Somatic Therapy Treatment Effects with Tsunami Survivors

Published in Traumatology, Vol. 14, No. 3, September 2008.
The journal version of the article can be ordered in electronic or hard copy format from
Somatic Therapy Treatment Effects with Tsunami Survivors
Catherine Parker, Ronald M. Doctor, and Raja Selvam1
This is an uncontrolled field study of the outcome effects of a somatically based therapy with tsunami victims in southern India. One hundred and fifty (150) participants, prescreened for trauma symptoms, received 75 minutes of somatic therapy and training in affect modulation and self-regulation. The results indicate a reliable and significant treatment effect at immediate, 4-week, and 8-month follow-up assessments. At the 8-month follow-up, 90% of participants reported significant improvement or being completely free of symptoms of intrusion, arousal, and avoidance. The results support the effectiveness and reliability of this modified version of Somatic Experiencing Therapy in working with trauma reactions and invite future controlled trials of this therapy .
Keywords: posttraumatic stress; somatic therapy; somatic experiencing therapy; IES; post-tsunami symptoms
1From the Department of Psychology, California State University, Northridge, California (CP, RMD); Pacific Graduate Institute, Carpinteria, California (RS).
Contributions to this research among the authors were equal.
Address correspondence to: Ronald M. Doctor, PhD, Department of Psychology, California State University, Northridge, 18111 Nordoff Street, Northridge, CA 91330; e-mail:
The authors wish to express their gratitude to members of the 2005 Trauma Vidya international trauma team to India: Emil Borgir, Daniel Bruce, Rosemary Carpendale, Alexandre Duarte, Giselle Genillard, Lisa LaDue, Beth Nielsen, Jacqueline Ramirez, Lucia Ribas, Jeanne Du Rivage, Lida Ruiter, and Ronnie San Jose. The authors also thank the members of the follow-up research team in India: K. Lakshmanan and L. Jeyanthi from Auroville.
This is a study examining the durability of a somatically based treatment approach to trauma on survivors of the devastating tsunami that hit southern India in the Tamil Nadu region in 2004. Catastrophic natural events such as this one produce high levels of terror, fear, despair, and loss, much of which eventually lead to the development of posttraumatic stress disorder (PTSD). PTSD is among the most common psychological reactions in survivors of disasters (Reyes & Elhai, 2004), but those who survive are also at risk for developing a variety of health problems and other psychological disorders, such as anxiety, depression, and substance abuse (American Psychological Association, 2006). In addition, studies suggest that PTSD has other serious and lasting effects, which include nonspecific stress, chronic problems in relationships, psychosocial resource losses, and problems specific to youth, such as separation anxiety, violence, and eating disorders (Norris, Perilla, & Murphy, 2001; Reyes & Elhai, 2004). 2
The rates of PTSD among survivors can vary significantly depending on the sample studied and the type of disaster (van der Kolk, McFarlane, & Weisaeth, 1996). But we know that situations where sudden, unpredictable life-threatening events pervade a community (such as hurricanes, tsunami, violence or floods, etc.) are likely to produce high levels of PTSD (Karamustafalioglu et al., 2006). For example, individuals involved in the Buffalo Creek disaster, where a dam broke and flooded an entire community, had a 59% incidence of PTSD symptoms among survivors and a lifetime rate of 25% symptomatic at a 14-year follow-up. Likewise, participants in the Vernberg, LaGreca, Silverman, and Prinstein (1996) study of factors predictive of PTSD symptomotology following Hurricane Andrew revealed effects on all five identified factors used to predict or mitigate PTSD development: frightening event, loss/disruption, social support, coping strategies, and supportive social environment. The hurricane caused most of these factors to occur, and, as expected, 86% of the children studied reported mild PTSD symptoms whereas 55% reported moderate to severe levels of symptoms. The tsunami that spread across Southeast Asia, India, and Africa was one of the deadliest natural disasters in history. The result was a death toll of more than 320,000 individuals (Bronisch et al., 2006). In studies of survivors, the following symptoms have been noted: dissociation, hyperarousal, flashbacks, sleep disturbances, illusions, loss of appetite, grief, suicidality, and difficulty concentrating (Bronisch et al., 2006).
According to Somatic Experiencing Therapy, a neurobiology-based somatic approach to working with trauma, trauma resides in the nervous system and not in the event itself (Heller & Heller, 2004). Lower brain centers become engaged during threatening events and executive functions become less active. A dominant neurological reaction occurs involving orienting, fight, flight, or freeze reactions. These reactions are conditioned to aspects of the life-threatening event and subsequent exposure to similar events can trigger an involuntary portion of the terror reaction in the body (Ledoux, 1996). Bodyoriented approaches, such as Somatic Experiencing Therapy (Levine & Frederick, 1997), attempt to gain careful access to these involuntary responses, build awareness of the bodily reactions, and actually “process” them to an “adaptive resolution.” Descriptive and subjective data collected in Thailand on tsunami survivors using Somatic Experiencing Therapy showed that immediately after this therapy 67% of the participants reported partial or complete remission and 90% reported partial or complete improvement at a 1-year follow-up (Leitch, 2007).
With this in mind, we set about using a modified form of Somatic Experiencing Therapy with survivors of the southern India area of Tamil Nadu where devastation was enormous. This study is the report of that effort. A more detailed description of the project may be obtained online (Selvam, 2005).

The participants were 204 volunteers from 13 fishing villages in 3 districts (Nagapattinam, Cuddalore, and Vilupuram) in Tamil Nadu. The final count, however, for these analyses was 150 participants, because the Vilupuram district did not receive all follow-up assessments and therefore was not included here. Of the 150 participants, 40 were men and 110 were women, with a mean age of 41.6 years. Participants had responded to notices and advertisements in local venues asking for volunteers who had been affected by the tsunami and wanted treatment. Each 3
trauma team member met individually with each participant for approximately 75 minutes.
Prior to treatment contact, each participant was asked to complete a brief set of assessment questionnaires. These questionnaire responses, before and after treatment, and at 4-week and 8-month follow-ups, constitute the data for this study. First, a 17-item Post-Tsunami Symptom Checklist was completed for symptoms that occurred after the tsunami. The Post-Tsunami Symptom Checklist items identified possible emotional and bodily stress reactions that had been gleaned from reports by tsunami survivors from nearby villages. Reliability measures were not taken on this list, but we assume there was good face and construct validity because of the nature of the item selection. Responses to this checklist were coded as dichotomous yes or no answers. Second, 5 of the 17 checklist items were taken from the Impact of Events Scale–Revised–Abbreviated (IES–R–A). These items provided a separate measure closely linked to the Diagnostic and Statistical Manual of Mental Disorders (fourth edition, text revision; American Psychological Association, 2004) main criteria for PTSD (see Horowitz, Wilner, & Alvarez, 1979) in terms of intrusion, hyperarousal, and avoidance criteria. Only volunteers who scored (i.e., said yes to) more than 8 of the 17 items were taken as treatment participants and assumed to be experiencing trauma-related symptoms and assumed to be under personal stress at the time of treatment. Third, a Presenting Post-Tsunami Symptoms completion task asked each participant to list up to three presenting symptoms from which they wanted relief from and whether they had received any medical attention for the same. Almost all tsunami survivors treated had received medical attention for their symptoms but without lasting relief to their symptoms. Because the number of presenting symptoms identified varied from one to three for each participant, the symptoms were combined and averaged into one score for each participant. Fourth, Overall Stress Improvement (change in overall stress from the tsunami), Presenting Symptoms, and IES–R–A items were measured with a modified Subjective Units of Distress (SUD) scale after the treatment and at each follow-up (Wolpe, 1990). The modified SUD units in this study, because of the nature of the population, ranged from 0 to 4, with 0 =worse, 1 =same, 2 =somewhat better, 3 =a lot better, and 4 =completely well. Additionally, at the 8-month follow-up alone, participants were asked to judge percentage changes in Overall Stress Improvement, Presenting Symptoms and IES–R–A items using 0 to 100 paises (in local currency, 100 paises equal 1 rupee). The Overall Stress Improvement scores were treated as a self-report global measure of overall change since treatment in their capacity for self-regulation and were treated as separate from changes in specific symptoms. At only the 8-month follow-up participants were asked to indicate their type of family, marital status, education, occupation, and types of loss experienced during the tsunami.
Design and Procedure
All participants received a modified1 form2 of Somatic Experiencing Therapy treatment from therapists trained for 3 years or more in this work. In all, 11 therapists participated, and assignment of respondents was on a random basis. Name, age, and sex of the participant were coded, but more personal questions were considered intrusive and insensitive in this culture and were not asked at this first session. A four-stage treatment protocol was developed for this single treatment intervention, and all therapists adhered to the protocol as judged by supervisor observations and therapist reports. The protocol consisted of four distinct stages. The first stage focused on containment of physiological arousal associated with tsunami and the aftermath experiences. Containment involved increasing awareness of bodily arousal reactions while, at the same time, building feelings of being grounded and connected and safe with the therapist. Bodily reactions were purposely slowed to avoid anxiety bursts. This first phase was emphasized in the treatment and participants were asked to practice it after the treatment session on a regular basis. Second, narratives were built of the tsunami experiences for what was remembered, felt, what 4
they did or did not do, and, most important, what they sensed in their bodies in the remembrances. “Pendulation” of awareness between being grounded and being safe and the terror-survival responses in their bodies was established on a gradual basis (Levine & Frederick, 1997). Pendulation eventually allowed for processing or release of these involuntary body reactions associated with the event. Third, participants were educated (as they pendulated) on the neurophysiology of stress and trauma in a simple manner that emphasized how the body stores experience and does not release it until the conditions are proper. Education was also aimed at normalizing the experience away from judgments and pathological categorizations and at emphasizing the need to practice regularly after the treatment experience. Finally, tracking of bodily experience helped guide participants toward discharge and better self-regulation of the body experience and state. Tracking the changes helped them see actual resolution of body-held experiences as a standard for future more stress-free behavior.
At each follow-up session, participants were asked to rate changes in Presenting Post-Tsunami Symptoms identified at the first session on a 5-point SUD-type scale previously described. Changes since the treatment in the items on the IES–R–A scale as well as their global Overall Stress Improvement were also rated on the 5-point scale at each follow-up. Two Indian field workers, trained in scale administration, conducted each follow-up session on an individual basis with participants. All questionnaire items and instructions had been translated into Tamil prior to data collection so that comprehension was not a problem.
The data were collapsed into three broad categories related to Degree of Loss and Degree of Traumatization for purposes of data analysis. Degree of Loss was partitioned into low, medium, and high ranges with low being 1 (loss of family, property, financial status, injury, or witness to significant loss), medium being 2 or 3, and high being 4 or 5 types of loss. Degree of Traumatization was determined by the number of yes symptoms checked on the Post-Tsunami Symptom Checklist. Here low represented 8 to 10 symptoms checked, medium represented 11 to 12, and high represented 13 or higher. Thus, two new variables were created, Degree of Traumatization and Degree of Loss.
The three primary dependent variables in this study were the Presenting Post-Tsunami Symptoms scores, Overall Stress Improvement scores, and the IES–R–A scores. Using SPSS, a mixed design analysis of variance (ANOVA) was performed for each dependent variable to determine changes during three assessment periods (immediate, 4 weeks, and 8 months). The grouping variables were by district, gender, Degree of Loss, and Degree of Traumatization. Assumptions regarding normality of sampling distributions, homogeneity of variance–covariance matrices, linearity, and multicollinearity were met in all cases.
The Time (of assessment) effect on Presenting Post-Tsunami Symptom scores was significant, F(2, 118) =14.42, p <.05, partial η2 =.20, indicating reliable improvement during the three follow-up assessment periods. Likewise, there was a significant district-by-time effect on Presenting Post Tsunami Symptom scores. Cuddalore improved at a greater rate than Nagapattinam, although both showed significant improvement. The test for gender effects was not significant and Degree of Loss and Degree of Traumatization scales did not show differential effects over Time although all levels of these variables improved significantly over Time. A mixed-design ANVOA was performed on the IES–R–A ratings at the 4-week and 8-month follow-ups but not the immediate post-treatment assessment. The post-treatment assessment was not considered an appropriate and meaningful measure because participants could not report on experiences that they had not yet had (such as sleep difficulties, etc.). District, gender, and Degree of Loss were used as grouping variables. Degree of Traumatization contained the IES–R–A items 5 so it was excluded from this set of analyses. There was a significant Time effect on IES–R–A scores, indicating reliable improvement over Time. However, district, gender, and Degree of Loss were not associated with IES–R–A score improvement over Time. The SUD ratings measuring Overall Stress Improvement produced significant effects by district, with Cuddalore showing greater improvement than Nagapattinam, F(1, 148) =8.64, p <.05, partial η2 =.06. Sex of the participant was not significant with Overall Stress Improvement over Time but Degree of Loss did show an interactive effect, F(2, 294) =8.26, p <.05, η2 =.05, with high and medium scores making greater improvement than low scores on Degree of Loss. Degree of Traumatization did not interact with Time but Overall Stress Improvement showed significant and sustained improvement over Time, F(2, 146) =6.88, p <.05, η2 =.09. Improvement Ratings
In Table 1, the frequency distributions of self-reported changes on Presenting Post-Tsunami Symptoms across participants at the 4-week and 8month follow-ups are presented. At 4 weeks, the participants reported that 74.2% of the presenting symptoms were somewhat better, a lot better, or completely resolved. At 8 months, the corresponding percentage was higher at 85.2%. In Table 2, the frequency distribution of percentage changes in the variable Overall Stress Improvement (improvement in overall stress from the tsunami), measured only during the 8-month follow-up, is presented. A total of 94.4% of the respondents at the 8-month follow-up reported 50% or more improvement in their overall level of stress from the tsunami.
Table 1. Participant Ratings on Improvement on Post-Tsunami Symptoms on Modified SUD at 4-Week and 8-Month Follow-up
Rating Category 4 Weeks 8 Months
0 =Worse 5.6% 10.7%
1 =The same 20.3% 4.1%
2 =Somewhat better 28.0% 16.4%
3 =A lot better 24.5% 41.8%
4 =Completely well 21.7% 27.0%