Abstract
Introduction:
The aftereffects of Hurricane Maria have been devastating to Puerto Ricans already overburdened by environmental fluctuations due to climate change, poverty, and government infrastructure in crisis. The hurricane contributed to the loss of housing, power, family and friends, and sense of security. This study aimed to identify the predictors of Puerto Ricans' perceived stress changes and mental health as self-reported after Hurricane Maria.
Methods:
Data collected by Kaiser Foundation-Washington Post in 2018 were obtained and analyzed using SAS v9.4 complex survey procedures. After the population weight was applied, multinomial logistic regressions were employed to identify predictors of participants' perceived stress changes and self-reported level of mental health.
Results:
Those whose general stress level was reported as worse after Hurricane Maria were more likely to have reported having generally poor mental health (OR 20.67; 95% CI 95% 8.60–49.69) and to have been impacted by the hurricane in various ways (OR 1.85; 95% [1.21,
Discussion:
A significant effect was shown for those who reported worse stress levels than those with better stress levels after the hurricane, and for those who reported good, fair, or poor general mental health.
Conclusion:
Owing to the impact of environmental disaster and current pandemic crisis on Puerto Rican citizens, it is imperative for both local and federal governments to enact plans to lessen resident mental health burden to reassure and facilitate access to mental health services after natural disasters.
INTRODUCTION
Hurricane Maria was the strongest storm to hit Puerto Rico in 85 years, leaving over a million residents without power on the island; many were left without power for >11 months. Environmental justice cannot be achieved on the island without addressing the impact of natural disasters and providing protection from environmental and health hazards. Hurricane Maria negatively impacted environmental justice efforts through water pollution, massive loss of electrical power throughout the island, and destruction of homes and roadways. This disaster exacerbated existing environmental disadvantages such as poor living conditions and poverty-related issues while compounding physical and mental health effects. Recovery efforts for this particular natural disaster are anticipated to continue for many years. 1 , 2 The short- and long-term effects of Hurricane Maria on Puerto Rico have resulted in the displacement and migration of >200,000 Puerto Ricans to the continental United States. There are long standing environmental justice issues with water and soil contamination on the island as well as loss of housing due to past hurricane disasters and economic problems that produce uninhabitable areas where the displaced and those affected by the hurricane suffer. 3 , 4
The lack of local and federal response to provide sufficient aid during the hurricane is reflected in the present COVID-19 emergency response. COVID-19 response has not been properly organized or addressed by the federal government. Out of the 2,183,400 total cases reported for the United States, 6195 of those are from Puerto Rico. 5 The lack of a firm response for the people of Puerto Rico who have already suffered from a natural disaster could create higher levels of psychological distress due to fear caused by the pandemic, which can lead to an increase in mental health issues and psychological distress in this population. 6
The effects of environmental injustices on stress and mental health need to be studied further in the case of disasters. There is a need to study the impact of Hurricane Maria's effects on the mental health and stress of Puerto Rican residents, especially those who have been displaced or have suffered ill effects due to the pandemic. Similar hurricanes, such as hurricanes Katrina and Sandy, were reported to have a significant effect on psychological distress, use of mental health services, and displacement of those directly affected by the hurricanes. A study conducted by Boscarino and colleagues showed a strong occurrence of post-traumatic stress, sleep problems, pain, and suicidal thoughts from participants impacted by Hurricane Sandy. 7 Wang and colleagues found that those who had psychological distress after Hurricane Katrina were not receiving mental health care due to factors such as low socioeconomic status and race/ethnicity while displaced. 8 There is also recent literature concerning the mental health outcomes of flooding on residents in the United Kingdom and in New Jersey, showing that there is a relationship between severe flooding, depression, and post-traumatic stress, as well as psychological impact due to displacement. These studies draw the conclusion that funding for short- and long-term crisis management needs to be increased. 9
The purpose of this study was to identify the factors associated with Puerto Ricans' perceived stress changes and the changes of self-reported level of mental health in Puerto Ricans who resided in Puerto Rico during and after Hurricane Maria. In addition, the effects of displacement and the impacts of the hurricane such as new or worse physical or mental health issues, major damage or destroyed home or vehicle damage, economic or job loss, water access, power loss for >3 months were also identified as potential predictors of perceived stress changes and the changes of self-reported level of mental health.
MATERIALS AND METHODS
This was a secondary data analysis of the 2018 Washington Post-Kaiser Family Foundation Survey of Puerto Rico Residents carried out a year after Hurricane Maria. The sample included 1500 adults over the age of 18 years residing in Puerto Rico during the time of the survey and during the time Hurricane Maria passed through the island from September 20 to 21, 2017, with harsh winds and rain.
Study region and population sample
In each region, 100 sampling points or census block groups were randomly sampled. In each of the census block groups, 15 adults were selected for interviews by contacting selected households using a systematic pattern of every fifth house (moving to the sixth house if fifth house did not respond). A multistage weighting process involving the computing of household weight and creation of a population weight was used. The population weight created was then matched to demographic estimates for Puerto Rico's adult population based on Census Bureau's 2012–2016 American Community Survey (ACS) to the randomly selected sample. Puerto Rico's reported population for demographics was used to reflect the estimated population in each of the five regions selected for sampling. Data were then balanced using overall 2016 Puerto Rico ACS estimates.
Collection of data
Data for the study were collected from July to August 2018 through in-person interviews. Interviews conducted for this survey were completed in Spanish and all questionnaire items were administered during interviews in the same order. At the completion of data collection, Kaiser Family Foundation proceeded to send the data to SSRS Full Service Survey and Market Research Firm in Pennsylvania to test validity and reliability through internal validity checks with the analysis yielding plausible data with no indication of duplicate data collected. 10 Participants with missing information concerning stress and mental health were excluded from the analysis conducted for this study (N = 1490).
Variables
Dependent variables
General level of stress
General level of stress was assessed using the following question: “Compared to before Hurricane Maria, is your general level of stress better, worse, or about the same today?” The response options for this question were better, worse, or about the same.
General mental health
General mental health of the participants was assessed by the following question: “Compared to before Hurricane Maria, in general, your mental health is excellent, very good, good, fair, or poor?” The response options for this question were excellent, very good, good, fair, or poor.
Independent variables
Independent variables included in analysis were age (under 18, 18–24, 25–44, 45–64 and 65+ years), whether problems with getting medical care after the hurricane (yes, no), unable to fill prescriptions after the hurricane (yes, no), trouble getting urgent or emergency care (yes, no), and having to go to different doctor or clinic than before hurricane (yes, no).
Additional independent variable questions included: “Did you evacuate or leave your home for any amount of time as a result of Hurricane Maria, or not?” (yes, no) and whether participant was impacted by the hurricane including new or worse physical or mental health issues due to storm, major damage, or destroyed home or vehicle damage, economic or job loss, water from natural source, gone without power for >3 months (yes, no).
Statistical analyses
Data were analyzed using SAS 9.4 complex survey procedures to account for the effect of weighting and sampling design. Frequencies, proportions, means, and standard deviations were calculated for all variables.
Dependent variable general level of stress
Chi-square analysis was conducted to assess associations for level of stress after Hurricane Maria with the following independent variables: age, gender, political party affiliation, reported mental health status after the hurricane, impact of the hurricane, and difficulty in getting medical care, unable to fill prescriptions, trouble getting urgent or emergency care, and having to change doctor or clinic after the hurricane.
Multinomial logistic regression was employed to examine associations between independent variables and change in level of stress posthurricane. The following independent variables were considered: age, gender, perception of mental health status, impact of the hurricane, difficulty in getting medical care, unable to fill prescriptions, trouble getting urgent or emergency care, and having to change doctor or clinic.
Dependent variable general mental health
Chi-square analysis was conducted to assess associations between reported general level mental health status after the hurricane and the following independent variables: age, impact of the hurricane, level of stress after Hurricane Maria, whether they evacuated their home as a result of the hurricane, difficulty in getting medical care, unable to fill prescriptions, trouble getting urgent care or emergency care, and having to change doctor or clinic.
Multinomial logistic regression was employed to examine associations between independent variables and change in general mental health status posthurricane. The following independent variables were considered: level of stress, age, gender, impact of the hurricane, difficulty in getting medical care, unable to fill prescriptions, trouble getting urgent or emergency care, and having to change doctor or clinic after the hurricane.
RESULTS
Descriptive analysis
All census-designated age groups reported their general level of stress after the hurricane to be worse, with age groups <18 (16.07%), 45–64 (20.28%), and ages >65 years (14.4%) having the highest levels as given in Table 1. Those who reported having worse levels of stress after the hurricane than before the hurricane also reported being impacted by the hurricane (45.23%). Those who did not evacuate or leave home as a result of the hurricane (44.2%) reported having worse levels of stress after than before the hurricane.
General Level of Stress Compared with Before Hurricane Maria in a Sample of Puerto Ricans from July to August of 2018
Weighted percentages.
Chi square test for association.
The majority of participants in all age groups reported good, fair, or poor mental health as given in Table 2. Those under the age of 18 years (10.87%), ages 45–64 years (10.98%), and ages 65 years or older (8.45%) reported good mental health status after compared with before the hurricane. The highest participants reporting of fair mental health were under the age of 18 years (8.55%), and 65 years or older (6.02%). Those impacted by the hurricane (20.20%) and those who did not evacuate or leave their home due to the hurricane (16.28%) reported having fair mental health after compared with before the hurricane.
General Level of Mental Health Compared with Before Hurricane Maria in a Sample of Puerto Ricans from July to August of 2018 a
Weighted percentages.
Chi square test for association.
Associations with stress
As given in Table 3, those whose general stress level was reported as worse compared with those who reported better after Hurricane Maria were less likely to be 65 years or older (OR 20.67; 95% CI 95% 8.60–49.69), more likely to have reported having generally poor (OR 20.67; 95% CI 95% 8.60–49.69), fair (OR 6.67; 95% [4.19, 10.62]), or good (OR 2.26; 95% [1.45, 3.50]) mental health, more likely to have reported having to go to a different doctor or clinic after Hurricane Maria (OR 1.64; 95% [1.04, 2.59]), more likely to have been impacted by the hurricane in various ways (health, damage/destruction of home or vehicle, monetary or job loss, loss of power, and water quality) (OR 1.85; 95% [1.21, 2.85]).
General Level of Stress Compared with Before Hurricane Maria in a Sample of Puerto Ricans from July to August of 2018
Significant values: ***p < 0.001, **p < 0.01.
Associations with mental health
As given in Table 4, the participants who reported their general level of mental health as fair, compared with excellent, were more likely to be 25 years or older (ages 25–44 years: OR 13.42, 95% CI [1.01–22.63]; 45–64 years: OR 4.15, 95% CI [1.97–8.75]; 65+ years: OR 2.47, 95% CI [1.57–3.87]), more likely to report their general stress level as worse (OR 4.26; 95% CI [2.55–7.10]).
General Level of Mental Health Compared with Before Hurricane Maria in a Sample of Puerto Ricans from July to August of 2018
Significant values: ***p < 0.001, **p < 0.01, *p < 0.05.
ND, no data.
Participants who reported their general level of mental health as poor compared with excellent were more likely to be 25 years or older (ages 45–64 years: OR 22.08, 95% CI [2.39–204.13]; 65+ years: OR 2.40, 95% CI [1.19–4.84]), more likely to report their general stress level as worse (OR 10.54; 95% CI [4.06–27.37]), to have reported having to go to a different doctor or clinic after Hurricane Maria (OR 2.49; 95% CI [1.20–5.65]), and more likely to have reported having to leave or evacuate for any amount of time as a result of Hurricane Maria (OR 2.44; 95% CI [1.20–5.65]). In addition, they were less likely to have been impacted by the hurricane in various ways (health, damage/destruction of home or vehicle, monetary or job loss, loss of power, water quality) (OR 0.19; 95% CI [0.04–0.87]).
DISCUSSION
Hurricane Maria devastatingly impacted Puerto Rico, leaving a trail of loss of family members, homes, jobs, electricity, and health services. The lack of coordination and insufficient aid while Puerto Rico struggles economically during a natural disaster are an environmental justice issue. In addition, the local and federal governments demonstrated that they are not prepared to tackle the multitude of disastrous consequences arising from an environmental disaster of this magnitude. Although in these instances having federal assistance is needed to diminish casualties, and address environmental and human health concerns, what Puerto Rico gets is short-term assistance for some, an agitated political climate, and an increase in human health problems due to displacement. The environmental injustices examined in this study include the impacts of the hurricane such as evacuation, new or worse physical or mental health issues, major damage or destroyed home or vehicle damage, loss of access to health care, economic or job loss, loss of water access, and power loss for >3 months. The environmental injustices caused here are a crucial political topic for environmentalists and other activist groups to address in order for the federal government to better respond to Puerto Rico's needs.
The results show how the impacts of hurricane and displacement affected the stress levels and mental health of those living in Puerto Rico during the hurricane. The lack of organized emergency response reported after the hurricane could have attributed to the stress and poor mental state of many in Puerto Rico indirectly during this time as well. 11 Smaller studies conducted after Hurricane Maria in specific areas of the island showed clinically significant symptom elevations for generalized anxiety, post-traumatic stress, and major depression. 12
Comparing stress levels from before to after Hurricane Maria and mental health from before to after Hurricane Maria
Stress when left unattended can lead to psychological distress and deterioration of health. 13 , 14 , 15 A significant effect on general levels of stress reported as worse for after the hurricane was associated with impacts due to the hurricane and reported general mental health as poor and fair. In addition, there is significant effect on general levels of mental health reported as poor after the hurricane that were associated with impacts due to the hurricane. Participants who stated evacuating Puerto Rico due to the hurricane also reported their stress levels as worse after the hurricane.
The results from the multinomial logistic regression for general stress level show that participants who reported having worse stress after the hurricane, compared with those who reported their stress as better, were less likely to be 65 years or older and more likely to have reported having generally poor, fair, or good mental health. The participants who reported worse stress after Hurricane Maria were also more likely to have reported having to go to a different doctor or clinic after Hurricane Maria and more likely to have been impacted by the hurricane in various ways (health, damage/destruction of home or vehicle, monetary or job loss, loss of power, and water quality) (Table 3).
One of the major effects of Hurricane Maria was flooding, 16 , 17 which has been incorporated to the impact question in the Kaiser Foundation and Washington Post survey. Other researchers have looked at long-term psychosocial impacts of flooding nationwide. The study conducted by Lamond and colleagues showed how different phases of the disasters were linked to factors such as relocation, adaptation, and recovery. Relocation and adaptation due to flooding were found to be related to mental health symptoms. Relocation during flood recovery was associated with a sixfold increase in mental health symptoms. Adaptation to limit future flood damage was related to lower levels of symptoms. 18
In line with the results reported in this study, other researchers found that displacement after Hurricane Maria poorly affected the mental health of those displaced. Scaramutti et al. assessed the mental health of Puerto Ricans who stayed on the island and those who moved to Florida after Hurricane Maria, reporting a significantly greater proportion of Puerto Ricans who stayed on the island living in rural/suburban areas had high levels of clinical anxiety and post-traumatic stress disorder. 19
Strengths and limitations of the study
The data collected by Kaiser Foundation and Washington Post and analyzed in this study were some of the most comprehensive data collected on Puerto Ricans concerning the impact of Hurricane Maria with a wide range of variables of potential impacts. This was also a large data set that is representative of the demographics of the population of Puerto Rico (N = 1490), representing the entire island, as opposed to other studies that only collected data in certain municipalities. In addition, the systematic sampling chosen for data collection by Kaiser Foundation and Washington Post reduced selection bias.
This study is limited as it is a secondary data analysis, the researchers were not able to reach out to those who participated in this study and collect additional data. The data collected are also focused on one specific event, Hurricane Maria, showing on amplification of stress and mental health levels, which is an important environmental phenomenon to study at all aspects but could lead to an inaccurate estimation of overall distress and poor mental health currently in this population. The psychological impacts of a category 5 hurricane of this magnitude are not always fully studied. Collected data are not always fully representative of a population living through a natural disaster such as the data collected by Kaiser Foundation and Washington Post. Future data collection for a similar phenomenon should take into consideration full representation of the population.
The data collected by Kaiser Foundation and Washington Posts were retrospectively a year after the hurricane, giving time for long-term implications of the hurricane to be studied. Although the data collection gave time for long-term impacts to be recorded, unpredictable future phenomena that would befall the island could not be factored in the data collection. The psychological distress after Hurricane Maria can also not be fully quantified at this time from past data due to the additional psychological distress caused by a lack of emergency response in Puerto Rico in response to COVID-19. The amplification of distress due to the COVID-19 pandemic at this time could also be an additional layer of stress that can play a factor in general mental health levels of those living on the island, who are not able to access mental health services easily due to the pandemic.
This data set had a long range of important questions concerning short- and long-term impacts of the hurricane as it pertains to physical health, mental health, and health care access. The data set also included questions concerning death of family members, Federal Emergency Management Agency responses, party affiliation, job, and monetary loss, as well as other questions. Although the range of questions allowed for a large possibility for analysis, due to the wording in some of the questions in the survey tool, data analysis was limited.
The number of variables stemming from the questions in this survey is extensive, but the question construction hindered the collection of data for certain variables. For instance, one of the variables analyzed in this study that showed no significance for this study asked participants what political party they were affiliated with (having federal and local government parties combined) and proceeded to have participants choose one option. Federal and local government parties are not mutually exclusive to each other, and if the survey question were split to collect both federal and local party information, it might have given a clearer picture of affiliations.
Wording for the general mental health question, “In general, your mental health is excellent, very good, good, fair, or poor?” did not show a fair balance in answer choices. FEMA questions were a missed opportunity: only those who acquired FEMA funding were asked about their experience with FEMA, and those who applied but did not acquire funding were not accounted for in the data collected. Those who did not know about funding were not asked follow-up questions about their awareness of FEMA funding. Another limitation is that stress and mental health were self-reported and not clinically assessed. Self-reporting can lead to response bias in which participants chose the answer they think the researcher wants or makes them look better over a more true response to the question. Also, the level of stress measure is not sensitive enough to fully capture high stress or low stress posthurricane. Information regarding these variables could have had a major impact on the data and analysis of mental health level and stress levels.
CONCLUSION
The data collected by Kaiser Foundation and Washington Post were one of the few data sets with such a vast amount of questions asked to Puerto Ricans after Hurricane Maria that took into consideration the island as a whole and not only the specific municipalities that were hit hardest by the hurricane. The questions seen in the survey are usually not asked in a sample of this size, and opened doors for analysis on specific variables that in the past would not be looked at together. In the future, looking to the COVID-19 pandemic and future environmental disasters, a survey of this size could be beneficial to understanding stress and mental health in Puerto Rico after a pandemic. Cross-examining a participant's self-reported levels of stress and mental health with that of a clinical assessment could enhance the measurement of these two variables. Natural disasters can have direct and indirect effects that need to be studied in the long term to improve access and deployment of services to those in greatest need. In addition, when collecting data on future natural disasters, researchers should consider collecting information on political party affiliation, trust in government services and programs, along with collecting information concerning mental and physical health effects.
HUMAN RESEARCH STATEMENT
This study used publicly available deidentified data and as such is not deemed as a human subjects research study.
Footnotes
ACKNOWLEDGMENT
We thank the Kaiser Family Foundation for releasing the data early for this study.
AUTHOR DISCLOSURE STATEMENT
No competing financial interests exist.
FUNDING INFORMATION
There were no funding sources used to support the study described in this article.