This article was downloaded by: [University of Nebraska, Lincoln] On: 13 April 2015, At: 17:38 Publisher: Routledge Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK
Journal of Trauma & Dissociation Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/wjtd20
Indian Ocean Tsunami: Relationships Among Posttraumatic Stress, Posttraumatic Growth, Resource Loss, and Coping at 3 and 15 Months a
b
David N. Sattler PhD , Sawitri Assanangkornchai MD , Adam M. a
c
Moller PhD , Wiworn Kesavatana-Dohrs PhD & James M. Graham PhD
a
a
Department of Psychology , Western Washington University , Bellingham , Washington , USA b
Click for updates
Department of Psychiatry , Prince of Songkhla University , Hat Yai , Thailand c
Department of Asian Languages and Literature , University of Washington , Seattle , Washington , USA Accepted author version posted online: 10 Jan 2014.Published online: 11 Mar 2014.
To cite this article: David N. Sattler PhD , Sawitri Assanangkornchai MD , Adam M. Moller PhD , Wiworn Kesavatana-Dohrs PhD & James M. Graham PhD (2014) Indian Ocean Tsunami: Relationships Among Posttraumatic Stress, Posttraumatic Growth, Resource Loss, and Coping at 3 and 15 Months, Journal of Trauma & Dissociation, 15:2, 219-239, DOI: 10.1080/15299732.2014.869144 To link to this article: http://dx.doi.org/10.1080/15299732.2014.869144
PLEASE SCROLL DOWN FOR ARTICLE Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content.
Downloaded by [University of Nebraska, Lincoln] at 17:38 13 April 2015
This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http://www.tandfonline.com/page/termsand-conditions
Journal of Trauma & Dissociation, 15:219–239, 2014 Copyright © Taylor & Francis Group, LLC ISSN: 1529-9732 print/1529-9740 online DOI: 10.1080/15299732.2014.869144
Indian Ocean Tsunami: Relationships Among Posttraumatic Stress, Posttraumatic Growth, Resource Loss, and Coping at 3 and 15 Months
Downloaded by [University of Nebraska, Lincoln] at 17:38 13 April 2015
DAVID N. SATTLER, PhD Department of Psychology, Western Washington University, Bellingham, Washington, USA
SAWITRI ASSANANGKORNCHAI, MD Department of Psychiatry, Prince of Songkhla University, Hat Yai, Thailand
ADAM M. MOLLER, PhD Department of Psychology, Western Washington University, Bellingham, Washington, USA
WIWORN KESAVATANA-DOHRS, PhD Department of Asian Languages and Literature, University of Washington, Seattle, Washington, USA
JAMES M. GRAHAM, PhD Department of Psychology, Western Washington University, Bellingham, Washington, USA
This study examines variables associated with posttraumatic stress symptoms (PTS) and posttraumatic growth among 2 independent samples of survivors following the Indian Ocean tsunami in Khao Lak, Thailand. Participants were exposed to unprecedented horror and loss of life and property. At 3 months participants (N = 248) were living in temporary shelters, and at 15 months a second sample (N = 255) was living in homes built after the tsunami. Prior traumatic experiences, life threat, loss of personal characteristic resources and condition resources, somatic problems, and social support accounted for close to half of the variance in PTS in each sample. At 3 months, emotion-focused coping and concerns about government favoritism also contributed to PTS. At 15 months, lack of prior disaster experience and loss of energy resources also contributed to PTS. Distress was higher among participants surveyed at 3 months than among those surveyed at 15 months. Posttraumatic Received 30 November 2012; accepted 28 September 2013. Address correspondence to David N. Sattler, PhD, Department of Psychology, Western Washington University, Bellingham, WA 98225-9172. E-mail: [email protected] 219
220
D. N. Sattler et al.
growth was positively associated with social support and problem– focused coping in both samples. The findings support conservation of resources stress theory (Hobfoll, 2012) and underscore how systemic issues affect mental health. The implications of the findings are discussed, as is the educational International Tsunami Museum designed by the first author to address systemic stressors.
Downloaded by [University of Nebraska, Lincoln] at 17:38 13 April 2015
KEYWORDS posttraumatic stress, resource loss, conservation of resources, posttraumatic growth, tsunami, disaster, systemic stressor, tsunami museum
On December 26, 2004, a 9.3-magnitude earthquake off Northern Sumatra, Indonesia, generated a powerful tsunami that devastated coastal communities bordering the Indian Ocean (Patniak, Kumar, Ramu, & Prasad, 2012). The death toll—226,323 lives lost—placed the tsunami as the fourth deadliest disaster in history. Two million people were left homeless. Hundreds of thousands of persons were exposed to extreme horror and witnessed people severely injured (American Red Cross, 2009). Estimated property damage and economic losses were more than $10 billion (Nidhiprabha, 2007; United Nations Office for Disaster Risk Reduction, 2009). In the wake of the catastrophic disaster, survivors were exposed to unprecedented systemic stressors, which we define as a loss or lack of essential community resources necessary for the effective functioning of citizens and the community and that creates enduring hardships. Systemic stressors occur when essential systems that help citizens navigate life’s challenges are insufficient to meet fundamental needs for an extended period. The tsunami affected virtually every aspect of life in coastal communities, including family functioning (e.g., loss of and injury to family members, loss of home and employment), basic survival needs (e.g., food, water, shelter), and infrastructure (e.g., damaged roads, businesses, schools, sanitation services). Survivors lacked financial resources to obtain supplies or rebuild, lacked insurance coverage, and had difficulty accessing social support networks. They saw constant reminders of the disaster and feared for their safety because of new disaster threats (Bangkok Post, 2012; United Nations Thailand, 2008). Survivors endured chronic hardship for years in the aftermath of the disaster.
RESOURCE LOSS, PSYCHOLOGICAL DISTRESS, AND POSTTRAUMATIC GROWTH Conservation of resources theory (COR) provides an important framework for understanding how resource loss as a result of natural disasters affects
Downloaded by [University of Nebraska, Lincoln] at 17:38 13 April 2015
Journal of Trauma & Dissociation, 15:219–239, 2014
221
individuals. COR posits that individuals strive to acquire, maintain, and protect resources they value and use them in ways consistent with their culture. Valued resources provide meaning, safety, and security; are necessary to acquire or maintain other resources; and help buffer against the loss of other resources. Resources fall into four general categories: condition (e.g., employment), personal characteristics (e.g., optimism), energy (e.g., time for activities), and object (e.g., house, physical possessions). When valued resources are threatened or lost, COR predicts that individuals will experience stress. COR advances that individuals must invest resources (e.g., time, money) to protect against and recover from resource loss. Thus, individuals with greater resources should be less susceptible to resource loss and recover more easily, whereas those with fewer resources may be more susceptible and less able to recover lost resources (Hobfoll, 2012). Persons with limited social support networks also tend to fare worse than those with richer networks (Norris & Alegria, 2005; Smith & Freedy, 2000). Resource caravans—the concurrent development or overlap of resources—are a central concept for COR. Threat or loss to specific resources may affect other resources because of overlap and can have a potentially wide-reaching impact on resources and well-being (Hobfoll, 2012). For example, a survivor with no property insurance who loses his or her home and job may also experience threats to personal characteristic resources (e.g., self-efficacy) and condition resources (e.g., social roles). He or she also may have increased difficulty accessing his or her social support network. As a result, resource loss spirals—the continuing loss of resources in one area that results in a loss of resources in other areas—can develop (Hobfoll, 2011). Resource loss spirals create additional stressors that tax remaining resources as well as mental health and explain, in part, the association of personal characteristic resource loss and condition resource loss with stress responses such as posttraumatic stress (Hobfoll, 2011; Sattler et al., 2006). For example, Cerdá et al. (2013) found that 18 months after a hurricane in Texas, ongoing stressors such as the loss of a job, decreased income, and the inability to live in one’s home were associated with posttraumatic stress symptoms (PTS). Resource loss spirals can contribute to the development and maintenance of systemic stressors that endure for months or years. Survivors may be especially concerned about governmental favoritism—the perception that resources are being distributed inequitably during recovery or that government (or relief) agencies are favoring certain groups with resources or services, especially when recovery is prolonged (Sattler et al., 1995). Such concerns may be a significant source of stress. Few studies have examined the associations of resource loss variables and posttraumatic stress at various points during recovery. To address this issue, this study examines responses at the end of the acute stage (defined as the first 3 months) and during the chronic stage (defined as 4 or more months) of recovery (Baum, 1991). Posttraumatic stress can develop after
Downloaded by [University of Nebraska, Lincoln] at 17:38 13 April 2015
222
D. N. Sattler et al.
exposure to psychological trauma and involves reexperiencing the event through flashbacks or nightmares, avoiding stimuli associated with the event, and experiencing increased arousal (Cerdá et al., 2013). As the challenges of recovery following the tsunami continued month after month, there was great potential for resource loss spirals to compromise personal characteristic, condition, and energy resources. Prior exposure to traumatic events, somatic problems, and the degree to which survivors believed they or loved ones would lose their lives or be injured during the disaster are associated with PTS following disasters (Smith & Freedy, 2000). Survivors may experience resource gains as disaster relief organizations help communities rebuild. COR predicts that resource gains can offset losses to some degree and serve as a protective factor against negative outcomes associated with stress (Hobfoll, 2011). Posttraumatic growth theory suggests that resource gains may be viewed as a positive experience following loss due to a traumatic event. For example, survivors may report increased awareness of life priorities and positive life changes as a result of processing the event (Groleau, Calhoun, Cann, & Tedeschi, 2012). People may learn the value of preparation, develop new coping skills, gain a new appreciation of life, and experience an enhanced sense of self-efficacy (Linley & Joseph, 2004; Sattler, 2003). These positive outcomes may offset negative experiences to some degree (Groleau et al., 2012; Linley & Joseph, 2004). We examined the association among level of education and posttraumatic growth, although prior findings are mixed (Linley & Joseph, 2004). Few studies have examined COR or posttraumatic growth following disasters in countries other than the United States (Sattler et al., 2006). Coping strategies may serve as a protective or risk factor for PTS given that natural disasters threaten survivors’ sense of control, predictability, safety, and trust (Baum, 1991) and may be associated with posttraumatic growth. Problem-focused coping involves attempting to directly change or manage the situation, and emotion-focused coping involves attempting to adapt emotions in order to reduce distress (Folkman & Lazarus, 1985). Problem-focused coping may help survivors reestablish feelings of control or obtain resources and is associated with resilience and posttraumatic growth (Wadsworth, Santiago, & Einhorn, 2009).
RESEARCH QUESTIONS Based on COR and prior research, we examined the following questions with two independent groups of survivors at 3 months and 15 months after the tsunami. First, how do demographic variables, prior traumatic experiences, experiences during the tsunami, resource loss, somatic problems, social support, coping, and concerns about tsunami threats and governmental favoritism relate to PTS and posttraumatic growth at 3 and 15 months?
Journal of Trauma & Dissociation, 15:219–239, 2014
223
Second, how do the relations between these predictors and posttraumatic growth differ across the two samples? Answers to these questions may assist theory development and the manner in which intervention and recovery programs provide assistance.
METHOD
Downloaded by [University of Nebraska, Lincoln] at 17:38 13 April 2015
Participants 3 months post-tsunami. The participants were 248 persons (97 men, 151 women) living in temporary shelters (age: M = 36, SD = 11, range = 18–72 years). All were Thai. Most were married (69%), followed by single (21%), widowed (7%), and separated/divorced (3%). Almost all were Buddhist (94%), followed by Muslim (4%) and Christian (2%). More than half had less than a secondary school education (56%), followed by completion of secondary school (20%), college degree (20%), and unknown (4%). The response rate was 94%, and those who declined to participate were not asked why. 15 months post-tsunami. The participants were 255 persons (148 men, 107 women; none participated in Study 1) who had previously lived in a temporary shelter and were living in homes in four neighborhoods built after the tsunami (age: M = 38, SD = 14, range = 18–84 years). Almost all were Thai (97%), followed by Laotian (2%) and Chinese (1%). Most were married (62%), followed by single (20%), widowed (5%), separated/divorced (2%), and unknown (11%). Most were Buddhist (90%), followed by Muslim (4%), Christian (2%), other (2%), and unknown (2%). More than half had less than a secondary school education (56%), followed by completion of secondary school (26%) and college degree (18%). The response rate was 95%, and those who declined to participate were not asked why. Two assessment instruments were partially completed and not analyzed. The demographics were similar to those of the sample collected at 3 months.
Assessment Instruments We used a version of the committee approach to translate the assessment instruments into Thai (Matsumoto & van de Vijver, 2011; van de Vijver & Leung, 1997) that we successfully used previously (Sattler et al., 2002, 2006). The fourth author translated the items and the second author revised the translation. Four medical nurses in Thailand evaluated the instruments, asked participants about readability, and reported that it was clear and understandable. A cover letter introduced the study, presented information on informed consent, and indicated that responses were anonymous. The instruments were presented in the following order.
Downloaded by [University of Nebraska, Lincoln] at 17:38 13 April 2015
224
D. N. Sattler et al.
Demographics and life threat. Seven items asked for demographic information and eight items asked about life threat during the tsunami, based on Sattler et al. (2006). Participants checked their choices or wrote in a number to indicate their answers. Four items asked whether participants were concerned that they or someone they knew would lose his or her life or be seriously injured during the tsunami. Respondents used a 4-point scale (1 = not at all to 4 = quite a bit) to indicate their answers (3-month sample: α = .84; 15-month sample: α = .83). Resource loss. An 18-item scale assessed loss of personal characteristic, object, condition, and energy resources in the past 3 months (Freedy, Saladin, Kilpatrick, Resnick, & Saunders, 1994; modified by Sattler et al., 2006). Participants used a 5-point scale (0 = no loss to 4 = great loss) to indicate their answers (3-month sample: α = .87; 15-month sample: α = .95). PTS and somatic problems. The 22-item Impact of Event Scale–Revised assessed symptoms associated with posttraumatic stress disorder in the past 7 days (Weiss & Marmar, 1996; 3-month sample: α = .94; 15-month sample: α = .94). Participants also answered four items about somatic problems (e.g., headaches, muscle pain) in the past 7 days (3-month sample: α = .76; 15month sample: α = .87). They used a 5-point scale (0 = not at all to 4 = extremely) to indicate their answers. Social support. Seven Social Support Index items assessed support (McCubbin & Thompson, 1991). An example is “If I had an emergency, even people I do not know in this community would be willing to help.” Participants used a 5-point scale (1 = not at all to 5 = extremely) to indicate their answers (3-month sample: α = .83; 15-month sample: α = .87). Coping. Eight items from the Family Crisis Oriented Personal Evaluation Scale assessed problem-focused coping (3-month sample: α = .74; 15-month sample: α = .77) and emotion-focused coping (3-month sample: α = .73; 15-month sample: α = .79; McCubbin & Thompson, 1991). An example is “I face problems head-on and try to get the solutions right away.” Participants used a 5-point scale (1 = not at all to 5 = extremely) to indicate their answers. Posttraumatic growth. Fifteen items from the Posttraumatic Growth Inventory assessed growth (Tedeschi & Calhoun, 1996). Participants used a 7-point scale (1 = great decrease to 4 = no change to 7 = great increase) to indicate their answers (3-month sample: α = .91; 15-month sample: α = .92). Concerns. One item asked about future tsunami threat concerns and one item asked about concern that the government was favoring some survivors by giving them more supplies than others. Participants used a 5-point scale (1 = strongly disagree to 5 = strongly agree) to indicate their answers.
Journal of Trauma & Dissociation, 15:219–239, 2014
225
Traumatic event exposure. Five items adapted from Freedy et al. (1994) asked about lifetime exposure to traumatic events (e.g., being physically attacked). Participants used a 2-point scale (1 = no, 2 = yes) to indicate their answers.
Downloaded by [University of Nebraska, Lincoln] at 17:38 13 April 2015
Procedure The project was approved by the Western Washington University Human Participants Research Committee and followed the American Psychological Association ethical guidelines. The first author trained four psychiatric nurses in questionnaire administration. Sampling occurred in two seaside villages near Khao Lak, a region that experienced extreme destruction and accounted for more than three quarters of all deaths in Thailand (residents and foreign tourists). The region’s economic base—fishing and tourism—and infrastructure virtually collapsed (Nidhiprabha, 2007). Participation was voluntary and no inducements were offered. 3 months post-tsunami. The sampling procedure involved approaching every other resident in five shelters, up to a sample size of 250, and returning one time if residents were not available. This procedure was modified slightly because 25 people asked and were allowed to complete the assessment instruments. These individuals overheard the nurses approach potential respondents and asked to participate. The criteria for inclusion were living in a shelter and being at least 18 years old. Most persons (60%) completed the instruments with some assistance (the nurses answered questions about a few items, if asked), followed by no assistance (20%) and having the assessment instruments read to them (20%). Assessment instruments were completed during the last 2 weeks of March 2005, and each took about 35 min. 15 months post-tsunami. The sampling procedure involved approaching every other residence in four communities until we achieved a sample of 250 participants. If after two visits no one participated, we sampled the adjacent house. Most participants (70%) completed the instruments with some assistance, followed by no assistance (20%) and having the instruments read to them (10%). Assessment instruments were completed during the last 2 weeks of March 2006, and each took about 35 min.
RESULTS Data Analysis Procedure We conducted hierarchical regression analyses to determine how the predictor variables related to PTS and posttraumatic growth (Freedy et al., 1994; Sattler et al., 2002, 2006). Based on COR and prior research, we first entered individual characteristics (e.g., gender), then variables relating to experiences prior to (e.g., trauma exposure) and during (life threat) the event, and finally
Downloaded by [University of Nebraska, Lincoln] at 17:38 13 April 2015
226
D. N. Sattler et al.
variables assessing responses. This approach allowed us to assess the variables in chronological order and to assess the effects of later variables while controlling for preexisting variables. We used a model-building approach to examine the relationships between the predictor and outcome variables within each sample and compare them across groups. First we examined the associations of predictor variables with PTS in a hierarchical regression analysis using data from the 3-month sample only. We replicated this analysis using the 15-month sample to examine the relationships 1 year later. Next, we tested the statistical equivalence of the regression coefficients by aggregating the samples into a single data file and distinguishing them with a dummy-coded time variable (coded as 0 = 3 months, 1 = 15 months). We created interaction terms between the time dummy code and the other predictors by grand-mean centering the predictor variables (to reduce multicollinearity between the interactions and parent terms) and multiplying each by the time dummy code (Aiken & West, 1991). We then replicated the previous regression model with both data sets combined, adding the time dummy code and the interaction terms in two additional steps. The interaction terms examined whether the relationship between a predictor and an outcome differed as a function of time. We repeated this model-building approach with posttraumatic growth as the outcome variable.
Experiences During the Tsunami and Property Damage 3 months post-tsunami. Most participants witnessed the tsunami (79%) and were concerned that they (50%) or someone they knew (65%) would lose their life. About one third (36%) were injured. Participants experienced significant damage to their homes: total loss (60%), major damage (20%), moderate damage (12%), minor damage (7%), and unknown (1%). Most did not have insurance coverage for their homes or possessions (93%) and had not been through another disaster (81%). They were without services for about 2 weeks: electricity (M = 13 days, SD = 18), running water (M = 18 days, SD = 22), and telephone (M = 13 days, SD = 20). The majority experienced irreparable damage to their homes and significant disruption of their lives. 15 months post-tsunami. Most participants witnessed the tsunami (74%) and were concerned that they (51%) or someone they knew (69%) would lose their life. About one third (36%) were injured. Participants experienced significant damage to their homes: total loss (50%), major damage (17%), moderate damage (16%), minor damage (8%), and no damage (9%). Most did not have insurance coverage (90%) and had not been through another disaster (86%). They were without services for almost 3 weeks: electricity (M = 19 days, SD = 22), running water (M = 22 days, SD = 22), and telephone (M = 22 days, SD = 25). The majority experienced irreparable damage to their homes and significant disruption of their lives.
Journal of Trauma & Dissociation, 15:219–239, 2014
227
Downloaded by [University of Nebraska, Lincoln] at 17:38 13 April 2015
PTS 3 months post-tsunami. Correlations among PTS, posttraumatic growth, demographics, traumatic event exposure, resource loss, somatic problems, social support, coping, and concerns are presented in Table 1. Table 2 shows that all predictor variables together explained 55% of the variance in PTS, F(17, 212) = 17.12, p < .001. PTS was positively associated with prior traumatic experiences; life threat during the tsunami; loss of personal characteristic resources, condition resources, and object resources; somatic problems; social support; emotion-focused coping; and concerns about government favoritism. There was no relationship among PTS (M = 37.11, SD = 16.91) and posttraumatic growth (M = 4.95, SD = 0.91). 15 months post-tsunami. Correlations among PTS, posttraumatic growth, demographics, traumatic event exposure, resource loss, somatic problems, social support, coping, and concerns are presented in Table 3. Table 2 shows that all predictor variables together explained 64% of the variance in PTS, F(17, 212) = 24.87, p < .001. PTS was associated with lack of prior disaster experience; prior traumatic experiences; life threat during the tsunami; loss of personal characteristic resources, condition resources, and energy resources; low object resource loss; somatic problems; and social support. There was no relationship among PTS (M = 26.48, SD = 15.28) and posttraumatic growth (M = 4.76, SD = 0.89). Comparison at 3 and 15 months. We combined the two samples and reran the analyses with time and the interactions between time and the other predictors. This allowed us to test the equivalence of beta-weights across samples. Table 2 shows that the Time × Predictor interaction terms explained an additional 3.2% of the variance in PTS beyond the main effects, F(17, 424) = 2.292, p = .003. The relation between object resource loss differed across samples. In the 3-month sample the effect was significantly positive, whereas in the 15-month sample the effect was significantly negative. Finally, although higher levels of emotion-focused coping were associated with more PTS in the 3-month sample, emotion-focused coping was unrelated to PTS in the 15-month sample.
Posttraumatic Growth 3 months post-tsunami. Table 4 shows that all predictor variables together explained 18% of the variance in posttraumatic growth, F(17, 212) = 4.05, p < .001. Posttraumatic growth was positively associated with education, prior trauma experience, somatic problems, social support, and problem-focused coping. 15 months post-tsunami. Table 4 shows that all predictor variables together explained 19% of the variance in posttraumatic growth, F(17, 216) = 4.20, p < .001.
228
2
3
.05
.08
.06
.25∗∗∗ .03
.08
.07
.14∗
.56∗∗∗
.03
.05 .03 .05 .26∗∗ .06 .11 .15∗ .23∗∗ .32∗∗∗ −.03 .29∗∗∗ −.10
−.07
.33∗∗∗ .12 .33∗∗∗ .57∗∗∗ .31∗∗∗ .39∗∗∗
.05
— .03 — .10 −.16∗∗ — .04 −.10 .02 −.04 .27∗∗∗ −.13∗ −.05 −.02 .10 .12 .17∗∗ −.27∗∗∗ .27∗∗∗ .01 .03 .04 .10 .48∗∗∗
1
Notes: PTS = posttraumatic stress symptoms. ∗ p < .05 ∗∗ p < .01 ∗∗∗ p < .001
PTS Posttraumatic growth Gender Age Education Disaster experience Trauma exposure Life threat concern Personal characteristic resource loss 10. Condition resource loss 11. Energy resource loss 12. Basic object loss 13. Object resource loss 14. Somatic problems 15. Social support 16. Problem-focused coping 17. Emotion-focused coping 18. Future tsunami concern 19. Government favoritism concern
1. 2. 3. 4. 5. 6. 7. 8. 9.
Variable
5
.04
.06
.13∗
.03 .13∗ .14∗ .03 −.02 .04
−.03
−.07
−.04
−.09 −.01
−.06
−.05 .05 −.05 −.03 .02 −.02
.08
— −.30∗∗∗ .09 −.06
6
−.04
.09 −.13 .01 −.08 .14∗ .00
−.02
— −.35∗∗ — −.04 −.00 −.08 .08 −.01 −.14 .07 −.09
4
— .14∗
8
.04
−.10
.04
.10 −.05 .03 .10 .03 .13∗
.18∗∗
.05
.18∗∗
.06 .24∗∗∗ .21∗∗∗ .19∗∗ .18∗∗ .25∗∗∗
.15∗ .48∗∗∗
— .04 .04
7
—
10
11
.10
−.01
.34∗∗∗
.05
.10
.30∗∗∗
— .42∗∗∗ .26∗∗∗ .07 .16∗
12
.13∗
.01
.15∗
.00
.31∗∗∗ .12
.40∗∗∗ .34∗∗∗ — .24∗∗∗ −.07 −.07 .43∗∗∗ .14∗ .20∗∗ .38∗∗∗ .17∗ .16∗ .17∗∗ .15∗ .19∗∗ .25∗∗∗ .24∗∗∗ .21∗∗∗
.42∗∗∗
—
9
14
15
—
16
.05
−.00
.26∗∗∗ .11
−.00 .30
.32∗∗∗
.14∗
.28∗∗∗ .33∗∗∗ .40∗∗∗ .40∗∗∗
— .28∗∗∗ — .23∗∗∗ .21∗∗∗ — .19∗∗ .30∗∗ .50∗∗∗
13
18 19
.90 .14∗ —
−.20 —
—
17
TABLE 1 3 Months Post-Tsunami: Correlations Among PTS, Posttraumatic Growth, Demographics, Traumatic Event Exposure, Resource Loss, Somatic Problems, Social Support, Coping, and Concerns (N = 248)
Downloaded by [University of Nebraska, Lincoln] at 17:38 13 April 2015
229
∗∗
p < .05 p < .01 ∗∗∗ p < .001
∗
Step 1 Gender (0 = male, 1 = female) Age Education Step 2: Prior experience Natural disaster Trauma Step 3 Life threat Step 4: Resource loss Personal characteristic Condition Energy Basic object Object Step 5 Somatic problems Step 6 Social support Step 7: Coping Problem focused Emotion focused Step 8: Concerns Future tsunami threat Government favoritism
Variable
0.28 1.38 1.59 1.30 1.11 0.90
2.92 2.69 7.02 0.59 1.95
1.20
4.67
2.12
2.90 1.27
−1.59 2.56
1.29 1.06 1.09 1.15 1.33
0.10 0.69
0.07 0.04
5.17 3.52 1.01 −0.60 2.71
2.31
SE B
3.28
B
.03 .11∗
.10 .29∗∗∗
.11∗
.41∗∗∗
.29∗∗∗ .21∗∗ .06 −.03 .13∗
.25
∗∗∗
−.04 .14∗
.04 .00
.10
β
.55
.54∗∗∗
.47∗
.46∗∗∗
.32∗∗∗
.07∗∗∗
.01
−.00
Adjusted R2
3 months post-tsunami (N = 248)
1.46 −0.00
0.92 0.66
3.50
2.20
4.07 1.96 4.06 1.95 −4.26
4.09
0.84 0.76
1.28 1.07
1.12
0.21
1.10 0.93 1.20 1.23 1.38
1.71
2.80 0.92
0.08 0.64
−0.03 −0.31 −9.72 4.33
1.98
SE B
0.84
B
.09 .00
.04 .03
.15∗∗
.50∗∗∗
.29∗∗∗ .13∗ .29∗∗ .13 −.29∗∗
.22
∗∗
−.22∗∗ .30∗∗∗
−.03 −.03
.03
β
.64
.64
.64∗∗
.62∗∗∗
.44∗∗∗
.18∗∗
.14∗∗∗
−.01
Adjusted R2
15 months post-tsunami (N = 255)
−.014 −.013
−.021 −.143∗∗∗
.047
.032
.006 −.001 .052 .119∗∗ −.136∗∗
−.023
−.034 .038
.024 .006
−.001
β
Time × Variable (N = 503)
TABLE 2 Predicting Posttraumatic Stress Symptoms with Demographics, Trauma Exposure, Resource Loss, Social Support, Coping, and Concerns
Downloaded by [University of Nebraska, Lincoln] at 17:38 13 April 2015
230
— −.08 −.17∗∗ .18∗∗ .02 −.03 .07 −.13∗
2
.24∗∗∗
.04
.38∗∗∗
.02
−.02
−.02
.03
.06
−.05 .02 −.02 −.04 −.07 −.02
.04
— .03 −.03 .11 −.14∗ .05 .08
3
.09
.50∗∗∗ −.02 .43 .00 .38∗∗∗ −.04 .68∗∗∗ −.10 .35∗∗∗ .39∗∗∗ .21∗∗∗ .40∗∗∗
.44∗∗∗ −.01
— −.06 .01 −.01 −.01 −.28∗∗∗ .31∗∗∗ .26∗∗∗ .53∗∗∗
1
Notes. PTS = posttraumatic stress symptoms. ∗ p < .05 ∗∗ p < .01 ∗∗∗ p < .001
PTS Posttraumatic growth Gender Age Education Disaster experience Trauma exposure Life threat concern Personal characteristic resource loss 10. Condition resource loss 11. Energy resource loss 12. Basic resource object 13. Object resource loss 14. Somatic problems 15. Social support 16. Problem-focused coping 17. Emotion-focused coping 18. Future tsunami concern 19. Government favoritism concern
1. 2. 3. 4. 5. 6. 7. 8. 9.
Variable
5
.12
.04
−.11
.00 .09 .14∗ −.05 −.24∗∗∗ −.18∗∗
.03
−.07
−.04
−.13∗
−.29∗∗∗ −.22∗∗∗ −.23∗∗∗ −.21∗∗∗ .03 −.03
−.20∗∗∗
−.14∗
−.16∗
7
−.06
−.07
.56∗∗∗
—
9
.38∗∗∗
.29∗∗∗
.21∗∗∗
.00
.25∗∗∗
.09 .63∗∗∗ .21∗∗∗ .65∗∗∗ .26∗∗∗ .64∗∗∗ .25∗∗∗ .31∗∗∗ .14∗ −.05 .21∗∗∗ −.01
.04
— .26∗∗∗
8
.18∗∗ .15∗
.16∗∗ .11 .20∗∗∗ .25∗∗∗ .19 .07
.15∗
— −.23∗∗∗ — −.16∗ −.09 −.20∗∗∗ .06
6
.16∗
−.04 −.10 −.12 −.05 .27∗∗∗ .20∗∗
−.02
— −.35∗∗∗ — −.03 −.01 −.13∗ .15∗ .13∗ −.11 .20∗∗∗ −.16∗
4
.02
−.10
.12
.54∗∗∗ .47∗∗∗ .46∗∗∗ .12∗ .13∗ .01
—
10
−.01
−.14∗
.18∗∗
— .69∗∗∗ .73∗∗∗ .33∗∗∗ .21∗∗∗ .06
11
.14∗
−.07
.07
— .76∗∗∗ .20∗∗∗ .09 .02
12
14
15
16
.06.
−.04
.14∗
17
.17∗
.07
.12
.08
.10
—
18 19
.28∗∗∗ .21∗∗∗ .33∗∗∗
.11
.37∗∗∗ .39∗∗∗ .50∗∗∗ —
— .27∗∗∗ — .10 .26∗∗∗ — .01 .13∗ .59∗∗∗ —
13
TABLE 3 15 Months Post-Tsunami: Correlations Among PTS, Posttraumatic Growth, Demographics, Traumatic Event Exposure, Resource Loss, Somatic Problems, Social Support, Coping, and Concerns (N = 255)
Downloaded by [University of Nebraska, Lincoln] at 17:38 13 April 2015
231
∗∗
p < .05 p < .01 ∗∗∗ p < .001
∗
Step 1 Gender (0 = male, 1 = female) Age Education Step 2: Prior experience Natural disaster Trauma Step 3 Life threat Step 4: Resource loss Personal characteristic Condition Energy Basic object Object Step 5 Somatic problems Step 6 Social support Step 7: Coping Problem focused Emotion focused Step 8: Concerns Future tsunami threat Government favoritism
Variable
.02 .09 .11 .09 .08 .07
.34 .24 .05
.14 −.04
.06
.10
.05
.15 .07
.09 .15
.08 .07 .07 .07 .08
.01 .04
−.00 .13
.03 .04 .03 .12 .00
.12
SE B
−.21
B
.11 −.04
.16∗ .04
.24∗∗∗
.19∗∗
.03 .04 .04 .12 .00
.10
.04 .15∗
−.02 .26∗∗∗
−.11
β
.18
.18
.17∗∗∗
.12∗∗
.10
.10
.09
.08
∗∗∗
Adjusted R2
3 months post-tsunami (N = 248)
−.08 −.06
.42 −.14
.07 .07
.11 .09
.10
.02
−.01 .55
.09 .07 .09 .10 .11
.08
.18 .06
.01 .04
.11
SE B
−.21 .05 .04 .07 .01
.10
.03 −.05
−.01 .07
.16
B
−.07 −.06
.29∗∗∗ −.11
.38∗∗∗
−.05
−.25∗ .06 .05 .08 .01
.09
.01 −.05
−.11 .13
.09
β
.19
.19∗∗
.15∗∗∗
.04
.04
.03
.03
.03
∗
Adjusted R2
15 months post-tsunami (N = 255)
−.087 −.020
.071 −.073
.051
−.116∗
.002 −.006 −.030 −.049 .059
.044
−.024 −.099∗
−.007 −.087
.106∗
β
Time × Variable (N = 503)
TABLE 4 Predicting Posttraumatic Growth with Demographics, Trauma Exposure, Resource Loss, Social Support, Coping, and Concerns
Downloaded by [University of Nebraska, Lincoln] at 17:38 13 April 2015
232
D. N. Sattler et al.
Downloaded by [University of Nebraska, Lincoln] at 17:38 13 April 2015
Posttraumatic growth was negatively associated with personal characteristic resource loss and positively associated with social support and problem-focused coping. Comparison at 3 and 15 months. Table 4 shows that the inclusion of the Time × Predictor interaction terms explained an additional 5.4% of the variance in posttraumatic growth above and beyond the main effects, F(17, 428) = 1.816, p = .024. The positive association between education and posttraumatic growth was weaker in the 15-month than in the 3-month sample. Although trauma was positively associated with posttraumatic growth in the 3-month sample, these factors were unrelated in the 15-month sample.
Comparison of Outcomes at 3 and 15 Months We performed independent samples t tests on the key outcome variables to compare the means across the 3-month and 15-month samples. Table 5 shows that scores were lower at 15 months than at 3 months for PTS, somatic problems, posttraumatic growth, resource loss, social support, problemfocused coping, belief in governmental favoritism, and concern about future tsunami threats.
DISCUSSION The results show that participants were exposed to unprecedented horror and loss of life and property and saw reminders daily of the devastation TABLE 5 Comparison of Variables at 3 and 15 Months 3 months (N = 248) Variable Posttraumatic stress Posttraumatic growth Natural disaster experience Traumatic experience Life threat/injury concern Personal characteristic resource loss Condition resource loss Energy resource loss Basic resource loss Object resource loss Somatic problems Social support Problem-focused coping Emotion-focused coping Future tsunami threat Government favoritism
15 months (N = 255)
M
SD
M
SD
df
t
p
d
37.11 4.95 1.81 0.62 2.66 1.70
1.02 0.57 0.39 0.93 0.92 0.95
26.44 4.75 1.86 0.58 2.62 1.74
1.02 0.50 0.35 1.05 0.78 1.06
495 499 500 498 497 501
7.38 2.49 −1.57 0.47 0.53 −0.27
.05 >.05 >.05
10.46 0.35
1.57 1.79 2.66 2.10 4.12 2.07 2.18 1.64 2.57 2.62
0.65 0.65 0.06 0.84 0.21 0.64 0.61 0.71 0.73 0.95
1.37 1.42 1.82 1.72 2.98 1.76 1.98 1.53 2.34 2.19
0.65 0.64 0.06 0.97 0.21 0.64 0.64 0.75 0.88 0.88
501 501 501 501 501 500 500 499 501 499
2.13 4.09 9.58 4.60 3.98 5.52 3.40 1.70 3.26 5.29
Journal of Trauma & Dissociation Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/wjtd20
Indian Ocean Tsunami: Relationships Among Posttraumatic Stress, Posttraumatic Growth, Resource Loss, and Coping at 3 and 15 Months a
b
David N. Sattler PhD , Sawitri Assanangkornchai MD , Adam M. a
c
Moller PhD , Wiworn Kesavatana-Dohrs PhD & James M. Graham PhD
a
a
Department of Psychology , Western Washington University , Bellingham , Washington , USA b
Click for updates
Department of Psychiatry , Prince of Songkhla University , Hat Yai , Thailand c
Department of Asian Languages and Literature , University of Washington , Seattle , Washington , USA Accepted author version posted online: 10 Jan 2014.Published online: 11 Mar 2014.
To cite this article: David N. Sattler PhD , Sawitri Assanangkornchai MD , Adam M. Moller PhD , Wiworn Kesavatana-Dohrs PhD & James M. Graham PhD (2014) Indian Ocean Tsunami: Relationships Among Posttraumatic Stress, Posttraumatic Growth, Resource Loss, and Coping at 3 and 15 Months, Journal of Trauma & Dissociation, 15:2, 219-239, DOI: 10.1080/15299732.2014.869144 To link to this article: http://dx.doi.org/10.1080/15299732.2014.869144
PLEASE SCROLL DOWN FOR ARTICLE Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content.
Downloaded by [University of Nebraska, Lincoln] at 17:38 13 April 2015
This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http://www.tandfonline.com/page/termsand-conditions
Journal of Trauma & Dissociation, 15:219–239, 2014 Copyright © Taylor & Francis Group, LLC ISSN: 1529-9732 print/1529-9740 online DOI: 10.1080/15299732.2014.869144
Indian Ocean Tsunami: Relationships Among Posttraumatic Stress, Posttraumatic Growth, Resource Loss, and Coping at 3 and 15 Months
Downloaded by [University of Nebraska, Lincoln] at 17:38 13 April 2015
DAVID N. SATTLER, PhD Department of Psychology, Western Washington University, Bellingham, Washington, USA
SAWITRI ASSANANGKORNCHAI, MD Department of Psychiatry, Prince of Songkhla University, Hat Yai, Thailand
ADAM M. MOLLER, PhD Department of Psychology, Western Washington University, Bellingham, Washington, USA
WIWORN KESAVATANA-DOHRS, PhD Department of Asian Languages and Literature, University of Washington, Seattle, Washington, USA
JAMES M. GRAHAM, PhD Department of Psychology, Western Washington University, Bellingham, Washington, USA
This study examines variables associated with posttraumatic stress symptoms (PTS) and posttraumatic growth among 2 independent samples of survivors following the Indian Ocean tsunami in Khao Lak, Thailand. Participants were exposed to unprecedented horror and loss of life and property. At 3 months participants (N = 248) were living in temporary shelters, and at 15 months a second sample (N = 255) was living in homes built after the tsunami. Prior traumatic experiences, life threat, loss of personal characteristic resources and condition resources, somatic problems, and social support accounted for close to half of the variance in PTS in each sample. At 3 months, emotion-focused coping and concerns about government favoritism also contributed to PTS. At 15 months, lack of prior disaster experience and loss of energy resources also contributed to PTS. Distress was higher among participants surveyed at 3 months than among those surveyed at 15 months. Posttraumatic Received 30 November 2012; accepted 28 September 2013. Address correspondence to David N. Sattler, PhD, Department of Psychology, Western Washington University, Bellingham, WA 98225-9172. E-mail: [email protected] 219
220
D. N. Sattler et al.
growth was positively associated with social support and problem– focused coping in both samples. The findings support conservation of resources stress theory (Hobfoll, 2012) and underscore how systemic issues affect mental health. The implications of the findings are discussed, as is the educational International Tsunami Museum designed by the first author to address systemic stressors.
Downloaded by [University of Nebraska, Lincoln] at 17:38 13 April 2015
KEYWORDS posttraumatic stress, resource loss, conservation of resources, posttraumatic growth, tsunami, disaster, systemic stressor, tsunami museum
On December 26, 2004, a 9.3-magnitude earthquake off Northern Sumatra, Indonesia, generated a powerful tsunami that devastated coastal communities bordering the Indian Ocean (Patniak, Kumar, Ramu, & Prasad, 2012). The death toll—226,323 lives lost—placed the tsunami as the fourth deadliest disaster in history. Two million people were left homeless. Hundreds of thousands of persons were exposed to extreme horror and witnessed people severely injured (American Red Cross, 2009). Estimated property damage and economic losses were more than $10 billion (Nidhiprabha, 2007; United Nations Office for Disaster Risk Reduction, 2009). In the wake of the catastrophic disaster, survivors were exposed to unprecedented systemic stressors, which we define as a loss or lack of essential community resources necessary for the effective functioning of citizens and the community and that creates enduring hardships. Systemic stressors occur when essential systems that help citizens navigate life’s challenges are insufficient to meet fundamental needs for an extended period. The tsunami affected virtually every aspect of life in coastal communities, including family functioning (e.g., loss of and injury to family members, loss of home and employment), basic survival needs (e.g., food, water, shelter), and infrastructure (e.g., damaged roads, businesses, schools, sanitation services). Survivors lacked financial resources to obtain supplies or rebuild, lacked insurance coverage, and had difficulty accessing social support networks. They saw constant reminders of the disaster and feared for their safety because of new disaster threats (Bangkok Post, 2012; United Nations Thailand, 2008). Survivors endured chronic hardship for years in the aftermath of the disaster.
RESOURCE LOSS, PSYCHOLOGICAL DISTRESS, AND POSTTRAUMATIC GROWTH Conservation of resources theory (COR) provides an important framework for understanding how resource loss as a result of natural disasters affects
Downloaded by [University of Nebraska, Lincoln] at 17:38 13 April 2015
Journal of Trauma & Dissociation, 15:219–239, 2014
221
individuals. COR posits that individuals strive to acquire, maintain, and protect resources they value and use them in ways consistent with their culture. Valued resources provide meaning, safety, and security; are necessary to acquire or maintain other resources; and help buffer against the loss of other resources. Resources fall into four general categories: condition (e.g., employment), personal characteristics (e.g., optimism), energy (e.g., time for activities), and object (e.g., house, physical possessions). When valued resources are threatened or lost, COR predicts that individuals will experience stress. COR advances that individuals must invest resources (e.g., time, money) to protect against and recover from resource loss. Thus, individuals with greater resources should be less susceptible to resource loss and recover more easily, whereas those with fewer resources may be more susceptible and less able to recover lost resources (Hobfoll, 2012). Persons with limited social support networks also tend to fare worse than those with richer networks (Norris & Alegria, 2005; Smith & Freedy, 2000). Resource caravans—the concurrent development or overlap of resources—are a central concept for COR. Threat or loss to specific resources may affect other resources because of overlap and can have a potentially wide-reaching impact on resources and well-being (Hobfoll, 2012). For example, a survivor with no property insurance who loses his or her home and job may also experience threats to personal characteristic resources (e.g., self-efficacy) and condition resources (e.g., social roles). He or she also may have increased difficulty accessing his or her social support network. As a result, resource loss spirals—the continuing loss of resources in one area that results in a loss of resources in other areas—can develop (Hobfoll, 2011). Resource loss spirals create additional stressors that tax remaining resources as well as mental health and explain, in part, the association of personal characteristic resource loss and condition resource loss with stress responses such as posttraumatic stress (Hobfoll, 2011; Sattler et al., 2006). For example, Cerdá et al. (2013) found that 18 months after a hurricane in Texas, ongoing stressors such as the loss of a job, decreased income, and the inability to live in one’s home were associated with posttraumatic stress symptoms (PTS). Resource loss spirals can contribute to the development and maintenance of systemic stressors that endure for months or years. Survivors may be especially concerned about governmental favoritism—the perception that resources are being distributed inequitably during recovery or that government (or relief) agencies are favoring certain groups with resources or services, especially when recovery is prolonged (Sattler et al., 1995). Such concerns may be a significant source of stress. Few studies have examined the associations of resource loss variables and posttraumatic stress at various points during recovery. To address this issue, this study examines responses at the end of the acute stage (defined as the first 3 months) and during the chronic stage (defined as 4 or more months) of recovery (Baum, 1991). Posttraumatic stress can develop after
Downloaded by [University of Nebraska, Lincoln] at 17:38 13 April 2015
222
D. N. Sattler et al.
exposure to psychological trauma and involves reexperiencing the event through flashbacks or nightmares, avoiding stimuli associated with the event, and experiencing increased arousal (Cerdá et al., 2013). As the challenges of recovery following the tsunami continued month after month, there was great potential for resource loss spirals to compromise personal characteristic, condition, and energy resources. Prior exposure to traumatic events, somatic problems, and the degree to which survivors believed they or loved ones would lose their lives or be injured during the disaster are associated with PTS following disasters (Smith & Freedy, 2000). Survivors may experience resource gains as disaster relief organizations help communities rebuild. COR predicts that resource gains can offset losses to some degree and serve as a protective factor against negative outcomes associated with stress (Hobfoll, 2011). Posttraumatic growth theory suggests that resource gains may be viewed as a positive experience following loss due to a traumatic event. For example, survivors may report increased awareness of life priorities and positive life changes as a result of processing the event (Groleau, Calhoun, Cann, & Tedeschi, 2012). People may learn the value of preparation, develop new coping skills, gain a new appreciation of life, and experience an enhanced sense of self-efficacy (Linley & Joseph, 2004; Sattler, 2003). These positive outcomes may offset negative experiences to some degree (Groleau et al., 2012; Linley & Joseph, 2004). We examined the association among level of education and posttraumatic growth, although prior findings are mixed (Linley & Joseph, 2004). Few studies have examined COR or posttraumatic growth following disasters in countries other than the United States (Sattler et al., 2006). Coping strategies may serve as a protective or risk factor for PTS given that natural disasters threaten survivors’ sense of control, predictability, safety, and trust (Baum, 1991) and may be associated with posttraumatic growth. Problem-focused coping involves attempting to directly change or manage the situation, and emotion-focused coping involves attempting to adapt emotions in order to reduce distress (Folkman & Lazarus, 1985). Problem-focused coping may help survivors reestablish feelings of control or obtain resources and is associated with resilience and posttraumatic growth (Wadsworth, Santiago, & Einhorn, 2009).
RESEARCH QUESTIONS Based on COR and prior research, we examined the following questions with two independent groups of survivors at 3 months and 15 months after the tsunami. First, how do demographic variables, prior traumatic experiences, experiences during the tsunami, resource loss, somatic problems, social support, coping, and concerns about tsunami threats and governmental favoritism relate to PTS and posttraumatic growth at 3 and 15 months?
Journal of Trauma & Dissociation, 15:219–239, 2014
223
Second, how do the relations between these predictors and posttraumatic growth differ across the two samples? Answers to these questions may assist theory development and the manner in which intervention and recovery programs provide assistance.
METHOD
Downloaded by [University of Nebraska, Lincoln] at 17:38 13 April 2015
Participants 3 months post-tsunami. The participants were 248 persons (97 men, 151 women) living in temporary shelters (age: M = 36, SD = 11, range = 18–72 years). All were Thai. Most were married (69%), followed by single (21%), widowed (7%), and separated/divorced (3%). Almost all were Buddhist (94%), followed by Muslim (4%) and Christian (2%). More than half had less than a secondary school education (56%), followed by completion of secondary school (20%), college degree (20%), and unknown (4%). The response rate was 94%, and those who declined to participate were not asked why. 15 months post-tsunami. The participants were 255 persons (148 men, 107 women; none participated in Study 1) who had previously lived in a temporary shelter and were living in homes in four neighborhoods built after the tsunami (age: M = 38, SD = 14, range = 18–84 years). Almost all were Thai (97%), followed by Laotian (2%) and Chinese (1%). Most were married (62%), followed by single (20%), widowed (5%), separated/divorced (2%), and unknown (11%). Most were Buddhist (90%), followed by Muslim (4%), Christian (2%), other (2%), and unknown (2%). More than half had less than a secondary school education (56%), followed by completion of secondary school (26%) and college degree (18%). The response rate was 95%, and those who declined to participate were not asked why. Two assessment instruments were partially completed and not analyzed. The demographics were similar to those of the sample collected at 3 months.
Assessment Instruments We used a version of the committee approach to translate the assessment instruments into Thai (Matsumoto & van de Vijver, 2011; van de Vijver & Leung, 1997) that we successfully used previously (Sattler et al., 2002, 2006). The fourth author translated the items and the second author revised the translation. Four medical nurses in Thailand evaluated the instruments, asked participants about readability, and reported that it was clear and understandable. A cover letter introduced the study, presented information on informed consent, and indicated that responses were anonymous. The instruments were presented in the following order.
Downloaded by [University of Nebraska, Lincoln] at 17:38 13 April 2015
224
D. N. Sattler et al.
Demographics and life threat. Seven items asked for demographic information and eight items asked about life threat during the tsunami, based on Sattler et al. (2006). Participants checked their choices or wrote in a number to indicate their answers. Four items asked whether participants were concerned that they or someone they knew would lose his or her life or be seriously injured during the tsunami. Respondents used a 4-point scale (1 = not at all to 4 = quite a bit) to indicate their answers (3-month sample: α = .84; 15-month sample: α = .83). Resource loss. An 18-item scale assessed loss of personal characteristic, object, condition, and energy resources in the past 3 months (Freedy, Saladin, Kilpatrick, Resnick, & Saunders, 1994; modified by Sattler et al., 2006). Participants used a 5-point scale (0 = no loss to 4 = great loss) to indicate their answers (3-month sample: α = .87; 15-month sample: α = .95). PTS and somatic problems. The 22-item Impact of Event Scale–Revised assessed symptoms associated with posttraumatic stress disorder in the past 7 days (Weiss & Marmar, 1996; 3-month sample: α = .94; 15-month sample: α = .94). Participants also answered four items about somatic problems (e.g., headaches, muscle pain) in the past 7 days (3-month sample: α = .76; 15month sample: α = .87). They used a 5-point scale (0 = not at all to 4 = extremely) to indicate their answers. Social support. Seven Social Support Index items assessed support (McCubbin & Thompson, 1991). An example is “If I had an emergency, even people I do not know in this community would be willing to help.” Participants used a 5-point scale (1 = not at all to 5 = extremely) to indicate their answers (3-month sample: α = .83; 15-month sample: α = .87). Coping. Eight items from the Family Crisis Oriented Personal Evaluation Scale assessed problem-focused coping (3-month sample: α = .74; 15-month sample: α = .77) and emotion-focused coping (3-month sample: α = .73; 15-month sample: α = .79; McCubbin & Thompson, 1991). An example is “I face problems head-on and try to get the solutions right away.” Participants used a 5-point scale (1 = not at all to 5 = extremely) to indicate their answers. Posttraumatic growth. Fifteen items from the Posttraumatic Growth Inventory assessed growth (Tedeschi & Calhoun, 1996). Participants used a 7-point scale (1 = great decrease to 4 = no change to 7 = great increase) to indicate their answers (3-month sample: α = .91; 15-month sample: α = .92). Concerns. One item asked about future tsunami threat concerns and one item asked about concern that the government was favoring some survivors by giving them more supplies than others. Participants used a 5-point scale (1 = strongly disagree to 5 = strongly agree) to indicate their answers.
Journal of Trauma & Dissociation, 15:219–239, 2014
225
Traumatic event exposure. Five items adapted from Freedy et al. (1994) asked about lifetime exposure to traumatic events (e.g., being physically attacked). Participants used a 2-point scale (1 = no, 2 = yes) to indicate their answers.
Downloaded by [University of Nebraska, Lincoln] at 17:38 13 April 2015
Procedure The project was approved by the Western Washington University Human Participants Research Committee and followed the American Psychological Association ethical guidelines. The first author trained four psychiatric nurses in questionnaire administration. Sampling occurred in two seaside villages near Khao Lak, a region that experienced extreme destruction and accounted for more than three quarters of all deaths in Thailand (residents and foreign tourists). The region’s economic base—fishing and tourism—and infrastructure virtually collapsed (Nidhiprabha, 2007). Participation was voluntary and no inducements were offered. 3 months post-tsunami. The sampling procedure involved approaching every other resident in five shelters, up to a sample size of 250, and returning one time if residents were not available. This procedure was modified slightly because 25 people asked and were allowed to complete the assessment instruments. These individuals overheard the nurses approach potential respondents and asked to participate. The criteria for inclusion were living in a shelter and being at least 18 years old. Most persons (60%) completed the instruments with some assistance (the nurses answered questions about a few items, if asked), followed by no assistance (20%) and having the assessment instruments read to them (20%). Assessment instruments were completed during the last 2 weeks of March 2005, and each took about 35 min. 15 months post-tsunami. The sampling procedure involved approaching every other residence in four communities until we achieved a sample of 250 participants. If after two visits no one participated, we sampled the adjacent house. Most participants (70%) completed the instruments with some assistance, followed by no assistance (20%) and having the instruments read to them (10%). Assessment instruments were completed during the last 2 weeks of March 2006, and each took about 35 min.
RESULTS Data Analysis Procedure We conducted hierarchical regression analyses to determine how the predictor variables related to PTS and posttraumatic growth (Freedy et al., 1994; Sattler et al., 2002, 2006). Based on COR and prior research, we first entered individual characteristics (e.g., gender), then variables relating to experiences prior to (e.g., trauma exposure) and during (life threat) the event, and finally
Downloaded by [University of Nebraska, Lincoln] at 17:38 13 April 2015
226
D. N. Sattler et al.
variables assessing responses. This approach allowed us to assess the variables in chronological order and to assess the effects of later variables while controlling for preexisting variables. We used a model-building approach to examine the relationships between the predictor and outcome variables within each sample and compare them across groups. First we examined the associations of predictor variables with PTS in a hierarchical regression analysis using data from the 3-month sample only. We replicated this analysis using the 15-month sample to examine the relationships 1 year later. Next, we tested the statistical equivalence of the regression coefficients by aggregating the samples into a single data file and distinguishing them with a dummy-coded time variable (coded as 0 = 3 months, 1 = 15 months). We created interaction terms between the time dummy code and the other predictors by grand-mean centering the predictor variables (to reduce multicollinearity between the interactions and parent terms) and multiplying each by the time dummy code (Aiken & West, 1991). We then replicated the previous regression model with both data sets combined, adding the time dummy code and the interaction terms in two additional steps. The interaction terms examined whether the relationship between a predictor and an outcome differed as a function of time. We repeated this model-building approach with posttraumatic growth as the outcome variable.
Experiences During the Tsunami and Property Damage 3 months post-tsunami. Most participants witnessed the tsunami (79%) and were concerned that they (50%) or someone they knew (65%) would lose their life. About one third (36%) were injured. Participants experienced significant damage to their homes: total loss (60%), major damage (20%), moderate damage (12%), minor damage (7%), and unknown (1%). Most did not have insurance coverage for their homes or possessions (93%) and had not been through another disaster (81%). They were without services for about 2 weeks: electricity (M = 13 days, SD = 18), running water (M = 18 days, SD = 22), and telephone (M = 13 days, SD = 20). The majority experienced irreparable damage to their homes and significant disruption of their lives. 15 months post-tsunami. Most participants witnessed the tsunami (74%) and were concerned that they (51%) or someone they knew (69%) would lose their life. About one third (36%) were injured. Participants experienced significant damage to their homes: total loss (50%), major damage (17%), moderate damage (16%), minor damage (8%), and no damage (9%). Most did not have insurance coverage (90%) and had not been through another disaster (86%). They were without services for almost 3 weeks: electricity (M = 19 days, SD = 22), running water (M = 22 days, SD = 22), and telephone (M = 22 days, SD = 25). The majority experienced irreparable damage to their homes and significant disruption of their lives.
Journal of Trauma & Dissociation, 15:219–239, 2014
227
Downloaded by [University of Nebraska, Lincoln] at 17:38 13 April 2015
PTS 3 months post-tsunami. Correlations among PTS, posttraumatic growth, demographics, traumatic event exposure, resource loss, somatic problems, social support, coping, and concerns are presented in Table 1. Table 2 shows that all predictor variables together explained 55% of the variance in PTS, F(17, 212) = 17.12, p < .001. PTS was positively associated with prior traumatic experiences; life threat during the tsunami; loss of personal characteristic resources, condition resources, and object resources; somatic problems; social support; emotion-focused coping; and concerns about government favoritism. There was no relationship among PTS (M = 37.11, SD = 16.91) and posttraumatic growth (M = 4.95, SD = 0.91). 15 months post-tsunami. Correlations among PTS, posttraumatic growth, demographics, traumatic event exposure, resource loss, somatic problems, social support, coping, and concerns are presented in Table 3. Table 2 shows that all predictor variables together explained 64% of the variance in PTS, F(17, 212) = 24.87, p < .001. PTS was associated with lack of prior disaster experience; prior traumatic experiences; life threat during the tsunami; loss of personal characteristic resources, condition resources, and energy resources; low object resource loss; somatic problems; and social support. There was no relationship among PTS (M = 26.48, SD = 15.28) and posttraumatic growth (M = 4.76, SD = 0.89). Comparison at 3 and 15 months. We combined the two samples and reran the analyses with time and the interactions between time and the other predictors. This allowed us to test the equivalence of beta-weights across samples. Table 2 shows that the Time × Predictor interaction terms explained an additional 3.2% of the variance in PTS beyond the main effects, F(17, 424) = 2.292, p = .003. The relation between object resource loss differed across samples. In the 3-month sample the effect was significantly positive, whereas in the 15-month sample the effect was significantly negative. Finally, although higher levels of emotion-focused coping were associated with more PTS in the 3-month sample, emotion-focused coping was unrelated to PTS in the 15-month sample.
Posttraumatic Growth 3 months post-tsunami. Table 4 shows that all predictor variables together explained 18% of the variance in posttraumatic growth, F(17, 212) = 4.05, p < .001. Posttraumatic growth was positively associated with education, prior trauma experience, somatic problems, social support, and problem-focused coping. 15 months post-tsunami. Table 4 shows that all predictor variables together explained 19% of the variance in posttraumatic growth, F(17, 216) = 4.20, p < .001.
228
2
3
.05
.08
.06
.25∗∗∗ .03
.08
.07
.14∗
.56∗∗∗
.03
.05 .03 .05 .26∗∗ .06 .11 .15∗ .23∗∗ .32∗∗∗ −.03 .29∗∗∗ −.10
−.07
.33∗∗∗ .12 .33∗∗∗ .57∗∗∗ .31∗∗∗ .39∗∗∗
.05
— .03 — .10 −.16∗∗ — .04 −.10 .02 −.04 .27∗∗∗ −.13∗ −.05 −.02 .10 .12 .17∗∗ −.27∗∗∗ .27∗∗∗ .01 .03 .04 .10 .48∗∗∗
1
Notes: PTS = posttraumatic stress symptoms. ∗ p < .05 ∗∗ p < .01 ∗∗∗ p < .001
PTS Posttraumatic growth Gender Age Education Disaster experience Trauma exposure Life threat concern Personal characteristic resource loss 10. Condition resource loss 11. Energy resource loss 12. Basic object loss 13. Object resource loss 14. Somatic problems 15. Social support 16. Problem-focused coping 17. Emotion-focused coping 18. Future tsunami concern 19. Government favoritism concern
1. 2. 3. 4. 5. 6. 7. 8. 9.
Variable
5
.04
.06
.13∗
.03 .13∗ .14∗ .03 −.02 .04
−.03
−.07
−.04
−.09 −.01
−.06
−.05 .05 −.05 −.03 .02 −.02
.08
— −.30∗∗∗ .09 −.06
6
−.04
.09 −.13 .01 −.08 .14∗ .00
−.02
— −.35∗∗ — −.04 −.00 −.08 .08 −.01 −.14 .07 −.09
4
— .14∗
8
.04
−.10
.04
.10 −.05 .03 .10 .03 .13∗
.18∗∗
.05
.18∗∗
.06 .24∗∗∗ .21∗∗∗ .19∗∗ .18∗∗ .25∗∗∗
.15∗ .48∗∗∗
— .04 .04
7
—
10
11
.10
−.01
.34∗∗∗
.05
.10
.30∗∗∗
— .42∗∗∗ .26∗∗∗ .07 .16∗
12
.13∗
.01
.15∗
.00
.31∗∗∗ .12
.40∗∗∗ .34∗∗∗ — .24∗∗∗ −.07 −.07 .43∗∗∗ .14∗ .20∗∗ .38∗∗∗ .17∗ .16∗ .17∗∗ .15∗ .19∗∗ .25∗∗∗ .24∗∗∗ .21∗∗∗
.42∗∗∗
—
9
14
15
—
16
.05
−.00
.26∗∗∗ .11
−.00 .30
.32∗∗∗
.14∗
.28∗∗∗ .33∗∗∗ .40∗∗∗ .40∗∗∗
— .28∗∗∗ — .23∗∗∗ .21∗∗∗ — .19∗∗ .30∗∗ .50∗∗∗
13
18 19
.90 .14∗ —
−.20 —
—
17
TABLE 1 3 Months Post-Tsunami: Correlations Among PTS, Posttraumatic Growth, Demographics, Traumatic Event Exposure, Resource Loss, Somatic Problems, Social Support, Coping, and Concerns (N = 248)
Downloaded by [University of Nebraska, Lincoln] at 17:38 13 April 2015
229
∗∗
p < .05 p < .01 ∗∗∗ p < .001
∗
Step 1 Gender (0 = male, 1 = female) Age Education Step 2: Prior experience Natural disaster Trauma Step 3 Life threat Step 4: Resource loss Personal characteristic Condition Energy Basic object Object Step 5 Somatic problems Step 6 Social support Step 7: Coping Problem focused Emotion focused Step 8: Concerns Future tsunami threat Government favoritism
Variable
0.28 1.38 1.59 1.30 1.11 0.90
2.92 2.69 7.02 0.59 1.95
1.20
4.67
2.12
2.90 1.27
−1.59 2.56
1.29 1.06 1.09 1.15 1.33
0.10 0.69
0.07 0.04
5.17 3.52 1.01 −0.60 2.71
2.31
SE B
3.28
B
.03 .11∗
.10 .29∗∗∗
.11∗
.41∗∗∗
.29∗∗∗ .21∗∗ .06 −.03 .13∗
.25
∗∗∗
−.04 .14∗
.04 .00
.10
β
.55
.54∗∗∗
.47∗
.46∗∗∗
.32∗∗∗
.07∗∗∗
.01
−.00
Adjusted R2
3 months post-tsunami (N = 248)
1.46 −0.00
0.92 0.66
3.50
2.20
4.07 1.96 4.06 1.95 −4.26
4.09
0.84 0.76
1.28 1.07
1.12
0.21
1.10 0.93 1.20 1.23 1.38
1.71
2.80 0.92
0.08 0.64
−0.03 −0.31 −9.72 4.33
1.98
SE B
0.84
B
.09 .00
.04 .03
.15∗∗
.50∗∗∗
.29∗∗∗ .13∗ .29∗∗ .13 −.29∗∗
.22
∗∗
−.22∗∗ .30∗∗∗
−.03 −.03
.03
β
.64
.64
.64∗∗
.62∗∗∗
.44∗∗∗
.18∗∗
.14∗∗∗
−.01
Adjusted R2
15 months post-tsunami (N = 255)
−.014 −.013
−.021 −.143∗∗∗
.047
.032
.006 −.001 .052 .119∗∗ −.136∗∗
−.023
−.034 .038
.024 .006
−.001
β
Time × Variable (N = 503)
TABLE 2 Predicting Posttraumatic Stress Symptoms with Demographics, Trauma Exposure, Resource Loss, Social Support, Coping, and Concerns
Downloaded by [University of Nebraska, Lincoln] at 17:38 13 April 2015
230
— −.08 −.17∗∗ .18∗∗ .02 −.03 .07 −.13∗
2
.24∗∗∗
.04
.38∗∗∗
.02
−.02
−.02
.03
.06
−.05 .02 −.02 −.04 −.07 −.02
.04
— .03 −.03 .11 −.14∗ .05 .08
3
.09
.50∗∗∗ −.02 .43 .00 .38∗∗∗ −.04 .68∗∗∗ −.10 .35∗∗∗ .39∗∗∗ .21∗∗∗ .40∗∗∗
.44∗∗∗ −.01
— −.06 .01 −.01 −.01 −.28∗∗∗ .31∗∗∗ .26∗∗∗ .53∗∗∗
1
Notes. PTS = posttraumatic stress symptoms. ∗ p < .05 ∗∗ p < .01 ∗∗∗ p < .001
PTS Posttraumatic growth Gender Age Education Disaster experience Trauma exposure Life threat concern Personal characteristic resource loss 10. Condition resource loss 11. Energy resource loss 12. Basic resource object 13. Object resource loss 14. Somatic problems 15. Social support 16. Problem-focused coping 17. Emotion-focused coping 18. Future tsunami concern 19. Government favoritism concern
1. 2. 3. 4. 5. 6. 7. 8. 9.
Variable
5
.12
.04
−.11
.00 .09 .14∗ −.05 −.24∗∗∗ −.18∗∗
.03
−.07
−.04
−.13∗
−.29∗∗∗ −.22∗∗∗ −.23∗∗∗ −.21∗∗∗ .03 −.03
−.20∗∗∗
−.14∗
−.16∗
7
−.06
−.07
.56∗∗∗
—
9
.38∗∗∗
.29∗∗∗
.21∗∗∗
.00
.25∗∗∗
.09 .63∗∗∗ .21∗∗∗ .65∗∗∗ .26∗∗∗ .64∗∗∗ .25∗∗∗ .31∗∗∗ .14∗ −.05 .21∗∗∗ −.01
.04
— .26∗∗∗
8
.18∗∗ .15∗
.16∗∗ .11 .20∗∗∗ .25∗∗∗ .19 .07
.15∗
— −.23∗∗∗ — −.16∗ −.09 −.20∗∗∗ .06
6
.16∗
−.04 −.10 −.12 −.05 .27∗∗∗ .20∗∗
−.02
— −.35∗∗∗ — −.03 −.01 −.13∗ .15∗ .13∗ −.11 .20∗∗∗ −.16∗
4
.02
−.10
.12
.54∗∗∗ .47∗∗∗ .46∗∗∗ .12∗ .13∗ .01
—
10
−.01
−.14∗
.18∗∗
— .69∗∗∗ .73∗∗∗ .33∗∗∗ .21∗∗∗ .06
11
.14∗
−.07
.07
— .76∗∗∗ .20∗∗∗ .09 .02
12
14
15
16
.06.
−.04
.14∗
17
.17∗
.07
.12
.08
.10
—
18 19
.28∗∗∗ .21∗∗∗ .33∗∗∗
.11
.37∗∗∗ .39∗∗∗ .50∗∗∗ —
— .27∗∗∗ — .10 .26∗∗∗ — .01 .13∗ .59∗∗∗ —
13
TABLE 3 15 Months Post-Tsunami: Correlations Among PTS, Posttraumatic Growth, Demographics, Traumatic Event Exposure, Resource Loss, Somatic Problems, Social Support, Coping, and Concerns (N = 255)
Downloaded by [University of Nebraska, Lincoln] at 17:38 13 April 2015
231
∗∗
p < .05 p < .01 ∗∗∗ p < .001
∗
Step 1 Gender (0 = male, 1 = female) Age Education Step 2: Prior experience Natural disaster Trauma Step 3 Life threat Step 4: Resource loss Personal characteristic Condition Energy Basic object Object Step 5 Somatic problems Step 6 Social support Step 7: Coping Problem focused Emotion focused Step 8: Concerns Future tsunami threat Government favoritism
Variable
.02 .09 .11 .09 .08 .07
.34 .24 .05
.14 −.04
.06
.10
.05
.15 .07
.09 .15
.08 .07 .07 .07 .08
.01 .04
−.00 .13
.03 .04 .03 .12 .00
.12
SE B
−.21
B
.11 −.04
.16∗ .04
.24∗∗∗
.19∗∗
.03 .04 .04 .12 .00
.10
.04 .15∗
−.02 .26∗∗∗
−.11
β
.18
.18
.17∗∗∗
.12∗∗
.10
.10
.09
.08
∗∗∗
Adjusted R2
3 months post-tsunami (N = 248)
−.08 −.06
.42 −.14
.07 .07
.11 .09
.10
.02
−.01 .55
.09 .07 .09 .10 .11
.08
.18 .06
.01 .04
.11
SE B
−.21 .05 .04 .07 .01
.10
.03 −.05
−.01 .07
.16
B
−.07 −.06
.29∗∗∗ −.11
.38∗∗∗
−.05
−.25∗ .06 .05 .08 .01
.09
.01 −.05
−.11 .13
.09
β
.19
.19∗∗
.15∗∗∗
.04
.04
.03
.03
.03
∗
Adjusted R2
15 months post-tsunami (N = 255)
−.087 −.020
.071 −.073
.051
−.116∗
.002 −.006 −.030 −.049 .059
.044
−.024 −.099∗
−.007 −.087
.106∗
β
Time × Variable (N = 503)
TABLE 4 Predicting Posttraumatic Growth with Demographics, Trauma Exposure, Resource Loss, Social Support, Coping, and Concerns
Downloaded by [University of Nebraska, Lincoln] at 17:38 13 April 2015
232
D. N. Sattler et al.
Downloaded by [University of Nebraska, Lincoln] at 17:38 13 April 2015
Posttraumatic growth was negatively associated with personal characteristic resource loss and positively associated with social support and problem-focused coping. Comparison at 3 and 15 months. Table 4 shows that the inclusion of the Time × Predictor interaction terms explained an additional 5.4% of the variance in posttraumatic growth above and beyond the main effects, F(17, 428) = 1.816, p = .024. The positive association between education and posttraumatic growth was weaker in the 15-month than in the 3-month sample. Although trauma was positively associated with posttraumatic growth in the 3-month sample, these factors were unrelated in the 15-month sample.
Comparison of Outcomes at 3 and 15 Months We performed independent samples t tests on the key outcome variables to compare the means across the 3-month and 15-month samples. Table 5 shows that scores were lower at 15 months than at 3 months for PTS, somatic problems, posttraumatic growth, resource loss, social support, problemfocused coping, belief in governmental favoritism, and concern about future tsunami threats.
DISCUSSION The results show that participants were exposed to unprecedented horror and loss of life and property and saw reminders daily of the devastation TABLE 5 Comparison of Variables at 3 and 15 Months 3 months (N = 248) Variable Posttraumatic stress Posttraumatic growth Natural disaster experience Traumatic experience Life threat/injury concern Personal characteristic resource loss Condition resource loss Energy resource loss Basic resource loss Object resource loss Somatic problems Social support Problem-focused coping Emotion-focused coping Future tsunami threat Government favoritism
15 months (N = 255)
M
SD
M
SD
df
t
p
d
37.11 4.95 1.81 0.62 2.66 1.70
1.02 0.57 0.39 0.93 0.92 0.95
26.44 4.75 1.86 0.58 2.62 1.74
1.02 0.50 0.35 1.05 0.78 1.06
495 499 500 498 497 501
7.38 2.49 −1.57 0.47 0.53 −0.27
.05 >.05 >.05
10.46 0.35
1.57 1.79 2.66 2.10 4.12 2.07 2.18 1.64 2.57 2.62
0.65 0.65 0.06 0.84 0.21 0.64 0.61 0.71 0.73 0.95
1.37 1.42 1.82 1.72 2.98 1.76 1.98 1.53 2.34 2.19
0.65 0.64 0.06 0.97 0.21 0.64 0.64 0.75 0.88 0.88
501 501 501 501 501 500 500 499 501 499
2.13 4.09 9.58 4.60 3.98 5.52 3.40 1.70 3.26 5.29
