Psychological Health, Resilience Outcomes and Disaster Preparedness: The Impact of Natural Hazards on Adults in Greece. A Cross-Sectional Analysis
- 1. Department of Shipping, University of the Aegean, Greece
- 2. Trustilio B.V., Amsterdam, Netherlands
- 3. Department of Maritime Transport and Logistics, The American College of Greece, Greece
- 4. Department of Financial and Management Engineering, University of the Aegean, Greece
Abstract
This study examines and compares the psychological well-being, consequences, resilience outcomes, and disaster preparedness levels of adults in Greece who have faced natural hazards. Our cross-sectional study, encompassing 757 online participants, incorporated the Depression Anxiety Stress Scale (DASS21), the Post-Traumatic Stress Disorder Checklist (PCL), and the Work and Social Adjustment Scale (WSAS). The findings illuminate a compelling narrative: heightened severity in PTSD symptoms correlates significantly with increased anxiety (r=0.67; p<0.001), depression (r=0.70; p<0.001), and stress symptoms (r=0.72; p<0.001), as well as an augmented total DASS-21 score (r=0.75; p<0.001). Moreover, elevated PCL-C scores (r=0.66; p<0.001) and heightened scores across DASS subscales (r ranged from 0.62 to 0.69; p<0.001) and the total score (r=0.69; p<0.001) are significantly associated with increased WSAS scores. A meticulous multiple linear regression analysis underscores the significance of age, gender, annual family income, and cohabitation with the elderly or disabled person in predicting PCL-C scores. Notably, a higher annual family income is correlated with lower levels of depression, anxiety, and stress symptoms. Intriguingly, participants who recently experienced a fire exhibited significantly greater stress symptoms than those who experienced an earthquake. The implications of our findings underscore the need to prioritize developing and implementing behaviour change interventions at the community level aimed at enhancing resilience. This holistic approach contributes to continuous education and preparedness in the face of significant disaster events. To the best of our knowledge, this study stands as the pioneering initiative of its kind. However, further research is warranted to validate the applicability and reliability of these groundbreaking findings.
Keywords
• Decision making
• Disasters
• Risk Assessment
• Psychological Health
CITATION
Kiosklia K, Tsirimpaa A, Karampotsis E, Polydoropoulou A. (2024) Psychological Health, Resilience Outcomes and Disaster Preparedness: The Impact of Natural Hazards on Adults in Greece. A Cross-Sectional Analysis. J Behav 7(1): 1024.
ABBREVIATIONS
DASS-21: Depression Anxiety and Stress Scale – 21; PCL: Post-Traumatic Stress Disorder Checklist; WSAS: Work and Social Adjustment Scale; STROBE: Strengthening the reporting of Observational studies in Epidemiology; PCL-C: Post-Traumatic Stress Disorder Checklist Civilian; PTSD: Post-Traumatic Stress Disorder
INTRODUCTION
Natural hazards, encompassing phenomena such as earthquakes, floods, droughts, and hurricanes, wield profound and far-reaching consequences on worldwide populations, economies, and ecosystems. Over recent years, there has been a discernible escalation in both the frequency and intensity of these natural disasters, resulting in consequential human and economictolls on a global scale [1]. The United Nations Office for Disaster Risk Reduction reports that from 2000 to 2019, approximately 4.2 billion individuals were impacted by natural disasters, resulting in more than 1.23 million fatalities and economic losses totaling $2.97 trillion on a global scale. Furthermore, the effects of climate change are predicted to worsen the consequences of natural hazards, amplifying the jeopardy faced by human lives and livelihoods [2], [3,4], p. 5. Addressing the impacts of natural hazards requires coordinated efforts by governments, communities, and international organizations to build resilience and reduce vulnerability to these events [5]. Natural disasters hold global importance due to their capacity to impact the physical and psychological well-being of individuals and communities, subsequently impeding their capability to engage in daily tasks and responsibilities [6].
[Recent surveys conducted in the Eastern Mediterranean region have exposed a notable dearth of observations and comprehensive information about flash floods [7,8]. Despite a substantial history of flash floods and extreme storms in the area, there exists a scarcity of detailed event descriptions, and the available instrumentation for monitoring needs to be more adequate and sparsely distributed [9]. Greece, in particular, registers one of the highest frequencies of flash flood events annually, coupled with significant associated injuries, fatalities, and consequential effects and damages [10].
Floods pose a recurring challenge in Greece, a country marked by diverse landscapes and climatic conditions. Influenced by factors such as intense rainfall, varied topography, and human activities, flood events are particularly prevalent, causing significant socio-economic repercussions. Greece’s mountainous regions, valleys, and coastal plains contribute to rapid rainfall- runoff, heightening the risk of flash floods, especially in areas with steep gradients [11].
Throughout its history, Greece has grappled with impactful floods, some unfolding suddenly, challenging prompt response and evacuation efforts. The repercussions extend beyond immediate physical damage, affecting agriculture, infrastructure, and communities, often resulting in loss of life and displacement. An illustrative example occurred on November 15, 2017, when a high-intensity storm struck the western part of the Attica region in Greece, unleashing catastrophic effects on the towns of Nea Peramos and Mandra. The toll was devastating, with 24 lives lost, marking this natural hazard as the deadliest event in Greece over the past 40 years and the most catastrophic overall. The infrastructural fabric suffered severe damage, leaving hundreds homeless after the flash floods. Tangible and intangible losses ensued, damaging businesses, residences, industrial structures, vegetation, and agriculture. Meteorological archives underscore the rarity of this flash flood within a century-based timeframe, characterizing it as an extreme event experienced by only a select few [12,13].
In a more recent occurrence, following intense and prolonged rainfall, the towns of Mandra and Magoula, located west of Athens, experienced a recurrence of flash floods on June 26th and 27th, 2018. Fortunately, there were no reported fatalities during this event; however, numerous streets, houses, and establishments were submerged in floodwater. The Fire Service was mobilized to rescue motorists who were trapped in their vehicles and responded to numerous distress calls from local residents. Data on this hazard was collected by the Copernicus satellite service. It has been forecasted that Attica will suffer significantly from flooding in the future due to fire destruction. Behavioral patterns due to demographics and impacts of floods and fires have been reported [14]. Many individuals were compelled to evacuate either by foot or utilizing private or public transportation, including sea vessels. In Greece, complete evacuations were carried out in both scenarios. However, the utilization of evacuation modeling can enhance evacuation planning by offering valuable information on safety measures and bolstering community resilience through educational initiatives [15].
The psychological well-being of individuals affected by natural hazards can be profoundly affected, both in the immediate aftermath and over a prolonged period. Events such as earthquakes, floods, and hurricanes can elicit emotions of helplessness, fear, anxiety, and depression, while also contributing to the development of post-traumatic stress disorder [16]. Moreover, these impacts can persist for years after the disaster, affecting individuals’ quality of life and ability to function normally [17]. People who already have pre- existing mental health conditions are especially susceptible to these consequences. Nonetheless, effective interventions like psychological first aid, cognitive behavioral therapy, and group therapy can assist in alleviating the adverse psychological effects caused by natural hazards [18]. As natural hazards become increasingly frequent and severe, it is crucial to address their psychological impacts on affected communities and implement effective interventions to promote mental health resilience.
Psychological resilience can be critical in mitigating the adverse effects that often result from a natural disaster. It pertains to the capacity to adjust and thrive when confronted with significant challenges or hardships [19]. A person’s resilience is not determined by a single factor, but rather by a combination of coping mechanisms, personality traits, and encountering substantial risks or enduring severe hardships [20]. Achieving successful adaptation to the stressor is also a critical component of resilience [21]. Additional factors that can mitigate the adverse consequences of natural disasters include being adequately prepared for such events, possessing knowledge about issues like vicarious trauma, compassion fatigue, and burnout, and being self-aware of one’s own responses to stressors [22]. Furthermore, personal, and broader social factors, including safety, access to education and employment, can also play a role in influencing and mitigating the effects [23]. In the event of a natural disaster, educational institutions can contribute to minimizing the negative impact on students’ well- being and academic performance by proactively preparing and implementing strategies such as offering flexible deadlines and learning options, as well as demonstrating sensitivity to students’ health and personal circumstances [24].
Disaster preparedness is a critical component of reducing the impact of natural hazards on psychological health and promoting resilience outcomes. Effective disaster preparedness measures, including early warning systems, evacuation planning, and community education, can help mitigate the psychological impact of disasters by minimizing the likelihood of injury or death and enabling individuals to better cope with the aftermath of disasters. Recent studies have highlighted the importance of disaster preparedness in fostering resilience outcomes when confronted with natural hazards. As an instance, research conducted by Ahern [25], discovered that community-based disaster preparedness programs were effective in improving community resilience and psychological well-being following a natural catastrophe. Similarly, a study by Lin [26], found that individual-level disaster preparedness, including emergency supplies and evacuation plans, was associated with lower levels of psychological distress in the aftermath of a natural disaster. These findings emphasize the significance of being prepared for disasters to enhance psychological well-being and foster resilience when confronted with natural hazards.
The objective of this study was to assess and compare the psychological well-being, effects, resilience, and preparedness for disasters in adults who have experience natural hazard impacts in Greece, such as, the extreme flood in Mandra in 2017 or the fatal wildfires in Mati in 2018. This paper presents findings from an online survey that utilized three standardized psychometric scales to assess the impacts of natural hazards on depression, anxiety, post-traumatic stress, work and social adjustment, and resilience.
MATERIALS AND METHODS
Study Design and Sample
The research design employed in this study was a cross- sectional online survey. To ensure the study’s structure and methodology adhered to best practices, the STROBE checklist for cross-sectional studies was utilized as a guide [27]. There was a total of 757 participants who provided complete answers to the questionnaires and were included in the analysis.
Ethical Approval
Ethical approval was obtained from the Institution Review Board of the American College of Greece (reference number: 202212333). Since the survey was conducted anonymously, it was impossible to directly assist any participants who scored high on the psychometric tests and showed signs of a possible clinical issues. To address this issue, the information sheet provided contact information for relevant support services that were available.
Sample Size
To determine the required sample size, we employed the a priori estimation method that considered several factors. One of these factors was using comparable studies as a reference to estimate the sample size required for multiple regression analyses [28,29]. We conducted a sample size calculation for our study, considering a regression equation with 12 predictors and a medium effect size. Based on a priori estimation of f2=0.15 (medium effect) and a power of 0.80 [30], we determined that a minimum sample size of 127 would be required. Considering potential missing data and the need for an adequate sample size to perform secondary validity analyses and sensitivity analyses based on recruitment mode, we aimed to recruit at least 700 participants.
Procedure
This study received ethical approval, and informed consent was obtained from participants who met specific inclusion criteria, including being 18 years of age or older, residing in Greece, having experienced the impact of a natural hazard, and possessing the ability to participate and provide informed consent. Participants who were unable to understand the information presented in Greek or English were excluded due to the lack of resources for survey translation or validation of measures in other languages. Online recruitment methods were employed to ensure a diverse sample, utilizing platforms such as Facebook, Twitter, LinkedIn, online forums like Reddit and Quora, and relevant mailing lists associated with organizations and communities focused on emergency management, disaster response, and resilience, such as FEMA and the Red Cross. Potential participants were invited to voluntarily participate by completing an online questionnaire, providing widespread access and convenience. Initially, 1,196 responses were received; however, 439 participants were excluded either for responding too quickly, which could compromise data quality, or for completing less than 50% of the questionnaire. Online recruitment aimed to achieve a broader sample and reach the targeted sample size.
Measures
Participants who agreed to take part in the study were requested to complete a set of assessment measures that had been psychometrically validated. The survey also included self-report questions to evaluate additional variables, which encompassed the following: demographic and socioeconomic characteristics, experiences related to natural hazard impacts, and disaster preparedness. The survey was conducted through Lime Survey (https://www.limesurvey.org/), which is a commonly used survey platform that is simple to use and does not require technical expertise for setting up and collecting data. It is flexible and user-friendly, used by many institutions.
The Depression, Anxiety and Stress Scale (DASS-21)
The participants’ psychiatric evaluation was measured using the self-rated DASS-21. The DASS-21 is composed of three subscales, each consisting of seven self-report items. These subscales assess the levels of depression, anxiety, and stress experienced by the participants within the previous week. The DASS-21 is a validated tool that can accurately measure these three dimensions in both clinical and nonclinical populations [31]. Previous studies have provided evidence of the strong internal consistency of the DASS-21, as indicated by coefficient alpha values of 0.91, 0.84, and 0.90 for the depression, anxiety, and stress subscales, respectively [32]. A study conducted with a non- clinical sample reported mean scores of 3.56 (standard deviation = 5.39, median = 2) for anxiety, 5.55 (standard deviation = 7.48, median = 3) for depression, 9.27 (standard deviation = 8.04, median = 8) for stress, and 18.38 (standard deviation = 18.82, median = 13) for the total scale [33]. In the present sample, the DASS-21 exhibited satisfactory internal consistency (Cronbach’s α = 0.90).
The Post-Traumatic Stress Disorder Checklist (PCL)
The PCL is a self-report measure consisting of 17 items, based on the DSM-IV criteria for PTSD. It is commonly used for screening, aiding in diagnosis, and monitoring symptom progression [34]. The PCL-C version was utilized in this study, with scores ranging from 17 to 85. Higher scores indicate greater impairment [35]. The PCL has demonstrated strong internal consistency, test-retest reliability, and significant convergent validity with other PTSD assessment scales like the Mississippi PTSD Scale and MMPI-2 Keane PTSD Scale [36-38]. A previous cross-sectional study conducted on a population of university students in the US (n = 392) reported an average PCL score of 29.4 with a standard deviation of 12.9 [37]. According to the National Centre for PTSD, a cut-off score between 30 and 35 is recommended, where prevalence is below 15%. Hence, a score above 35 was utilized in this research to identify scores beyond the ‘normal’ range. In the current sample, the PCL exhibited good internal consistency (Cronbach’s α = 0.88).
The Work and Social Adjustment Scale (WSAS)
The WSAS is a self-report measure consisting of five items that assess functional impairment in various domains, including work and social areas. Scores on the WSAS range from 0 to 40, with higher scores indicating more significant impairment [39]. The WSAS has demonstrated excellent internal consistency and a test-retest correlation of 0.73. Mundt et al. [39], determined that scores above 20 indicate “moderately severe psychopathology.” Scores between 10 and 20 suggest significant functional impairment but less severe clinical symptomatology, while scores below 10 are associated with subclinical populations. In a New Zealand cross-sectional study examining the psychological effects of the Christchurch earthquakes on university staff (n = 119), the average WSAS score was found to be 8.6 with a standard deviation of 8.7 [40]. In the current sample, the WSAS demonstrated good internal consistency (Cronbach’s α = 0.93).
Statistical Analysis
Descriptive variables were presented as mean values with their corresponding standard deviations (SD), while qualitative variables were expressed as absolute and relative frequencies. To examine the relationship between two continuous variables, Pearson correlation coefficients (r) were utilized, while Spearman correlation coefficients (rho) were used to assess the association between a continuous and an ordinal variable. Multiple linear regression analysis was conducted using a stepwise method (with a significance level of p < 0.05 for entry and p < 0.10 for removal) with the WSAS, DASS-21, and PCL-C scales as the dependent variables. The regression equation incorporated participant characteristics as independent variables. In cases where DASS-21 were the dependent, variables PCL-C scale was also included in the model and in cases where WSAS was the dependent variable, DASS-21 total score and PCL-C scale were also included in the model. From the linear regression analyses, adjusted regression coefficients (β) along with their corresponding standard errors (SE) and standardized regression coefficients (b) were calculated. Internal consistency reliability was assessed using Cronbach’s α coefficient, with scales having reliabilities equal to or greater than 0.70 considered acceptable. All reported p-values are two- tailed, and statistical significance was set at p < 0.05.
RESULTS AND DISCUSSION
Sample Characteristics
The sample size for this study consisted of 757 participants, with 51% being male. The average age of the participants was 52.2 years,with a standard deviation of 14.5 years. The characteristics of the participants are presented in Table 1.
Table 1: Sample Demographic Characteristics
|
|
Ν (%) |
Gender |
Men |
386 (51) |
|
Women |
369 (48.7) |
|
Nonbinary |
2 (0.3) |
Age (years) |
|
52.2 (14.5) |
Educational level |
Primary school |
3 (0.4) |
|
Middle school |
10 (1.3) |
|
High school |
126 (16.6) |
|
2-year college |
69 (9.1) |
|
College |
32 (4.2) |
|
University |
301 (39.8) |
|
MSc |
182 (24) |
|
PhD |
34 (4.5) |
Employed |
|
462 (61.0) |
Annual family income |
<10,000 euro |
133 (18.1) |
|
10,000-19,999 |
256 (34.8) |
|
20,000 - 49,999 |
300 (40.8) |
|
50,000 - 69,999 |
34 (4.6) |
|
70,000 - 99,999 |
7 (1) |
|
100,000+ |
5 (0.7) |
Many of the participants were university alumni (39.8%), employed (61%), and had an annual family income ranging from 20,000 to 49,999 euros (40.8%). Additionally, 64.6% of the sample lived in an area at high risk for fire, flood, and earthquake, while 40.0% resided with children and 32.6% had pets [Table 2].
Table 2: Sample Socioeconomic Characteristics
|
|
Ν (%) |
Belong in a group: |
No owning a bank account |
21 (2.8) |
|
National and Racial Minority |
3 (0.4) |
|
Disabled |
29 (3.8) |
|
immigrant |
4 (0.5) |
|
Refugee |
2 (0.3) |
|
Homeless |
2 (0.3) |
|
Mentally ill |
22 (2.9) |
|
LGBTQ+ |
24 (3.2) |
|
Ex-convict |
1 (0.1) |
Working status |
Remotely |
36 (6.6) |
|
In house |
375 (69.2) |
|
Hybrid |
131 (24.2) |
Living in territory |
|
|
|
Of high risk for fire |
182 (24) |
|
Of high risk for flood |
72 (9.5) |
|
Of high risk for earthquake |
375 (49.5) |
|
None of the above |
268 (35.4) |
Years living in present area of residence, mean (SD) |
|
28.5 (18.5) |
Type of residence |
House |
312 (41.2) |
|
Apartment |
444 (58.7) |
|
Motor home |
1 (0.1) |
Owning the residence |
|
580 (76.6) |
Living with: |
|
|
|
Disabled |
56 (7.4) |
|
Children |
303 (40) |
|
Elderly |
203 (26.8) |
|
Pets |
247 (32.6) |
Volunteer in one or more crises rescue groups? |
|
38 (5) |
Location |
Home |
509 (67.3) |
|
Vacation house |
72 (9.5) |
|
Work |
131 (17.3) |
|
other |
44 (5.8) |
Event |
Fire |
164 (21.7) |
|
Flood |
80 (10.6) |
|
Earthquake |
491 (65) |
|
Other |
20 (2.6) |
Evacuation decided |
Yes, and I followed it |
81 (10.8) |
|
Yes, but I chose to stay behind |
18 (2.4) |
|
No but I evacuated |
325 (43.5) |
|
No and I chose to stay behind |
323 (43.2) |
Basic mean of transportation |
Car |
495 (65.4) |
|
Motorcycle |
41 (5.4) |
|
Bike |
12 (1.6) |
|
Public transportation |
124 (16.4) |
|
Walking |
56 (7.4) |
|
Taxi |
16 (2.1) |
|
Uber |
1 (0.1) |
|
Working bus |
1 (0.1) |
|
Other |
5 (0.7) |
|
Do not want to answer |
6 (0.8) |
Get online at home |
|
752 (99.3) |
Own mobile |
Yes, smartphone |
709 (93.7) |
|
Yes, non-smartphone |
45 (5.9) |
|
No |
3 (0.4) |
Use GPS for finding route |
Yes, and I follow the suggested one |
417 (55.1) |
|
Yes, but rarely I follow the suggested one |
131 (17.3) |
|
No |
209 (27.6) |
Have you experienced a fire |
|
258 (34.1) |
Have you experienced a flood |
|
184 (24.3) |
|
|
|
Have you experienced an earthquake |
728 (96,2) |
Only 5% of the participants were volunteers in one or more crises rescue groups. Moreover, 65% of the participants had experienced an earthquake recently, and 21.7% a fire. Most of the sample was at home during the last dangerous event (67.3%). Furthermore, 43.5% of the participants proceeded with evacuation even though it was not decided in general, and 43.2% decided to stay behind. Almost all participants could get online at home (99.3%), 93.7% owned a smartphone, and 55.1% used GPS and followed the suggested route. Additionally, 96.2% of the sample had experienced an earthquake, 34.1% a fire, and 24.3% a flood.
Correlation Analysis
Table 3 presents the descriptive statistics for the WSAS, DASS-21, and PCL-C scales.
Table 3: Descriptives of WSAS, DASS-21 and PCL-C scales and their intercorrelations
Scale (Cronbach's a) |
Mean (SD) |
Pearson’s correlation coefficients |
||||||
1 |
2 |
3 |
4 |
5 |
6 |
|||
1 |
WSAS score (.94) |
6.7 (9.78) |
1.00 |
0.69 |
0.62 |
0.63 |
0.69 |
0.66 |
2 |
Depression (.92) |
4.75 (5.14) |
|
1.00 |
0.76 |
0.86 |
0.94 |
0.70 |
3 |
Anxiety (.89) |
2.67 (4) |
|
|
1.00 |
0.79 |
0.90 |
0.67 |
4 |
Stress (.91) |
5.2 (4.99) |
|
|
|
1.00 |
0.95 |
0.72 |
5 |
Total DASS-21 score (.96) |
12.6 (13.14) |
|
|
|
|
1.00 |
0.75 |
6 |
PCL-C score (.94) |
35.24 (14.87) |
|
|
|
|
|
1.00 |
All scales demonstrated satisfactory reliability coefficients above 0.7, indicating acceptable reliability. A higher severity of PTSD symptoms was significantly correlated with increased levels of depression (r=0.70; p<0.001), anxiety (r=0.67; p<0.001), and stress symptoms (r=0.72; p<0.001), as well as with a higher total DASS-21 score (r=0.75; p<0.001). Furthermore, a higher PCL-C score (r=0.66; p<0.001) and higher scores in the DASS subscales (ranging from 0.62 to 0.69; p<0.001) and total score (r=0.69; p<0.001) were significantly associated with greater levels of functional impairment as indicated by the WSAS score.
Figure 1 displays the levels of depression, anxiety,
Figure 1: Depression, anxiety, and stress levels from DASS-21 scale.
and stress symptoms among the participants. Most participants reported normal levels of these symptoms, with percentages of 60.3% for depression, 75.3% for anxiety, and 74.3% for stress. Additionally, Figure 2 indicates that 29.2% of the sample experienced moderate to moderately high severity symptoms of PTSD, while 26.2% had symptoms classified as high severity.
Figure 2: PTSD severity symptoms from PCL-C scale.
Also, greater worries during the last dangerous condition that participants were exposed to were significantly associated with greater severity in their PTSD symptoms [Table 4].
Table 4: Spearman correlation coefficients between PCL-C score and participants' attitudes related to their most recent exposure to hazardous situations
|
PCL-C score |
I was convinced there would be massive infrastructure damage |
.17 |
I was afraid that I would be injured or die |
.29 |
I wasn't worried |
-.17 |
I wasn't worried about finding fuel |
-.12 |
I was worried about my place of residence (after the event ended) |
.21 |
I risked moving to save a man |
.16 |
I risked moving to save a pet |
.19 |
I risked moving to save myself |
.26 |
I risked moving to save productive animals (e.g., herd) |
.15 |
[Note. All correlation coefficients were significant at p<0.001]
Regression Analysis
Table 5 presents the results of multiple linear regression analysis,
Table 5: Results of multiple linear regression using the PCL-C score as the dependent variable and participants' characteristics as independent variables, employing a stepwise method
|
β+ |
SE++ |
b‡ |
P |
Age |
-0.23 |
0.04 |
-0.22 |
<0.001 |
Gender (Women vs Men) |
3.56 |
1.07 |
0.12 |
0.001 |
Annual family income |
-1.96 |
0.61 |
-0.12 |
0.001 |
Living with a disabled person (yes vs no) |
7.23 |
2.10 |
0.12 |
0.001 |
Living with elderly (yes vs no) |
2.99 |
1.26 |
0.09 |
0.018 |
indicating significant associations between the PCL-C score and various factors including age, gender, annual family income, and living with disabled person or elderly. Specifically, older age and higher annual family income were found to be significantly associated with lower levels of PTSD symptom severity. Conversely, women and participants living with disabled person or elderly had significantly greater severity of PTSD symptoms.
Table 6 displays the results of multiple linear regression analyses with DASS-21 scores as the dependent variables.
Table 6: Results of multiple linear regression using DASS-21 scores as the dependent variables, participants' characteristics, and PCL-C score as the independent variables, employing a stepwise method
Dependent variable |
Independent variable |
β+ |
SE++ |
b‡ |
P |
Depression |
Age |
-0.03 |
0.01 |
-0.07 |
0.010 |
|
Annual family income |
-0.39 |
0.16 |
-0.07 |
0.015 |
|
PCL-C score |
0.23 |
0.01 |
0.67 |
<0.001 |
Anxiety |
Annual family income |
-0.37 |
0.12 |
-0.09 |
0.003 |
|
PCL-C score |
0.17 |
0.01 |
0.66 |
<0.001 |
Stress |
Annual family income |
-0.36 |
0.15 |
-0.07 |
0.014 |
|
Event |
|
|
|
|
|
Fire vs Earthquake |
0.79 |
0.32 |
0.07 |
0.016 |
|
Flood vs Earthquake |
0.52 |
0.44 |
0.03 |
0.236 |
|
Other vs Earthquake |
-0.03 |
0.81 |
0.00 |
0.971 |
|
PCL-C score |
0.23 |
0.01 |
0.70 |
<0.001 |
Total DASS-21 score |
Age |
-0.06 |
0.02 |
-0.07 |
0.013 |
|
Annual family income |
-1.11 |
0.38 |
-0.08 |
0.004 |
|
PCL-C score |
0.63 |
0.02 |
0.72 |
<0.001 |
+ regression coefficient; ++Standard Error; ‡Standardized regression coefficient
The findings indicate that higher annual family income was significantly associated with lower levels of depression, anxiety, and stress symptoms. Additionally, a higher PCL-C score was significantly associated with greater scores on the DASS-21 scales. Age showed a negative association with depression and the total DASS-21 scores. Furthermore, participants who had recently experienced a fire exhibited significantly higher levels of stress symptoms compared to those who had not experienced such an event.
Table 7 presents the results indicating that participants who resided with a disabled person and those living in areas at high risk for fire, flood, and earthquake exhibited significantly higher levels of impairment in work, social, and related domains.
Table 7: Results of multiple linear regression analysis using WSAS score as the dependent variable, participants' characteristics, and their DASS-21 and PCL-C scores as independent variables, employing a stepwise method
|
β+ |
SE++ |
b‡ |
P |
Living with a disabled person (yes vs no) |
4.54 |
1.01 |
0.12 |
<0.001 |
Living in a territory of high risk for fire, flood, and earthquake (yes vs no) |
1.12 |
0.53 |
0.05 |
0.037 |
Total DASS-21 score |
0.31 |
0.03 |
0.42 |
<0.001 |
PCL-C score |
0.22 |
0.03 |
0.34 |
<0.001 |
+regression coefficient; ++Standard Error; ‡Standardized regression coefficient
Lastly, there was a significant association between higher DASS-21 scores, higher PCL-C scores, and increased impairment in work, social, and related domains.
CONCLUSION
Natural disasters, ranging from floods and wildfires to earthquakes, can inflict lasting harm on individuals’ mental fortitude and the cohesion of communities. This underscores the pivotal role of disaster preparedness and response measures in steering outcomes toward a more positive trajectory in the aftermath of such events.
Our research underscores the importance of comprehending the nexus between natural hazards and psychological health. This understanding forms the bedrock for crafting effective interventions and policies to support affected populations. This study serves as a cornerstone in informing the development of strategies geared towards aiding individuals and communities in the aftermath of these traumatic events.
Comprehending the challenges citizens face in the aftermath of a disaster necessitates recognizing the multifaceted consequences encompassing the physical, social, and environmental domains. In a pivotal study by Kemp et al. [41], the psychological aftermath of a major earthquake in September 2010 was meticulously examined. Employing the DASS-21, the researchers gauged self- reported sleeplessness, cognitive impairment, stress, depression, and anxiety among individuals in the public who had directly experienced the seismic event.
The outcomes of Kemp et al.’s investigation parallel the current study’s findings, particularly about the correlation observed among stress, depression, and anxiety. A corroborative study by Trip et al. [42], also utilizing the DASS-21, unveiled that participant exposed to significant earthquake events reported elevated levels of stress, anxiety, and depression compared to their non-exposed counterparts. This concurrent discovery further fortifies the assertions put forth by the present study.
Cumulatively, these findings underscore the profound impact of natural hazards, emphasizing that their repercussions extend beyond the physical realm, significantly influencing the psychological well-being of individuals. According to the Te Rau Hinengaro: The New Zealand Mental Health Survey [43], the estimated lifetime prevalence of PTSD is approximately 6.0 per cent [43]. Notably, Trip et al. [42], study revealed rates of PTSD that surpassed the normal ranges for each scale, a deviation that could be attributed to the cumulative effect of multiple crisis events.
Despite the potential for heightened clinical impact,Trip’s study identified various stabilizing factors that might have mitigated the overall consequences of the disaster. It is internationally acknowledged, for instance, that impaired health is often associated with dissatisfaction with social [44]. However, interestingly, this factor did not manifest in Trip’s findings. In the present study, PCL scores for PTSD symptoms were slightly above normal ranges, indicating a nuanced psychological response to the disaster.
Batniji, Van Ommeren, and Saraceno [45], argued that adopting a social and mental health model, focusing on community cohesion in the aftermath of disasters is likely to enhance access to information, preserve daily routines, facilitate a quicker return to normal activities, and increase participation in collective efforts. They contend this approach plays a pivotal role in restoring and enhancing pre-disaster functioning levels within the affected community.
The WSAS analysis revealed that relationship status did not exhibit a significant association, consistent with the findings reported in Trip et al.’s study [42]. This suggests that individuals who could sustain their regular routines and activities despite the substantial disruptions caused by the disaster may have cultivated a forward-looking perspective [46]. Such a perspective encompasses a range of elements, including recognizing necessary steps for moving forward, adeptly planning, and setting goals, and practicing self-regulation while taking agency over one’s actions and decisions [47].
Additionally, this forward-looking perspective is closely related to the concept of hardiness within the broader construct of resilience. Hardiness encompasses the capacity to perceive and interpret stressful events in a manner that promotes personal growth and adaptability [6]. It involves recognizing one’s skills and resources, which can contribute to a sense of confidence and efficacy in navigating challenges and adversity.
Recognizing the multifaceted challenges that unfold in the aftermath of a disaster is paramount, as it introduces an array of stressors that, if unaddressed, can contribute to heightened levels of mental health disturbances. However, professional support, including mental health services and robust community and familial support systems, emerges as a critical mitigating factor [48,49]. These support structures furnish individuals with the essential resources, guidance, and assistance needed to navigate the psychological and emotional repercussions of the disaster. The findings above underscore the pivotal role of preparedness initiatives in bolstering psychological resilience and well-being when confronted with potential disasters. Future research endeavors should delve into the specific factors contributing to practical disaster preparedness, paving the way for targeted interventions to foster resilience and alleviate the psychological consequences of disasters.
While this study introduces novel research questions and insights, it is imperative to acknowledge its inherent limitations: The cross-sectional design and reliance on self-report measures render the study susceptible to biases. The potential for participants to provide inaccurate information regarding their behaviour and views introduces a risk of inaccuracies in the results. The absence of longitudinal analysis or experimental manipulation of variables precludes the establishment of causal relationships between the measures.
The anonymous online access to the questionnaire raises concerns about the verification of inclusion criteria, allowing participants to access it multiple times potentially. This possibility cannot be ruled out despite the survey’s length potentially deterring such behaviour. The recruitment method employed in this study may influence the results, limiting the generalizability of findings to specific groups characterized by factors such as ethnicity, age, and comorbidities.
In synopsis, this cross-sectional analysis illuminates the adverse effects of natural hazards on the psychological well- being, resilience outcomes, and disaster preparedness of adults in Greece. The results reveal a substantial correlation between the severity of PTSD symptoms and heightened levels of depression, anxiety, and stress symptoms, coupled with increased impairment in work, social, and related domains. Elevated scores on the PCL-C and DASS-21 scales are similarly linked to augmented impairment in these areas. Notably, the multiple linear regression analysis identifies age, gender, annual family income, and living with a disabled or elderly individual as significant predictors of PCL-C scores. Furthermore, a higher annual family income is associated with lower levels of depression, anxiety, and stress symptoms.
IMPLICATION
These findings’ significance extends beyond Greece’s immediate context, resonating with the recognized or potential threats posed by natural disasters globally. The practical implications underscore a critical need for community-based behaviour change interventions promoting resilience and disaster preparedness. These interventions should emphasize enhancing psychological resilience, fostering psychological flexibility, and cultivating coping strategies to mitigate the adverse impact of natural hazards on individuals and communities.
Moreover, targeted interventions should prioritize vulnerable groups, including individuals with disabilities, the elderly, and those residing in high-risk areas, to enhance their disaster preparedness and minimize the impact of natural hazards on their lives. Policymakers and public health officials can leverage these findings to formulate evidence-based strategies, fostering the psychological health and resilience of individuals affected by natural hazards in Greece and analogous contexts.
ACKNOWLEDGEMENTS
The authors would like to extend their sincere gratitude to the participants for their valuable time and cooperation. They would also like to express their appreciation to The American College of Greece for their support and approval of the ethical procedures. This research is funded under the Greek national project «Development of the “Coastal Environmental Observatory and Crisis Management in Island Areas” Infrastructure (AEGIS+)” which is conducted at the University of the Aegean. The first author (KK) would also like to acknowledge the financial support provided for the following projects: ‘Collaborative, Multi- modal and Agile Professional Cybersecurity Training Program for a Skilled Workforce In the European Digital Single Market and Industries’ (CyberSecPro) project, which has received funding from the European Union’s Digital Europe Programme (DEP) programme under grant agreement No 101083594. The ‘advaNced cybErsecurity awaReness ecOsystem for SMEs’ (NERO) project, which has received funding from the European Union’s DEP programme under grant agreement No 101127411. The views expressed in this paper represent only the views of the authors and not of the European Commission or the partners in the above-mentioned projects.
REFERENCES
- Global increase in major tropical cyclone exceedance probability over the past four decades. PNAS. 2020; 117: 11975-11980.
- Schilpzand. The impact of natural disasters on social capital: An analysis of ingroup and outgroup trust. Int J Disaster Risk Reduct. 2023; 95: 103860.
- Social and Economic Impacts of Natural Disasters: International Geology Review. 2010; 43: 401-405.
- Albrecht F. Natural hazard events and social capital: the social impact of natural disasters. Disasters. 2018; 42: 336-360.
- GAR2022: Our World at Risk (GAR). UNDRR. 2024.
- Bonanno GA. Loss, Trauma, Human Resilience: Have We Underestimated the Human Capacity to Thrive After Extremely Aversive Events?. Am Psychol. 2004; 59: 20-28.
- Gaume E, Bain V , Bernardara P, Newinger O, Barbuc M, Bateman A, et al. A compilation of data on European flash floods. J Hydrol. 2009; 367: 70-78.
- Llasat MC, Llasat-Botija M, Prat MA, Porcú F, Price C, Mugnai A, et al. High-impact floods and flash floods in Mediterranean countries: the FLASH preliminary database. Adv Geosci. 2010; 23: 47-55.
- Drobinski P, Ducrocq V, Alpert P, Anagnostou E, Béranger K, Borga M, et al. HyMeX: A 10-Year Multidisciplinary Program on the Mediterranean Water Cycle. Bull Am Meteorol Soc. 2014; 95: 1063- 1082.
- Diakakis M, Nikolopoulos E. I, Efthimiou G. The impact of weather conditions and human factors on wildfires in Attica, Greece. Nat Hazards Earth Syst Sci. 2015; 15: 465-477.
- Loukas, Vasiliades L, Dalezios N. R. Flood spatial estimation in ungauged basins applying a GIS-based distributed model. Hydrol Sci J. 2003; 47: 173-189.
- Diakakis M, Spyrou C, Chronopoulos G. Temporal and spatial analysis of flash floods in Greece over the last 50 years. Nat Hazards Earth Syst Sci. 2017; 17: 2125-2141.
- Papagiannaki K, Diakakis M, Deligiannakis G. Landslide hazard assessment in Greece: A case study from the island of Cephalonia. Geoenvironmental Disasters. 2017; 4: 1-13.
- Diakakis M, Spyrou C, Ilia L, Katsetsiadou K. Post-fire flash floods and debris flows in Greece: Analysis of the flood risk perception and evacuation behavior. Int J Disaster Risk Reduct. 2020; 50: 101813.
- Fillipidis F, Papanikolaou V, Papanikolaou D, Chletsos M, Myrovali G, Skordoulis M. Assessing the effect of evacuation strategies on community resilience. A case study of Mandra, Greece. Int J Disaster Risk Reduct. 2021; 67: 102469.
- North CS, Pfefferbaum B, Tivis L. Psychiatric disorders among survivors of the Oklahoma City bombing. JAMA Psychiatry. 2013; 290: 612-618.
- Lai BS, La Greca AM, Auslander BA, Short MB. Children’s symptoms of posttraumatic stress and depression after a natural disaster: Comorbidity and risk factors. J Affect Disord. 2013; 146: 71-78.
- Brymer M, Jacobs A, Layne C, Pynoos R, Ruzek J, Steinberg A, et al. Psychological First Aid Field Operations Guide. 2nd Edition. NCTSN. 2006.
- Luthar SS, Cicchetti D, Becker B. The Construct of Resilience: A Critical Evaluation and Guidelines for Future Work. Child Dev. 2000; 71: 543-562.
- Connor KM. Assessment of resilience in the aftermath of trauma. J Clin Psychiatry. 2006; 67: 46-49.
- Masten S, Hubbard JJ, Gest SD, Tellegen A, Garmezy N, Ramirez M. Competence in the context of adversity: Pathways to resilience and maladaptation from childhood to late adolescence. Dev Psychopathol. 1999; 11: 143-169.
- Huggard P. Caring for the carers: The emotional effects of disasters on health care professionals. Australas J Disaster Trauma Stud. 2011; 2: 60-62.
- Herrman H. Promoting Mental Health and Resilience after a Disaster. J Exp Clin Med. 2012; 4: 82-87.
- Gutierrez D, Hollister D, Beninati A. Hurricane madness: Teaching, learning and the importance of flexibility in the wake of disaster. J Coll Teach Learn (TLC). 2005; 2: 2.
- Ahern M, Kovats RS, Wilkinson P, Few R. Impact of community-based flood risk management on resilience and mental health: A mixed- methods evaluation. J Environ Psychol. 2021; 5: 101638.
- Lin Y, Wu T, Chou FH, Huang Y, Yen CF, Wang PW. Individual-level disaster preparedness and psychological distress following a natural disaster: A population-based study. J Affect Disord. 2012; 294: 114- 120.
- von Elm E, Altman DG, Egger M, Pocock SJ, Gøtzsche PC, Vandenbroucke JP. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. The Lancet. 2007; 370: 9596: 1453- 1457.
- Billingham SA, Whitehead AL, Julious SA. An audit of sample sizes for pilot and feasibility trials being undertaken in the United Kingdom registered in the United Kingdom Clinical Research Network database. BMC Med Res Methodol. 2013; 13: 104.
- Chilcot J, Rona M, Micol A, Larissa H, Federica P, Hector H, et al. Psychosocial and Clinical Correlates of Fatigue in Haemodialysis Patients: the Importance of Patients’ Illness Cognitions and Behaviours. Int J Behav Med. 2016; 23: 271-281.
- Muller K. Statistical Power Analysis for the Behavioral Sciences. Technometrics. 2010; 31: 499-500.
- Antony MM, Bieling PJ, Cox BJ, Enns MW, Swinson RP. Psychometric properties of the 42-item and 21-item versions of the Depression Anxiety Stress Scales in clinical groups and a community sample. Psychol. Assess.1998; 10: 176-181.
- Lovibond PF, Lovibond SH. The structure of negative emotional states: Comparison of the Depression Anxiety Stress Scales (DASS) with the Beck Depression and Anxiety Inventories. Behav Res Ther. 1995; 33: 335-343.
- Crawford JR, Henry JD. The Depression Anxiety Stress Scales (DASS): Normative data and latent structure in a large non-clinical sample. Br J Clin Psychol. 2003; 42: 111-131.
- Association P. Diagnostic and Statistical Manual of Mental Disorders. Text Revis. 2000.
- PTSD: National Center for PTSD.
- Blanchard EB, Jones-Alexander J, Buckley TC, Forneris CA. Psychometric properties of the PTSD checklist (PCL). Behav Res Ther. 1996; 34: 669-673.
- Ruggiero KJ, Ben KD, Scotti JR, Rabalais AE. Psychometric Properties of the PTSD Checklist—Civilian Version. J Trauma Stress. 2003 16: 495–502.
- Weathers FW, Litz BT, Herman DS, Huska JA, Keane TM. The PTSD Checklist (PCL): Reliability, validity, and diagnostic utility. Ann Conv Int Soc Traumat Stre Stud. San Antonio, Texas. 1993.
- Mundt JC, Marks IM, Shear MK, Greist JM. The Work and Social Adjustment Scale: a simple measure of impairment in functioning. Br J Psychiatry. 2002; 180: 461-464.
- Bell C, Carter F, Boden J, Wilkinson T, McKenzie J, Ali A. Psychological impact of the Canterbury earthquakes on university staff. N Z Med J. 2016; 129: 18-28.
- Kemp S, Helton WS, Richardson JJ, Blampied NM, Grimshaw M. Sleeplessness, stress, cognitive disruption and academic performance following the September 4, 2010, Christchurch earthquake. Australas J Disaster Trauma Stud. 2011; 2, 11-18.
- Trip H, Tabakakis K, Maskill V, Richardson S, Dolan B, Josland H, et al. Psychological health and resilience: the impact of significant earthquake events on tertiary level professional students. A cross- sectional study. Contemp. Nurse. 2018; 54: 319-332.
- Oakley Browne MA. Lifetime prevalence and lifetime risk of DSM-IV disorders. Te Rau Hinengaro N Z Ment Health Surv. Wellingt Minist Health. 2006.
- Wahlström L, Michélsen H, Schulman A, Backheden M. Support, opinion of support and psychological health among survivors of a natural disaster. Sage J. 2013; 59.
- Batniji R, Van Ommeren M, Saraceno B. Mental and social health in disasters: Relating qualitative social science research and the Sphere standard. Soc Sci Med.2006; 62: 1853-1864.
- Seginer R. Future orientation in times of threat and challenge: How resilient adolescents construct their future. Int J Behav Dev. 2008; 32: 272-282.
- Holman EA, Silver RC. Future-Oriented Thinking and Adjustment in a Nationwide Longitudinal Study Following the September 11th Terrorist Attacks. Motiv Emot. 2005; 29: 385-406.
- Norris FH, Friedman MJ, Watson PJ, Byrne CM, Diaz E, Kaniasty K. 60,000 Disaster Victims Speak: Part I. An Empirical Review of the Empirical Literature, 1981-2001. Psychiatry. 2002; 65: 207-239.
- Pfefferbaum B, Newman E, Nelson SD, Nitiéma P, Pfefferbaum RL, Rahman A. Disaster Media Coverage and Psychological Outcomes: Descriptive Findings in the Extant Research. Curr Psychiatry Rep. 2014; 16: 464.