Introduction
Access to medical care in flood-prone regions can be complicated by co-occurring disasters or emergencies. Harris County, TX, is in a highly populated region that floods frequently.Reference Miranda, Callender and Canales 1 During the COVID-19 pandemic, medical care access was further restricted by flooding and freezing events. Flooding can complicate strategies designed to prevent COVID-19 infection and add layers of challenges. For example, evacuations and exposure to environmental hazards related to flooding may increase the need for health care services but may also come with the risks of increased exposure to pathogens in the emergency room or in crowded conditions at evacuation sites. The inverse is also true, where the economic damage from the COVID-19 pandemic may limit community members’ ability to absorb unexpected costs due to flooding.Reference Simonovic, Kundzewicz and Wright 2 In Houston, TX, communities that experienced storm-related damages to their homes or income loss related to Hurricane Harvey in late 2017 were also more likely to have lost income due to COVID-19.Reference Callender, Canales, Avendano, Craft, Ensor and Miranda 3 Houstonians severely impacted by Hurricane Harvey were also 5 times more likely to have severe anxiety during the COVID-19 pandemic. Although this storm occurred more than 2 years before the COVID-19 pandemic started, many communities were still recovering from Hurricane Harvey flooding and other smaller flooding events that occurred in the interim. The community perspective of access to health care is critical to the development of policy and prevention plans, particularly in regions where flooding is a common occurrence.
As natural disasters, such as flooding events, become more frequent and severe worldwide, examining the impact of natural disasters, including challenges with health care access, is increasingly important.Reference Rajkhowa and Sarma 4 The adverse effects of increased flooding have been exacerbated by urbanization, especially due to increased paved surfaces and extended floodplain boundaries.Reference Cao, Zhou, Güneralp, Li, Zhao and Zhang 5 , Reference Feng, Zhang and Bourke 6 The cost of flooding is estimated to reach billions of dollars within the next few decades in the United StatesReference Fant, Jacobs and Chinowsky 7 and expected to continue to cause extensive displacement and public health crises worldwide.Reference Kam, Aznar Siguan and Schewe 8 Unfortunately, flooding issues continue to arise even during peaks in COVID-19 infections in urban areas, such as Houston, which contributes to continued barriers to health care access, as coinciding events cause increased need for health care even as access becomes more limited.
The repercussions of flooding are particularly problematic for health care access. Flooding causes health care disruptions by closing access to health care facilities due to damaged infrastructure,Reference Salam, Wireko and Jiffry 9 and health care professionals often face displacement and damage to their personal property, making it difficult to get to central locations where they are needed most.Reference Ginex, Dickman and Elia 10 Furthermore, disasters, such as flooding, disrupt supply chains, including supply transportation, contributing to shortages of essential medical supplies and medications.Reference Syahrir and Suparno 11 There are also significant challenges in obtaining and sharing patient information during disasters, including flooding, making it harder to ensure not only the proper treatment of individual patients but also broader operational decisions and future disaster responses.Reference Sauer, Resnick, Links, Garibaldi and Rutkow 12
During the COVID-19 pandemic, there were similar problems with diminished access to health care. Common challenges faced by individuals included limited access to medication and health care services.Reference Shad, Colón-López, Pérez and López-Cepero 13 Some health care facilities suspended services,Reference Hood 14 and hospitals experienced overcrowding as they became increasingly used for primary care and lacked necessary supplies.Reference Plagg, Piccoliori, Oschmann, Engl and Eisendle 15 A shift to digital health technologies (DHTs) helped address some limitations.Reference Ndayishimiye, Lopes and Middleton 16 Still, challenges related to power outages, damaged telecommunications, and varying technological literacy of health care professionals significantly impact the effectiveness of DHTs.
While the current literature focuses on the immediate effects of natural disasters on health care access, there is a gap regarding how combined disaster events affect access to health care more generally from the community perspective.Reference Deen 17 , Reference Lee and Eom 18 This combination is particularly important, as the combination of flooding and COVID-19 infection spikes continues to affect the region 5 years after the pandemic onset. This study aims to understand the individualized and combined effects of personal flooding-related losses and COVID-19 on access to health care services in Houston.
Methods
Study Design
This survey study used a convenience sample of adults (N = 206) from community clinics, including primary care clinics run by the Baylor College of Medicine (20.39) and freestanding clinics (25.525%), (45.64%), including those providing care to uninsured patients, and libraries (60.69%) in the Houston metro area to assess how access to health and support services were affected by the COVID-19 pandemic and flooding. Libraries in Houston are popular community spaces, with more than 2.1 million annual visits from county residents in 2022. The locations chosen for survey distribution were in regions that experienced flooding. The survey was conducted between 3/1/2022 and 10/7/2022 which included weather events that caused several floods in the region. The COVID-19 pandemic was less severe, but still of concern, during this timeframe. The survey consisted of 98 questions on topics including sociodemographic characteristics, health status and conditions, health care access, flooding, COVID-19, and alcohol use. Participants received a small gift (<$5 in value) for completing the survey, such as pocket flashlights, small first aid kits, or school supplies. This study was approved by the Baylor College of Medicine Institutional Review Board.
Flooding
For flooding, if participants responded with yes when asked if “Lost an apartment or home due to flooding,” “Had to relocate to a different apartment or home due to flooding,” or “Had a car, truck, or work vehicle damaged due to flooding,” they were classified as being affected by flooding. Only respondents who answered “yes” to any of the previous questions were classified as having been affected by flooding. In total, there were 15 questions related to flooding and health care access that were assessed. A variable was created counting each “yes” to the questions related to flooding and health care access. Questions with responses that had cell sizes <5 were removed from analyses.
COVID-19
Participants were classified as having had a previous COVID-19 diagnosis if they had responded yes to the question “Has a doctor or other health professional ever told you that you had coronavirus or COVID-19?” In total, there were 13 questions related to COVID-19 and health care access. A variable was created counting each “yes” to the questions related to COVID-19 and health care access.
Flooding and COVID-19
Participants were classified as having been affected both COVID-19 and flooding if they answered yes to the question “Has a doctor or other health professional ever told you that you had coronavirus or COVID-19?” and yes, to either “Lost an apartment or home due to flooding,” “Had to relocate to a different apartment or home due to flooding,” or “Had a car, truck, or work vehicle damaged due to flooding.” A variable was created counting each “yes” to the questions related to flooding and health care access, and COVID-19 and health care access.
Data Analysis
Bivariate comparisons were made between those diagnosed with COVID-19 and those who were not, and those affected by flooding (homes/vehicles damaged) and those who were not, with chi-square analyses and t-tests. Logistic regression analyses, unadjusted and adjusted for comorbidities and race/ethnicity, were then used to compare the odds of having access to health care and other services. Poisson regression analyses, unadjusted and adjusted for comorbidities and race/ethnicity, were conducted using the number of types of health care access lost due to flooding, COVID-19, and flooding and COVID-19. Significance was assessed at P <.05 for logistic and Poisson regression analyses and P <.10 for chi-square and t-test analyses.
Results
Among 206 adults, 42 (20.39%) had lost homes or vehicles due to flooding (Table 1). The mean age in the sample was 48.26 years, and most participants were female (72.8%). In this sample, 38.54% were non-Hispanic (NH) White, 18.05% were NH Black, 20.49% were Hispanic, and 22.93% were classified as NH other, which is roughly representative of the diversity in the Houston region. Comorbidity was high, with 57.35% reporting one or more comorbidities. No significant differences in demographics, including SES, were found between those who reported they were personally affected by flooding as compared to those who were not.
Descriptive characteristics of those affected by flooding
In bivariable comparisons examining access to health care by flooding status, there were several significant differences (Table 2). Although an overall high proportion of participants reported that they were delayed getting medical care, close to twice the proportion of those who were affected by flooding reported this issue (57%) as compared to those not affected (33%). There was an association between being affected by flooding and needing medical care but not getting it (P <.01) and being delayed in filling a prescription (P <.05). There were also significant barriers to other services in the community among those affected by flooding, with significant associations found between participants being affected by flooding and reporting their grocery store closed (P <.001), that they could not get food (P <.01), having lost income (P <.001), their church was flooded (P <.05), or their community experienced significant damage (P <.001).
Health care access by flooding status
After adjusting logistic regression models for comorbidity and race/ethnicity, many of the bivariable associations remained (Table 3). Participants affected by flooding had more than 7 times higher odds of reporting delayed access to medical care (OR: 7.71; 95% CI: 2.87–20.66). Respondents affected by flooding also had higher odds of not being able to get the medical care that they needed (OR: 5.75; 95% CI: 1.48–22.26) or had delays filling a prescription (OR: 5.85; 95% CI: 1.21–28.17) as compared to those who were unaffected by flooding. Associations also remained significant for other services in the community. Those affected by flooding had higher odds of reporting closure of the closest grocery store (OR: 5.38; 95% CI: 2.35–12.28), were unable to get food (OR: 3.48; 95% CI: 1.37–8.85), lost income (OR: 5.36; 95% CI: 2.20–13.08), their church flooded (OR: 3.22; 95% CI: 1.23–8.40), their community had experienced significant damages (OR: 10.35; 95% CI: 4.57–23.46), their doctor’s office had closed (OR: 3.06; 95% CI: 1.22–7.67), and a closure in the hospital closest to them (OR: 4.72; 95% CI: 1.19-18.66) as compared to those personally unaffected by flooding.
Logistic regression analyses of health care access by flooding status
Demographic characteristics were similar among those who reported having been diagnosed with COVID-19 as compared to those without a prior diagnosis, with the exception of comorbidities (P <.05), as 69% of those who reported a COVID-19 diagnosis had a comorbidity as compared to 51% of those without a diagnosis (Table 4). An examination of the association between COVID-19 diagnosis and access to health care and community services revealed that there were no significant differences with the exception of hospital closures (P <.05) and a lack of access to a support group due to the COVID-19 pandemic (P <.05; Table 5). None of these associations remained significant after controlling for comorbidities and race/ethnicity (Table 6).
Descriptive characteristics of those who have had COVID-19
Health care access by COVID-19 diagnosis
Logistic regression analyses of health care access by COVID-19 status
In the Poisson regression analyses (Table 7), after adjusting for race/ethnicity and comorbidities, those personally affected by either COVID-19 or flooding reported higher number of health care services lost due to COVID-19 or flooding than those who were not affected by flooding or COVID-19 (P <.05). The association between personally experiencing both COVID-19 and flooding was even more strongly associated with the frequency of health care services lost by participants (P <.05).
Poisson regression analysis of health care access loss by number of adverse events
Discussion
Our findings indicate that there are differences in the effects that more localized flooding has on health care access and community services. At the same time, the COVID-19 pandemic affected services across communities, regardless of who was personally affected at the time of the survey. In addition, those with a previous COVID-19 diagnosis or who were previously affected by flooding lost a higher number of health care access services due to flooding or COVID-19. Although there are similarities in how the different situations may be addressed, it is important to consider developing individualized plans to provide health care access across a range of disaster scenarios in different communities, including those that occur at the same time. This is especially important as urban areas become more crowded and older populations move into regions susceptible to flooding and storms.Reference Bukvic, Gohlke, Borate and Suggs 19 , Reference Lane, Charles-Guzman, Wheeler, Abid, Graber and Matte 20
In flood-prone areas, such as Houston, TX, the occurrence of COVID-19 can reduce access to health care services, particularly when these events co-occur with periods of higher flooding risks, as has been true in the past 3 years. In the current study, participants with prior experiences with either flooding or COVID-19 had reduced access to certain health care services compared to those without these experiences. This is more pronounced with the participants who have experienced both flooding losses and a COVID-19 diagnosis. Multi-hazard occurrences lead to several issues, such as the economic impact of COVID-19 limiting one’s ability to absorb the unexpected cost of flooding damages or flooding evacuation, and relocating to shelters can make social distancing for contagious diseases, such as COVID-19, extremely difficult.Reference Simonovic, Kundzewicz and Wright 2 In Houston, those severely impacted by Hurricane Harvey had 5 times higher odds of reporting severe anxiety during the COVID-19 pandemic and nearly 4 times higher odds of COVID-19-related income loss.Reference Callender, Canales, Avendano, Craft, Ensor and Miranda 3 Implementation of an effective multi-hazard management plan is necessary in areas such as Houston, where flooding is a reoccurring event.
In a review of 27 US studies, 65% identified racial, ethnic, or socioeconomic disparities, with poorer physical and mental health outcomes observed among those living in low-income households.Reference Flores, Sullivan, Yu and Friedrich 21 After Hurricane Harvey caused unforeseen flooding in 2017, it was determined that lower socioeconomic status (SES) households disproportionately experienced the most extensive flooding in Houston.Reference Collins, Grineski, Chakraborty and Flores 22 Unfortunately, those neighborhoods also often have greater health care needs and experience greater barriers to receiving aid. Our study, combined with previous findings, indicates that resources need to prioritize health care access in affected communities, particularly if they have a high proportion of low-income households. This is especially important in instances, such as the Houston region, when safety-net services are in zones that often experience flooding events. Although it was a unique event, Hurricane Harvey caused an urban safety-net county hospital to lose 133 of 179 inpatient beds (74%) from water damage. The loss of beds increased wait times and overcrowding, leading to inefficient access to necessary care regardless of whether patients had comorbidities or not in a hospital that serves low-income community members.Reference Chambers, Husain and Chathampally 23 The area remains susceptible to flooding and subsequent power outages and other infrastructure-related issues that could affect services. Most studies focus on a general overview of health care access after flooding occurs rather than individual-level problems experienced due to both major and relatively minor, but still impactful, disaster events. Inequitable access can affect a variety of necessary health care services, as specialty clinics, such as dialysis centers, are often isolated and may not have the same priority for repairs as larger consolidated medical facilities.Reference Horney, Gangwal and Dong 24 Complications and adverse outcomes may arise from chronic diseases due to reduced access to such clinics in areas that experienced localized flooding.
There are several possible solutions to address the reduced access to health care that may arise as a result of flooding. In South Carolina, mobile integrated health teams are activated in rapid response and have an array of functions, such as access to medications, alternative housing options, and resources for outpatient management of chronic illnesses.Reference Gainey, Brown and Gerard 25 After Hurricane Katrina, in response to potential major hurricanes in the future, several at-risk regions have proposed solutions such as integrating social services, including local government programs and non-profits, care coordination interventions, and improving social support through local groups, particularly religious organizations and neighborhood associations.Reference Raker, Arcaya, Lowe, Zacher, Rhodes and Waters 26 In the city of Houston, the Stormwater Master Plan was created to improve effective drainage during flooding events, and the Texas Medical Center improved infrastructure, such as flood gates and water pumps, and utilized a specific protocol titled the Hospital Emergency Incident Command System.Reference Hines and Reid 27 Although these initiatives and solutions have been enacted, there are still many issues to address in order to ensure that patients receive equitable and effective care during periods of flooding.
Differences observed among those who reported COVID-19 diagnoses as compared to those without diagnoses may be explained by the prevalence and speed of transmission of the virus, which caused health care systems around the world experience service-related stress and resource allocation away from other forms of care.Reference Karan and Wadhera 28 This led several hospitals to refrain from offering elective surgeries, which were sometimes forced due to lack of equipment, PPE, or medical staff.Reference Muellers, Andreadis, Ancker, Horowitz, Kaushal and Lin 29 In Canada, widespread delays in surgical glaucoma due to relocation of medical staff at the beginning and middle of the pandemic were observed.Reference Zhang, McIntyre, Abu-Dieh and Toren 30 This reallocation of resources and health care staff during the COVID-19 pandemic may have led to reduced access to several health care services particularly among those with chronic health conditions.
Online chronic disease management programs can help address problems faced by those with comorbidities and strengthen the effectiveness of the health care system during pandemics.Reference Kendzerska, Zhu and Gershon 31 Community mobile health clinics were shown to be effective in helping marginalized and vulnerable groups get tested and gain vital access to primary care.Reference Baker, Cadet and Mani 32 Mobile clinics can aid in providing medical care during pandemics by expanding availability to people and areas who would otherwise be inhibited by personal and external factors. Expanding these services and making them more accessible during co-occurring disasters, such as flooding, may ensure communities retain access.
Furthermore, the current study found differences in delays in filling prescriptions between those who have previously been diagnosed with COVID-19 and those who had not. The pandemic facilitated a surge of medications that treated chronic and COVID-19-related symptoms, causing pharmacies to decrease in efficiency and decrease in availability of other drugs.Reference Romano, Galante, Figueira, Mendes and Rodrigues 33 Pharmacies had to contend with new safety protocols and staff shortages, also contributing to the decrease in efficiency of prescription fulfillment witnessed.Reference Khan, Dignos, Papadopoulos, Nowrouzi-Kia, Sivanthan and Gohar 34 Furthermore, border closures and reduction of transportation around the world caused disruptions of key ingredients essential to drug production.Reference Mircheva 35 Possible solutions for prescription interruptions during viral infections may include a more streamlined process and expansion of telehealth services like tele-pharmacies.Reference Aruru, Truong and Clark 36 Access to prescription medication is essential to the treatment of not just COVID-19 but also chronic condition symptom management.
The current study found significant differences in being able to receive counseling or care for alcohol or drug use between those who have previously been diagnosed with COVID-19 and those who had not. Additionally, while not significant, the adjusted odds ratio for not receiving counseling or care for alcohol or drug use due to COVID-19 was large. After adjusting for comorbidities, those who had been diagnosed with COVID-19 had 7.22 times higher odds for not receiving counseling for alcohol or drug use. After controlling for race and ethnicity, we saw the same elevated odds among those who had been diagnosed with COVID-19 for not receiving counseling. However, results did not reach statistical significance due to a small sample size. This argues for potential, but unconfirmed, disparities in access to substance disorder treatments for those with COVID-19.
The findings of this study align with previous research, as the pandemic caused impediments for treatment programs of substance abuse disorders by creating barriers in staffing, infrastructure, service provisions, and safety protocols.Reference Hemasinghe, Jayamaha and Herath 37 There was a 23.5% decrease in substance use disorder (SUD) treatment admissions from 2019 to 2020.Reference Cantor, Whaley, Stein and Powell 38 Comorbidities associated with poorer COVID-19 outcomes, such as cardiovascular disease, respiratory illnesses, and diabetes, are common in those with SUD.Reference Wang, Kaelber, Xu and Volkow 39 Furthermore, it is also important to provide and fund initiatives to improve digital literacy and provide technical support training for health care professionals treating SUD.Reference López-Pelayo, Aubin and Drummond 40
Conclusion
These findings highlight the community-level needs for planning related to expanded access to care and support services during emergency situations and co-occurring adverse events. Communities susceptible to various adverse events need a proactive disaster management plan to mitigate the greater loss in health care services. These types of plans can also help to mitigate disaster fatigue, which has been described in the Gulf Counties of Texas among residents and disaster planners. 41 In particular, cities such as Houston that are already prone to flooding must be proactive against future adverse events that may regularly occur. Communities with a high proportion of uninsured or low socioeconomic status residents are susceptible to localized damages resulting from flooding. Finally, this study offers insight into the increasing probability of co-occurring emergencies in certain geographic regions, particularly as it has become apparent that COVID-19 peaks occur at the same time as flooding risks increase in the Houston region. This manuscript aims to improve awareness among disaster researchers that communities may be affected by multiple events that can seriously strain their ability to meet their health care needs, particularly for those with chronic conditions. Since the different events can be expected to affect access differently, it could also be expected that coinciding events may have an effect on different layers of access to care and treatment. Plans for communities that flood frequently in Houston should include how to expand access for those with comorbidities who may have restricted ability to get needed services, particularly for services that may be overlooked, such as mental health care and pharmacies. While the COVID-19 pandemic created more generalized issues related to health care access, it mostly affected those with comorbidities in accessing health care services, such as hospitals restricting access due to the COVID-19 pandemic and lacking access to support groups.
Strengths and Limitations
Limitations to the study include convenience sampling and a small sample size. We could not calculate the response rate. This limits the generalizability of the study and may not be representative of the Houston metro area. The study was also subject to recall bias. Participants may be unable to fully recall all services they did not have access to. Also, the question of COVID-19 infection may not have captured those who developed a COVID-19 infection but were not diagnosed by a physician or other health care worker. It is also possible that the response to this question was affected by social desirability bias, as many who had COVID-19 in the region may not have wanted to report it on a survey, although the anonymous nature of the survey may have reduced this potential issue. Finally, this study was not published sooner because the effects of COVID-19 were originally expected to wane or reach peaks during periods when flooding was not as significant a concern, such as in the winter. However, since the survey, peak COVID-19 has occurred regularly during peak storm seasons, and flooding has coincided with peak infection periods regularly since, indicating the need for continued consideration of these issues.
Some strengths from the study are that it allows us to examine how health care access is reduced at a personal level due to flooding and COVID-19. Some strengths from the study are that it allows us to examine how health care access is reduced at a personal level due to flooding and COVID-19.
Author Contributions
Substantial contributions to the conception or design of the work; or the acquisition, analysis, or interpretation of data for the work: Willie Rajvong, Wilford Robinson, Justin Nguyen, Christine Walther, Roger Zoorob, and Jacqueline Hirth. Drafting the work or reviewing it critically for important intellectual content: Willie Rajvong, Wilford Robinson, Justin Nguyen, Christine Walther, Roger Zoorob, and Jacqueline Hirth. Final approval of the version to be published: Willie Rajvong, Wilford Robinson, Justin Nguyen, Christine Walther, Roger Zoorob, and Jacqueline Hirth. Agreement to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved: Willie Rajvong, Wilford Robinson, Justin Nguyen, Christine Walther, Roger Zoorob, and Jacqueline Hirth.
Funding Statement
This work was supported by the Department of Health and Human Services, Health Resources and Services Administration, Grant number T32 HP10031.
Competing Interests
The authors have no conflicts of interest to declare.
