Symptoms of Posttraumatic Stress, Depression and Anxiety among Youths Exposed to a Massive Fire Disaster in Greece

___________________________________________________________________________________________ ABSTRACT Aim: To evaluate the prevalence of symptoms of posttraumatic stress disorder (PTSD), depression and anxiety among children and adolescents exposed to the 2007 fire disaster in Greece along with the relationships of these symptoms with disaster-related stressors and sociodemographic characteristics four months after the fire. Methodology: A sample of 343 youths aged 9-18 years from schools in an area severely affected by the fire completed self-reported questionnaires. The Children’s Post-Traumatic Stress Disorder-Reaction Index-Revised (CPTSD-RI-R), the Children’s Depression Inventory (CDI) and the Screen for Child Anxiety Related Emotional Disorders (SCARED) were used to assess relative symptoms, while fire-related stressors were also examined in respondents. symptoms. Conclusion: A significant proportion of children and adolescents exposed to 2007 Greek forest fires experienced mental health problems four months after the fire. Different types of stressors were associated with PTSD, depressive and anxiety symptoms. These findings highlight the importance of preventive and treatment mental health services in the aftermath of similar disasters.


INTRODUCTION
Massive forest fires broke out in several areas across Greece throughout the summer of 2007, which was the worst fire season on record in the past 50 years. In total 84 people lost their lives and the fire destroyed 1,000 houses, 1,100 other buildings and 2,700 square kilometers (670,000 acres) of forest, olive groves and farmland (Maditinos and Vassiliadis, 2011). People in fire-affected areas were exposed to several stressors, such as risk of death, loss of property, housing adversity and community disruption.
The experience of a major disaster has been shown to provoke a range of stress responses, maladaptive reactions, mental health symptoms and diagnosable psychiatric disorders among children and adolescents (Bromet and Dew, 1995;Norris et al., 2002). However, relatively little research has been published regarding the impact of a fire disaster on youths' mental health (Dyregrov et al., 2003;Jones and Ribbe, 1991;Jones et al., 1994;Jones et al., 2002;Kim et al., 2009;Lundin and Jansson, 2007;Maja et al., 2008;March et al., 1997;McDermott and Palmer, 2002;Reijneveld et al., 2003;Yelland et al., 2010). In the first two months to over a year after a fire, the prevalence of posttraumatic stress disorder (PTSD) among children and adolescents has been estimated at 11.9% to 27% (Dyregrov et al., 2003;Kim et al., 2009;March et al., 1997;Yelland et al., 2010). The prevalence rates of depression and internalizing problems have been shown to range from 18.5% to 27% four to seven months post-fire (Dyregrov et al., 2003;Reijneveld et al., 2003). Exposure to disasterrelated stressors, such as property loss, physical adversity, housing adversity, physical illness or injury and life-threatening experience, has been shown to be a significant predictor of mental health problems among children and adolescents post-disaster (Eksi and Braun, 2009;Kolaitis et al., 2003;McLaughlin et al., 2009;Yelland et al., 2010). Identification of factors that distinguish children who present serious symptoms from those who experience rather mild problems seems to be of great importance for developing post-disaster interventions. Female sex, younger age, and parent psychopathology have been also identified as predictors of more serious symptoms in children and adolescents after natural disasters (McLaughlin et al., 2009;Pynoos et al., 1993;Shannon et al., 1994). These symptoms may last for several months or years, interfering with children's and adolescents' growth and development. This, in turn, may lead to negative long-term health outcomes and hamper normal development (Lundin and Jansson, 2007;McDermott and Palmer, 2002;McFarlane et al., 1987).
To date, there is no published report examining the effects of 2007 Greek forest fires on child and adolescent mental health. The present study was an effort to examine the prevalence of youth mental health symptoms in an affected area given the possible large extent of the public health issue and the relative lack of preventive and treatment resources in the affected area. The study also evaluates the association between exposure to various disaster-related stressors and mental health problems. Sociodemographic factors are examined too as predictors of youth symptomatology.

Participants and Procedure
The study was conducted four months after the 2007 fire disaster in Peloponnese, Greece. A sample of 343 youths (156 males, 187 females) aged 9-11 (N = 60), 12-13 (N = 90), 14-15 (N = 87), 16-20 (N = 106) volunteered to take part from schools in Lakonia of Peloponnese, Greece. Youths filled in the questionnaire at school. Inclusion criteria for the youths were to be able to read and complete the questionnaires themselves. After approval by relevant ethics committees, parents provided written informed consent for youth participation.

PTSD symptoms
The Children's Post-Traumatic Stress Disorder-Reaction Index-Revised (CPTSD-RI-R; Steinberg et al., 2004) was used to screen for serious PTSD symptoms. The CPTSD-RI-R is a 20-item questionnaire assessing posttraumatic stress reactions in children and adolescents. Items referenced the recent fire as the traumatic event. Youths indicated how often they experienced each symptom (0 = none of the time; 4 = most of the time). The measure yields a total severity score, with scores falling within the categories of doubtful (0-11), mild (12-24), moderate (25-39), severe (40-59), or very severe (60+). The respondents who received a CPTSD-RI-R score of 12 or more were classified as having high levels of PTSD symptoms. The PTSD-RI has been widely used and has good convergent validity, test-retest reliability, and internal consistency (Steinberg et al., 2004). In this study, internal consistency was 0.88.

Depressive symptoms
The Children's Depression Inventory (CDI; Kovacs, 1985) was used to screen for serious depressive symptoms. There are 27 items quantifying symptoms such as depressed mood, hedonic capacity, vegetative functions, self-evaluation and interpersonal behaviours. It covers the consequences of depression as they relate to children and functioning in school and with peers. For each item the child has three possible answers; 0 indicating the absence of symptoms, 1 the mild symptoms, and 2 the definite symptoms. The total score ranges from 0 to 54. The respondents who received a CDI score of 15 or more were classified as having high levels of depressive symptoms (Giannakopoulos et al., 2009a). Previous research has shown that the CDI is adequately reliable and valid with respect to depressive symptoms (Myers and Winters, 2002).

Anxiety symptoms
The Screen for Child Anxiety Related Emotional Disorders (SCARED; Birmaher et al., 1997) was used to screen for serious anxiety problems. The SCARED is a self-report scale for children suffering from anxiety disorders. It contains 38 items, describing different emotions and behaviors. For each item, the child is asked to report the frequency of that emotion or behavior, over a 3-point scale (0 = never, 2 = frequently). Five scores can be obtained, including somatic problems, general anxiety, separation anxiety, social phobia, and school phobia. Summing all items reveals a total anxiety score. The respondents who received a SCARED score of 25 or more were classified as having high levels of anxiety symptoms. This scale has demonstrated acceptable internal and test-retest consistency (Birmaher et al., 1997).

Disaster-related stressors
The respondents were asked 20 questions regarding exposure to fire-related stressors. Seven categories of stressors were sufficiently common to be included in the analysis: housing adversity (experiencing moves post-fire), staying without both the parents, separation from parents post-fire, property loss, worry for a loved one during fire, injury of a loved one, and life-threatening experience of a loved one during fire.

Sociodemographic characteristics
Age, sex and number of siblings were examined in relation to PTSD, depressive and anxiety symptoms at the time of interview.

Data Analysis
Continuous variables are presented with mean ± standard deviation while categorical variables are presented with absolute and relative frequencies. For comparisons of proportions, chi-square and Fisher's exact tests were used. Pearson correlations coefficients were used to explore the association of study scales. P-values for trends in the severity of disorders were computed for order categories. In order to find the factors independently associated with depression or PTS symptoms stepwise multiple logistic regression analysis (p for removal was set at 0.1 and p for entry was set at 0.05) was performed. Adjusted odds ratios (Ors) with 95% confidence intervals were computed from the results of logistic regression analyses. All p-values reported are two-tailed. Statistical significance was set at 0.05 and analyses were conducted using SPSS statistical software (version 13.0).

Sociodemographic Characteristics and Disaster-Related Stressors
Sociodemographic characteristics are presented in Table 1. Approximately 5% of the participants had stayed without both their parents after the fire disaster, while 4.4% of the youths in the sample had been separated from their parents post-fire. Almost one-third (30%) of the sample had been worried for a family member or a close friend during fire and 13.1% had at least one loved person who was trapped during the fire.

PTSD Symptoms
High levels of PTSD symptoms were found in 45.4% of the youths in the sample. Mild PTSD symptoms were found in 22.0 % of children and adolescents, moderate symptoms were found in 11.7%, severe symptoms in 10.8% and very severe PTSD symptoms in 0.9% of the sample. The proportion of children with mild to very severe PTS symptoms decreased by age (61.7% for those aged from 9 to 11 years, 51.1% for those aged from 12 to 13 years, 41.4% for those aged from 14 to 15 years and 34% for those aged from 16 to 20 years, p for trend < 0.001). No gender differences were found in PTSD symptoms. Staying without both parents post-fire, the number of siblings, being afraid to stay at home, being worried for a loved one during fire, injury of a loved one and life-threatening experience of a loved one during fire were significantly associated with mild to very severe PTSD symptoms in univariate analysis (Table 2). Sociodemographic characteristics are presented in Table 1. Approximately 5% of the participants had stayed without both their parents after the fire disaster, while 4.4% of the youths in the sample had been separated from their parents post-fire. Almost one-third (30%) of the sample had been worried for a family member or a close friend during fire and 13.1% had at least one loved person who was trapped during the fire.
Multiple logistic regression analysis showed that age, staying without both parents post-fire (OR=3.88; 95% CI: 1.21-2.51), being worried for a loved one during fire (OR=3.09; 95% CI: 1.81-5.26), and a life-threatening experience of a loved one during fire (OR=3.57; 95% CI: 1.63-7.81) were independently associated with increased likelihood for having PTSD symptoms. Furthermore, the larger number of siblings was associated with lower odds for having high levels of PTSD symptoms (Table 3).

Depressive Symptoms
High levels of depressive symptoms were found in 34.1% of the youths (117 cases). No age or sex differences were found for serious depressive symptoms both in univariate and multiple analyses. Separation from parents post-fire and the number of siblings were associated with serious depressive symptoms only in univariate analysis (Table 2). Multiple logistic regression analysis revealed that staying without both the parents post-fire (OR=5.57; 95% CI: 1.09-28.58), property loss (OR=4.39; 95% CI: 1.09-17.60) and injury of a loved one (OR=2.24; 95% CI: 1.04-4.82) independently increased the odds for high levels of depressive symptoms (Table 3).

Anxiety Symptoms
High levels of anxiety symptoms were found in 32.2% of the youths. The proportion of children and adolescents with high levels of anxiety symptoms increased as the magnitude of PTSD symptoms increased; high levels of anxiety symptoms were present at 0.0% in children with doubtful PTSD symptoms, 12.8% with mild PTSD symptoms, 45.2% with moderate PTSD symptoms, 75.7% with severe PTSD symptoms and at 33.3% in youths with very severe PTSD symptoms, p for trend < 0.001). Also, 18.1% of the children and adolescents (62 cases) had both high levels of depressive and anxiety symptoms. Being female was significantly associated with high levels of anxiety symptoms both in univariate analysis (Table 2) and multiple logistic regression analysis (OR=2.42, CI: 1.47 -3.99). Number of siblings was associated with high levels of anxiety symptoms only in univariate analysis (Table 2). Being worried for a loved one and injury of a loved one were associated with high levels of anxiety symptoms in univariate analysis. Furthermore, logistic regression analysis indicated that injury of a loved one was a significant predictor for high levels of anxiety symptoms (OR=3.73; CI: 1.68 -8.28).

DISCUSSION
The estimated prevalence rates of high levels of PTSD, depressive and anxiety symptoms among children and adolescents exposed to 2007 Greek forest fires four months after the disaster are approximately 45%, 34% and 32%, respectively. These are in contrast to the 4% to 9% prevalence of diagnosable mental disorders reported in epidemiological studies (Bloom and Cohen, 2007;Costello et al., 1996;Giannakopoulos et al., 2009b;Giannakopoulos et al., 2009c). The prevalence of PTSD, depressive and anxiety symptoms, has been previously reported after fire disasters with varying results. The prevalence of PTSD among children and adolescents was estimated at 11.9% to 27% two months to over a year after a fire in previous research (Dyregrov et al., 2003;Kim et al., 2009;March et al., 1997;Yelland et al., 2010). Similarly, the prevalence rates of depression and internalizing problems have been shown to range from 18.5% to 27% four to seven months post-fire (Dyregrov et al., 2003;Reijneveld et al., 2003). The estimated prevalence of PTSD, depressive and anxiety symptoms in the present sample is significantly higher, but these estimates are difficult to compare mainly because each study used different measures of psychopathology and recruited different age groups of youths. Nonetheless, the high prevalence of these symptoms among children and adolescents exposed to 2007 Greek forest fires suggests that the impact of the disaster on youth mental health was at least similar to previous fire and natural disasters elsewhere (Shannon et al., 1994;Yelland et al., 2010). Moreover, in the present study, younger age and small number of siblings is associated with more symptoms of PTSD, while female sex is associated with elevated symptoms of anxiety. These relations have been typically reported in other studies of natural disasters (Green et al., 1991;Yelland et al., 2010). However, the absence of other relations between sociodemographic factors and levels of symptoms may suggest that the intensity of trauma and distress associated with the fire increased risk for mental health problems in children and adolescents across many groups.
The study also examined exposure to the various component stressors related to the fire disaster and finds multifaceted relations between these stressors and the prevalence of PTSD, depressive and anxiety symptoms. Staying without both parents after the fire is strongly associated with more PTSD symptoms, while housing adversity and loss of property are most strongly associated with elevated depressive symptoms. Life-threatening experience of a loved one and worry for a loved one predict higher levels of PTSD symptomatology, whereas injury of a loved one is associated with high levels of depressive and anxiety symptoms. Although these findings are not surprising since exposure to disaster-related stressors has been associated with child psychopathology in previous research, the present analysis suggests that different aspects of fire disaster exposure is associated with different aspects of child psychopathology. Differentiating the effects of stressors on various types of symptoms provides some guidance for the types of treatment that may be most effective for exposed children and adolescents. For example, perception of loss seems to be critical in the development of depressive symptoms, while perception of harm to self or a loved one is most strongly associated with PTSD and anxiety-related symptomatology. Future research may enhance our understanding about the pathways that underlie the complex associations between fire-disaster stressors and youths' mental health problems.
The current results demonstrate that elevated PTSD, depressive and anxiety symptoms among the children and adolescents exposed to 2007 Greek forest fires are quite common and thus suggest a need for preventive and treatment health services among natural disaster-exposed youths. Beyond responding to economic adversity and treating physical illness or injury, policy makers and public health sector should take into account the complex cluster of mental health problems and needs which possibly burden youths and eventually remain unmet.
A number of limitations of the present study must be acknowledged. First, PTSD, depressive and anxiety symptoms were estimated using screening scales rather than a diagnostic interview. Thus, the findings could not be indicative for presence or absence of PTSD, depression and anxiety-related disorders, but merely suggest a high probability of serious levels of symptoms that need to be further evaluated through clinical assessment and validated structured diagnostic interviews. Second, data were based on youth self-reports. Although children and adolescents are considered the best informants for internalizing symptoms, future research might include multiple informants. Third, the evaluation was conducted at one time point, missing to monitor trajectories of risk and recovery over time.
Fourth, assessment of fire-related stressors was retrospective, although evidence suggests that reports of acute traumatic events are reliable. Finally, unmeasured confounders, such as family income and social support, may be responsible for the reported associations, which should be interpreted with caution.

CONCLUSION
• A significant proportion of children and adolescents exposed to 2007 Greek forest fires experienced mental health problems four months after the fire. • Different types of stressors were associated with PTSD, depressive and anxiety symptoms. • These findings highlight the importance of preventive and treatment mental health services in the aftermath of similar disasters.