 millibars.
Table 2.
Characteristics of Counties by Proximity to Hurricane Path
| County Characteristic | County in Hurricane Path | County Near Hurricane Path | County Remote from Hurricane Path |
|---|---|---|---|
| Total number of FEMA-designated disaster counties for the 7 hurricanes | 48 | 316 | 421 |
| Number (%) of counties in study | 21 (44) | 184 (58) | 139 (33) |
| Total population of in FEMA-designated disaster counties for the 7 hurricanes | 15,571,789 | 73,671,930 | 38,737,408 |
| Total (%) population in counties in study | 13,354,505 (86) | 53,974,967 (73) | 21,307,790 (55) |
| Total number of EDs in FEMA-designated disaster counties for the 7 hurricanes | 183 | 776 | 769 |
| Number (%) of EDs in study | 114 (62) | 488 (63) | 302 (39) |
| Rate per 10,000 population of ED visits in 4 weeks prior to hurricane | 80.0 | 79.5 | 88.3 |
| Counties labeled storm ready, % | 43 | 52 | 65 |
| Mean maximum sustained wind speed, mph | 90 | -- | -- |
| Mean minimum central pressure, millibars | 953 | -- | -- |
| County had flood or flash flood, % | 67 | 68 | 52 |
| County had storm surge tide, % | 52 | 13 | 8 |
Abbreviations: ED, emergency department; FEMA, Federal Emergency Management Agency.
Source: Agency for Healthcare Research and Quality, Healthcare Cost and Utilization Project, State Emergency Department Databases and State Inpatient Databases, 11 states, 2005–2016, and National Oceanic and Atmospheric Administration Best Track data for Hurricanes Dennis (Florida; 2005), Wilma (Florida; 2005), Gustav (Florida; 2008), Irene (Maine, Maryland, Massachusetts, New Jersey, New York, North Carolina, Rhode Island, and Vermont; 2011), Isaac (Florida; 2012), Sandy (Maryland, Massachusetts, New Jersey, New York, and Rhode Island; 2012), and Matthew (Florida, Georgia, North Carolina, and South Carolina; 2016).
Overall ED Utilization
The overall rate of weekly ED visits for counties affected by the hurricanes (80.0 for in-path counties, 79.5 for near-path counties, and 88.3 for remote counties) either did not change significantly or decreased less than 4% from the prehurricane period to the hurricane week or any of the 3 posthurricane weeks (Figure 1, Supplemental Table 1).
Figure 1.
Percent Change in Weekly ED Visits per 10,000 from Prehurricane to the Hurricane Week and 3 Weeks Posthurricane, by Hurricane Proximity and Patient Age
Patient Age
Compared with the prehurricane rate, children in in-path and remote counties demonstrated a delayed increase in utilization in posthurricane week 3 (11.3% and 9.6%, Figure 1, Supplemental Table 1). In contrast, adults aged 65 years and older in in-path counties had a significant increase in utilization during the hurricane week, with the magnitude of change decreasing as the distance from the hurricane increased from in-path to remote counties. For adults aged 65–79 years, the rate of ED visits during the hurricane week increased by 17.7% for in-path counties, 9.4% for near-path counties, and 4.1% for remote counties. Larger increases were seen for adults aged 80 years and older (22.1% for in-path counties, 14.4% for near-path counties, and 5.2% for remote counties). Posthurricane ED visit rates for adults aged 80 years and older from in-path counties remained significantly higher than prehurricane rates for the 3 weeks studied, but rates for adults aged 65–79 years from in-path counties returned to baseline by the second posthurricane week.
Disease Categories
Figures 2 through 4 (Supplemental Tables 2 through 6) depict changes in ED utilization rates for the six disease categories by patient age. For children in in-path counties, only the rate of weekly visits involving respiratory disorders significantly increased during the hurricane week, and the magnitude of the change decreased with the distance from the hurricane (22.4% for in-path counties, 9.3% for near-path counties, and 5.2% for remote counties) (Figure 2, Supplemental Table 2). The largest increases in visits involving respiratory disorders occurred between the prehurricane period and posthurricane week 3 (56.4% for in-path counties, 24.2% for near-path counties, and 29.0% for remote counties). The most common respiratory conditions for children during the hurricane week were bronchitis, acute tonsillitis, acute upper respiratory infections, and asthma. For pediatric ED visits involving injuries, there was about a 10% increase in weekly ED visit rates from the prehurricane period to posthurricane weeks 2 and 3 for in-path and remote counties. Near-path counties had decreased rates for injuries during most posthurricane weeks.
Figure 2.
Percent Change in Weekly ED Visits per 10,000 from Prehurricane to the Hurricane Week and 3 Weeks Posthurricane, by Hurricane Proximity, Patients Aged 17 Years and Younger
Figure 4.
Percent Change in Weekly ED Visits per 10,000 from Prehurricane to the Hurricane Week and 3 Weeks Posthurricane, by Hurricane Proximity, Patients Aged 65–79 Years and Aged 80 Years and Older
ED visits involving respiratory disorders increased in varying patterns for adults aged 18–44 years and 45–64 years (Figure 3, Supplemental Tables 3 and 4). ED visits involving respiratory disorders for adults aged 45–64 years significantly increased during the hurricane week. The magnitude of the change decreased with the distance from the hurricane from 17.5% for in-path counties, to 13.6% for near-path counties, and to 6.2% for remote counties. In addition, increased rates were observed for this age group for posthurricane weeks 1 and 3. Younger adults aged 18–44 only showed an increase in ED visits for respiratory disorders in posthurricane weeks 2 and 3 for in-path counties (10.0% and 22.4%, respectively) and posthurricane week 3 in near-path and remote counties (6.4% and 14.6%, respectively). The largest increases in ED visits involving respiratory disorders occurred from in-path counties in posthurricane week 3 (22.4% for adults aged 18–44%, 19.7% for adults aged 45–64 years). During the hurricane week, the most common respiratory conditions for adults aged 18–64 years were acute pharyngitis, asthma, bronchitis, acute upper respiratory infections, and pneumonia.
Figure 3.
Percent Change in Weekly ED Visits per 10,000 from Prehurricane to the Hurricane Week and 3 Weeks Posthurricane, by Hurricane Proximity, Patients Aged 18 to 44 Years and Aged 45 to 64 Years
ED visits involving injuries for adults aged 45–64 years significantly increased during the hurricane week (Figure 3, Supplemental Table 4). The magnitude of the change decreased with the distance from the hurricane from 32.0% for in-path counties, to 13.4% for near-path counties, and to 5.6% for remote counties. Increased rates continued for in-path and remote counties until posthurricane week 1 and then either returned or declined from the prehurricane average rates of ED visits.
In-path counties showed either no change or a decrease in weekly ED visits for adults aged 18–44 and 45–64 years for the other four diagnoses, circulatory disorders, digestive disorders, genitourinary disorders, and nervous system disorders (Figure 3, Supplemental Tables 3 and 4). ED visits for adults aged 18–44 years in near-path counties showed a significant 4 to 12% decrease in visits during the hurricane week and posthurricane week 1 for all four diagnoses. ED visits for adults 45–65 years showed significant decrease in the four diagnoses for varying posthurricane weeks, with no notable pattern. Remote counties also showed a significant decrease of 3 to 8% in ED visits for adults aged 18–44 years for the four diagnoses during some of the hurricane or posthurricane weeks. Only ED visits for adults 45–65 years for genitourinary disorders showed a significant decrease in remote counties in the third posthurricane week.
For adults aged 65–79 and aged 80 years and older, ED visits involving respiratory disorders significantly increased during the hurricane week; the magnitude decreased with distance from the hurricane (Figure 4, Supplemental Tables 5 and 6). For adults aged 65–79 years, ED visit rates involving respiratory disorders increased during the hurricane week by 34.1% in in-path counties, 33.1% in near-path counties, and 9.5% in remote counties. For adults aged 80 years and older, ED visit rates involving respiratory disorders increased during the hurricane week by 38.1% in in-path counties, 32.9% in near-path counties, and 9.9% in remote counties. The most common respiratory disorders for adults aged 65 years and older during the hurricane week were bronchitis, pneumonia, chronic obstructive pulmonary disease, and respiratory failure.
For adults aged 65–79 years, ED visit rates involving injury increased during the hurricane week by 44.1% in in-path counties, 30.0% in near-path counties, and 15.1% in remote counties. A similar pattern of increased rates for ED visits involving injury during the hurricane week with decreased magnitude with distance from the hurricane was observed for adults aged 80 years (52.9% for in-path counties, 29.7% in near-path counties, and 8.7% in remote counties). For both older age groups (65–79 and 80+ years), the ED visit rates related to injuries returned to the prehurricane rate by the third posthurricane week.
For adults aged 80 years and older, ED visit rates involving genitourinary disorders increased during the hurricane week by 17.1% in in-path counties and 11.5% in near-path counties. The most common conditions during the hurricane week were urinary tract infections, acute renal failure, and calculus of the ureter or kidney.
ED visit rates related to disorders of the nervous system increased during the hurricane week in both in-path and near-path counties for adults aged 65–79 years (21.4% and 7.2%, respectively), but only in near-path counties for adults aged 80 years and older (16.7%). The most common conditions for adults aged 65 years and older reported during the hurricane week were Alzheimer’s disease, transient ischemic attack, epilepsy, vertigo, and chronic pain. For both older age groups (65–79 and 80+ years), the ED visit rates for nervous system disorders returned to the prehurricane rate by the second posthurricane week.
Discussion
In aggregate, changes in ED visits appear relatively small during the hurricane and 3 posthurricane weeks. However, these small changes mask striking changes in ED utilization by specific subpopulations. Individuals younger than 18 years and older than 44 years had an increased need for ED respiratory services the week of the hurricane. The magnitude of the increased need declined with the distance from the patient’s county of residence to the hurricane’s path. Adults older than 44 years also had significantly increased need for ED services related to injury the week of the hurricane with the magnitude of the change decreasing with the distance from the hurricane. Only adults aged 65 years and older had a significantly increased need for ED services for disorders of the genitourinary and nervous systems the weeks during or following the hurricane.
Some changes in ED utilization were not in an expected direction. For example, rather than diminishing after the hurricane passed, ED visits involving respiratory disorders were at their highest rates during posthurricane week 3 for patients younger than 65 years. This delay may reflect a latency period after exposure to fungi, lack of air conditioning because of power outages, and other factors.14
Decreases in ED visits for children and adults in near-path counties were also unexpected. Further research is needed, but this finding may suggest that ED volume following a hurricane is affected by migration out of in-path areas and, in some cases, temporary hospital closures.
Conclusions
This analysis shows that ED utilization data can reveal striking changes in the use of specific categories of ED services by subgroups of the population in the aftermath of hurricanes. Findings on the increased use of ED visits related to respiratory conditions by people of all ages and for services related to injury and disorders of the genitourinary and nervous systems by older people may be useful in short- and longer-term service plans of facilities located in areas that are frequently affected by these types of events. The availability of ED utilization data makes further work in this area highly feasible. Additional analyses should examine the association between hurricane characteristics and the epidemiology of conditions. Knowing the short- and long-term impact on ED utilization for specific types of storms by clinical specialty would be valuable to emergency planners and local health care providers.
Supplementary Material
Acknowledgments:
The authors gratefully acknowledge Thomas Flottemesch, PhD (while with IBM Watson Health) and Minya Sheng, MS (IBM Watson Health) for assistance with data management and programming; Frank Yoon (IBM Watson Health) and Hal Skinner (while with IBM Watson Health) for statistical guidance; Emma Mollenhauer (IBM Watson Health) for assistance with preparation of figures; and Linda Lee, PhD (while with IBM Watson Health) for providing editorial review of the manuscript. We also wish to acknowledge the eleven HCUP Partner organizations that contributed to the HCUP State Databases used in this study: Florida Agency for Health Care Administration, Georgia Hospital Association, Maine Health Data Organization, Maryland Health Services Cost Review Commission, Massachusetts Center for Health Information and Analysis, New Jersey Department of Health, New York State Department of Health, North Carolina Department of Health and Human Services, South Carolina Revenue and Fiscal Affairs Office, Rhode Island Department of Health, and the Vermont Association of Hospitals and Health Systems.
Funding: This research was supported by the Agency for Healthcare Research and Quality (AHRQ), Healthcare Cost and Utilization Project (HCUP) contract No. HHSA-290-2018-00001-C.
Abbreviations:
- FEMA
Federal Emergency Management Agency
- ED
emergency department
- HCUP
Healthcare Cost and Utilization Project
Footnotes
Publisher's Disclaimer: The views expressed in this article are those of the authors and do not necessarily reflect those of the Agency for Healthcare Research and Quality or the U.S. Department of Health and Human Services.
References
- 1.Centers for Medicare and Medicaid Services. Emergency Medical Treatment & Labor Act (EMTALA) web page. www.cms.gov/Regulations-and-Guidance/Legislation/EMTALA. Last modified March 26, 2012. Accessed May 13, 2018.
- 2.Gotanda H, Fogel J, Husk G, et al. Hurricane Sandy: impact on emergency department and hospital utilization by older adults in lower Manhattan, New York (USA). Prehosp Disaster Med. 2015;30(5):496–502. [DOI] [PubMed] [Google Scholar]
- 3.Greenstein J, Chacko J, Ardolic B, et al. Impact of Hurricane Sandy on the Staten Island University Hospital emergency department. Prehosp Disaster Med. 2016;31(3):335–339. [DOI] [PubMed] [Google Scholar]
- 4.Kim H, Schwartz RM, Hirsch J, et al. Effect of Hurricane Sandy on Long Island emergency department visits. Disaster Med Public Health Prep. 2016;10(3):344–350. [DOI] [PubMed] [Google Scholar]
- 5.Lee DC, Smith SW, Carr BG, et al. Geographic distribution of disaster-specific emergency department use after Hurricane Sandy in New York City. Disaster Med Public Health Prep. 2016;10(3):351–361. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Lin CJ, Pierce LC, Roblin PM, et al. Impact of Hurricane Sandy on hospital emergency and dialysis services: a retrospective survey. Prehosp Disaster Med. 2014;29(04):1–6. [DOI] [PubMed] [Google Scholar]
- 7.Malik S, Lee DC, Doran KM, et al. Vulnerability of older adults in disasters: emergency department utilization by geriatric patients after Hurricane Sandy. Disaster Med Public Health Prep. 2017;12(2):184–193. [DOI] [PubMed] [Google Scholar]
- 8.Mortensen K, Dreyfus Z. How many walked through the door? The effect of Hurricane Katrina evacuees on Houston emergency departments. Med Care. 2008;46(9):998–1001. [DOI] [PubMed] [Google Scholar]
- 9.Federal Emergency Management Agency. FEMA Disaster Declarations website. https://www.fema.gov/disasters/. Accessed August 25, 2018.
- 10.National Oceanic and Atmospheric Administration. Storm Events Database website. https://www.ncdc.noaa.gov/stormevents/ftp.jsp. Accessed February 19, 2018.
- 11.National Oceanic and Atmospheric Administration. Best Track Data (HURDAT2) web page. https://www.nhc.noaa.gov/data/hurdat/hurdat2-1851-2017-050118.txt. Updated May 1, 2018. Accessed April 9, 2018.
- 12.Agency for Healthcare Research and Quality, Healthcare Cost and Utilization Project (HCUP). HCUP User Support website. https://www.hcup-us.ahrq.gov/databases.jsp. Accessed December 12, 2018. [PubMed]
- 13.National Oceanic and Atmospheric Administration. National Hurricane Center’s Tropical Cyclone Reports website. https://www.nhc.noaa.gov/data/tcr/. Accessed May 7, 2020.
- 14.Rath B, Young E, Harris A, et al. Adverse respiratory symptoms and environmental exposures among children and adolescents following Hurricane Katrina. Public Health Rep. 2011;126(6):853–860. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.