• adolescent;
  • depression;
  • disasters;
  • grief;
  • post-traumatic stress disorder


  1. Top of page
  2. Aims

The aim of the present study was to examine the psychological impact on adolescent survivors of a maritime disaster that resulted in the deaths of nine people, including four high school students, and the effects of psychiatric intervention for the survivors.

Methods:  Long-term multidimensional intervention consisting of psychoeducation, hospital treatment, family support and day care, was provided for nine adolescent survivors. To evaluate these effects, the survivors were also assessed using self-rating scales (Impact of Event Scale, General Health Questionnaire and Self-rating Depression Scale) and psychiatric structured interviews (Clinician-Administered Post-Traumatic Stress Disorder [PTSD] Scale) at 2, 8, 14, 26, and 38 months after the accident.

Results:  Prevalence of PTSD among adolescent survivors was much higher than in adult survivors at the 2-month examination (78% vs 12%, respectively). Although the observed prevalence remained high until the 14-month examination, remarkable improvement occurred thereafter and none was diagnosed with PTSD at the 38-month examination.

Conclusion:  Adolescents may have a specific vulnerability to PTSD and community-based intervention is effective for adolescents with serious symptoms of PTSD.

SINCE THE INDUSTRIAL revolution of the late 18th century, transportation disasters, such as air, sea, and rail accidents, have affected many people. Several studies have found that transportation disasters can also provoke post-traumatic stress disorder (PTSD) or major depression among survivors.1 For example, investigations into maritime disasters have outlined that approximately two-thirds of survivors from two major car ferry accidents had symptoms of serious traumatic responses at 3–30 months after the accident.2,3 Furthermore, Yule et al., using standardized interviews, examined 217 young survivors of the Jupiter cruise ship accident and demonstrated that 51.7% had developed PTSD within 8 years after the accident.4 Compared with natural disasters, however, few longitudinal studies exist on the psychological influence of transportation disasters, due to the difficulty involved in performing a long-term study of the victims of transportation disasters, owing mainly to their dispersion after accidents.

The Ehime Maru disaster, which involved a Japanese training vessel, occurred in the Pacific Ocean in 2001 and resulted in nine deaths, four of which were of high school students. A total of nine high school students survived the accident and witnessed the deaths of their classmates. In order to examine the long-term influence of this disaster on the survivors, particularly the students, mental health care was provided for 26 survivors soon after the accident, and a longitudinal study was conducted. Some of the results of that study have previously appeared in a monograph,5 but more detailed data and analysis of the adolescent (students) and adult (crewmembers) survivors have not yet been presented. The aims of the present study were (i) to examine the psychological impact of the accident on the adolescent survivors (students), in comparison with the adults (crew members); and (ii) to examine the effect of psychiatric intervention, based on a community care system, on the adolescent survivors.


  1. Top of page
  2. Aims

The accident

The Ehime Maru, a Japanese training vessel, sailed from Japan on 10 January 2001 to the Hawaiian Islands, as a training excursion for the students of Uwajima Fisheries High School. In total 20 crew members (the captain, engineers, and navigators), two teachers and 13 students were aboard the ship. The Ehime Maru arrived in Hawaii after a 1-month-long voyage, without calling at any other port. Following a short vacation, the Ehime Maru set sail to return to Japan on 10 February. Soon after, the ship was hit by the nuclear submarine USS Greenville approximately 18 km off Honolulu, Hawaii, and sank quickly with 35 people on board. The Coast Guard rescued 26 people, but four high school students, two teachers, and three crew members were reported missing and presumed dead. Approximately 8 months after the accident, the wreck of the Ehime Maru was raised and towed closer to Oahu Island. On 14 October 2001 the wreck was set down in 35 m of water, and divers attempted to recover the bodies. On completion of the retrieval work, the Ehime Maru was towed back out to sea and scuttled on 25 November.


All survivors (nine adolescents, 17 adults) were individually examined at the health center after the accident. The adolescent survivors had been studying together for approximately 2 years since enrolling in the high school, and prior to the accident, had hoped to become seamen or work in a fishery. This was their first long-distance training excursion. All nine adolescent survivors participated in all of the examinations, but the adult survivors participated in only the 2-month examination because they dispersed around the country seeking work after the accident. At the first examination all adolescent survivors were 17 years old, while the mean age of the adult survivors was 45.9 ± 11.6 years.


Shortly after the accident, district nurses and psychiatrists belonging to the Uwajima Health Center and Uwajima Psychiatric Hospital in Uwajima City began the provision of intensive care for the adolescent survivors, in collaboration with their school teachers. At the same time two expert psychiatric teams, consisting of several psychiatrists and clinical psychologists from Kurume University Hospital and the Hyogo Institute for Traumatic Stress, were created (the authors were all team members). These teams assisted the activities of the local care staff, and also engaged in the assessment of the post-traumatic responses among the survivors and the evaluation of the effects of the care programs provided for them.

Three weeks after the accident the team members received a report concerning the mental states of the surviving students from the local staff. This report indicated that most of the students were very irritable or depressive, and tended to stay indoors. It was therefore necessary for us to provide adequate information about traumatic responses and appropriate psychiatric support for the survivors and their families as soon as possible. Thus, the following intervention programs were provided for the survivors, particularly all adolescent survivors and their families: (i) home visiting services for the survivors and their families (once every 2 or 3 weeks), provided by district nurses; (ii) intensive psychiatric treatment at the clinic (several of the adolescent survivors were also admitted to the psychiatric hospital due to an increased risk of suicide); (iii) group meetings for students' families, which were held once a week for approximately 2 years after the accident, in order to perform psychoeducation and facilitate the self-help function of the survivors and their family members; and (iv) day care programs designed specifically for the adolescent survivors (twice a week), carried out at the Uwajima Health center following graduation. Every staff member had considerable experience in care for people with major psychosis such as schizophrenia but no experience for those with severe post-traumatic responses.


Examinations were carried out at 2, 8, 14, 26, and 38 months after the accident. At each examination all participants completed the following three self-rating scales: the Japanese version of the Impact of Event Scale–Revised (IESR) to assess PTSD symptoms,6 the 28-item General Health Questionnaire (GHQ) to assess general health problems,7,8 and the Self-rating Depression Scale (SDS) to assess depressive symptoms.9,10 The recommended cut-off points for these self-rating scales, which were obtained from previous Japanese research, were 25 for the IESR, 48 for the SDS, and 6 for the GHQ.8–10 Furthermore, the participants were assessed and diagnosed by the outside psychiatric teams using two structured interviews: the Clinician-Administered PTSD Scale (CAPS)11,12 and the Mini-International Neuropsychiatric Interview (MINI).13,14 CAPS is a practical and reliable structured interview for assessing PTSD.15 MINI is also useful as a short, structured, diagnostic interview for DSM-IV and ICD-10 psychiatric disorders.


Because most of the survivors were clearly upset as a result of the accident, the Ehime prefectural government immediately decided to provide mental health care for all survivors, and asked the authors to cooperate in the creation of a care plan. Once arrangements had been made with local staff, the examinations described in the previous section were conducted with the survivors in order to clarify the severity of their traumatic responses. All participants, as well as the students' parents, were informed of the aim of this examination and provided their written consent prior to participation. In addition, respondents who were judged to be in need of urgent psychiatric help were referred to a local psychiatric clinic.

Statistical analysis

Mean scores on the IESR, SDS, GHQ, and CAPS at the 2-month examination were compared between the adolescent and adult survivors using an unpaired t-test. The difference between the number of adolescent survivors diagnosed with psychiatric disorders at the 2-month examination and the number of adult survivors diagnosed with psychiatric disorders at the 2-month examination was evaluated using Fisher's exact test. Furthermore, the relationship among the mean total scores on the IESR, SDS, GHQ and CAPS at the 2-month examination was investigated using Pearson's correlation coefficient. Because the number of adolescent participants was small, repeated measures analysis of variance (anova) was used, in conjunction with Dunnett's procedure, to compare the mean scores of IESR, SDS, GHQ, and CAPS at the 8-, 14-, 26-, and 38-month examinations with those at the 2-month examination. Statistical significance was established at the P < 0.05 level. When the sphericity assumption was not met, the Huynh-Feldt correction was applied. Data were analyzed using SAS version 9.13 for Windows (SAS Institute Japan, Tokyo, Japan).


  1. Top of page
  2. Aims

Care program participants

The adolescent survivors received various treatment and care programs, as illustrated in Fig. 1. Although it was recommended that most of the adolescent survivors visit the clinic, they initially hesitated. They changed their minds, however, after receiving psychoeducation from local staff through the home visiting service. Seven adolescent survivors underwent medication therapy (mainly antidepressants such as paroxetine and sertraline) and ordinary supportive psychotherapy at the clinic, and three were admitted to psychiatric hospital due to the aggravation of depressive symptoms. Since graduating from high school, eight adolescent survivors have attended the day care programs.


Figure 1. Care programs provided for the adolescent survivors. (—) Period during which more than half of the adolescent survivors participated; (- - -) period during which fewer than half of the adolescent survivors participated.

Download figure to PowerPoint

Comparison between adolescent and adult survivors

The mean scores of the adolescent survivors were high on all of the self-rating scales: IESR, 51.2; SDS, 51.5; GHQ, 13.1 (Table 1). All of the scores of the self-rating scales at first examination were far greater than the cut-off points. Regarding the data of the students, a highly significant positive correlation was observed between the total CAPS score and the three self-rating scales (IESR, GHQ and SDS) at the 2-month examination, with Pearson's correlation coefficients of r = 0.72, 0.99 and 0.85, respectively. In contrast, the mean scores of IESR, SDS and GHQ observed in the adult survivors were 23.0, 46.0 and 8.3, respectively, which were all significantly lower than those of the adolescent survivors (unpaired t-test, P < 0.001).

Table 1.  Students vs crew at 2-month examination
 Students (n = 9)Crew (n = 17) 
  • *

    P < 0.05;

  • **

    P < 0.001.

  • †Fisher's exact test;

  • unpaired t-test.

  • A, avoidance and numbing; CAPS, Clinician-Administered PTSD Scale; IESR, Impact of Event Scale–Revised; GHQ, 28-item General Health Questionnaire; H, hyperarousal; MDD, major depressive disorder; PTSD, post-traumatic stress disorder; R, re-experiencing; SDS, Self-rating Depression Scale.

Not PTSD215*
Not MDD316*
Mean age (years)17.0 ± 0.045.9 ± 11.6 
IESR51.2 ± 18.223.0 ± 20.4**
SDS51.5 ± 11.246.0 ± 8.2**
GHQ13.1 ± 8.68.3 ± 8.0**
CAPS: total80.0 ± 25.223.1 ± 22.2**
CAPS: R22.1 ± 8.67.6 ± 6.6**
CAPS: A30.2 ± 12.98.4 ± 7.5**
CAPS: H27.7 ± 7.96.9 ± 7.0**

Seven adolescent survivors (77.8%) were diagnosed with PTSD and six (66.7%) with major depressive disorder (MDD; Table 1). Furthermore, the mean total CAPS score of the adolescent survivors was extremely high (80.0), as were the mean scores for each of the three subscales of PTSD (re-experiencing, avoidance and numbing, and hyperarousal).

Regarding the adult survivors, only two (11.8%) were diagnosed with PTSD and one (5.9%) with MDD, significantly fewer than the adolescent survivors diagnosed with PTSD/MDD (Fisher's exact test, P < 0.05). The mean total CAPS score and the means of the three subscale scores observed among the adult survivors, as well as other rating scales, was also found to be significantly lower than those among the adolescent survivors.

Symptom outcome

Mean scores on all three self-rating scales declined from 2- to 38-month examination (51.2–12.5 for the IESR score, 51.5–42.9 for the SDS score, and 13.1–6.2 for the GHQ). As a result of repeated-measures anova, main effects were observed for the IESR (F(3,24) = 15.28, P < 0.005) and GHQ (F(3,18) = 3.27, P < 0.05), but not for SDS (F(2.01,14.09) = 2.65, P = 0.105). The post-hoc Dunnett's test indicated significant decreases in IESR scores between the 2-month examination and the 14-, 26- and 38-month examinations, and significant decreases in GHQ scores between the 2-month examination and the 26- and 38-month examinations.

Although the mean total CAPS scores indicated severe PTSD among the adolescent survivors until the 14-month examination, these scores improved over time (Fig. 2). The repeated measures anova indicated a significant main effect for CAPS (F(3,24) = 15.72, P < 0.005), and the post-hoc Dunnett's test showed significant decreases between the 2-month examination and the 26- and 38-month examinations. Results of CAPS and MINI interviews indicated that, at 14 months, seven adolescent survivors still had PTSD and four adolescent survivors still had MDD (Table 2). The prevalence of PTSD and MDD among the adolescent survivors, however, was remarkably reduced at the 26-month examination and none of the adolescent survivors was diagnosed with either PTSD or MDD at the 38-month examination.


Figure 2. Changes in Clinician-Administered Post-Traumatic Stress Disorder Scale (CAPS) after the accident. *Significant decrease compared to 2-month examination (P < 0.05, repeated measures ANOVA, post-hoc Dunnett's test).

Download figure to PowerPoint

Table 2.  Changes of prevalence among students
Time after the accident (months)PTSD n = 9MDD n = 9
  1. MDD, major depressive disorder; PTSD, post-traumatic stress disorder.



  1. Top of page
  2. Aims

Vulnerability of adolescent survivors

During the early stages of the present study a considerably high prevalence and symptom severity of PTSD were observed among the adolescent survivors. At the 2- and 14-month examinations 77.8% of the adolescent survivors were diagnosed with PTSD. To our knowledge the highest prevalence of PTSD in past studies of maritime disasters was 51.5% at 5 months, which was obtained using CAPS on the survivors of the Jupiter cruise ship disaster in Greek waters.4 The prevalence of PTSD among the adolescent survivors of the Ehime Maru accident was much higher than for the Jupiter accident, which demonstrates the massive effect that the Ehime Maru accident had on the adolescent survivors. Furthermore, high comorbidity of PTSD and MDD was found in the students, and the results of CAPS and SDS total score indicate that severity of PTSD symptoms strongly correlated with those of depressive symptoms. These findings support the common view that people who are diagnosed with PTSD also often have MDD.16,17

In contrast, the prevalence of PTSD and MDD among the adult survivors was remarkably lower than that of the adolescent survivors, at 11.8% and 5.9%, respectively. Although students and crew members died as a result of this accident (as did some teachers) and all of the survivors narrowly escaped death, the accident seems to have had a surprisingly different effect on the adult and adolescent survivors. This suggests that, compared with adult survivors, adolescent survivors are somewhat more vulnerable to PTSD and MDD.

We speculate that there are two major reasons why such a large difference between the post-traumatic responses of adolescent and adult survivors occurred. First, the difference might be a result of the amount of experience at sea. That is to say, the crew's preparedness for a maritime accident might have mitigated their post-traumatic responses.

Another possible reason for the disparity in the responses is that the undefined identity of the adolescent survivors could have played an important role in the development of their serious post-traumatic responses. It is important for adolescents to acquire self-certainty and establish various identities. They also have a strong tendency to gather in gender-specific groups.18 In a related study by van der Kolk, it was found that the prevalence of PTSD among 18-year-old soldiers engaged in the Vietnam War was much higher than that of older soldiers.19 Young soldiers reacted more violently to the death of a friend, which caused them uncontrollable grief and rage. The responses of the adolescent survivors of the Ehime Maru accident are more likely to resemble the responses of these young soldiers than the passengers of a wrecked ship. The adolescent survivors, like the young soldiers, formed a strong group identity through the dedicated guidance and training of the teachers and the crew amid the harsh winter conditions of the Pacific Ocean.

Effects of intervention

A large body of studies has demonstrated that PTSD symptoms are likely to continue for a long time. In the general population, Kessler et al. found that more than one-third of people with an index episode of PTSD failed to recover even after a period of approximately 4 or 5 years.16 In a recent longitudinal epidemiological study in Germany, 48% of young adults with PTSD showed no significant remission of PTSD symptoms during the follow-up period (34–50 months).20 Furthermore, several investigations into the long-term outcome of PTSD following a war or disaster have demonstrated that survivors with PTSD suffered various symptoms of PTSD for many years.21–23 With regard to maritime accidents, Yule et al. also reported that approximately one-third of the young survivors of the Jupiter accident continued to suffer symptoms of PTSD 5–8 years after the accident.4

Although these studies demonstrated prolongation of PTSD symptoms, the adolescent survivors of the Ehime Maru accident had a different outcome. The scores for the self-rating scales and CAPS among these survivors were extremely high until the 8-month examination. A significant decrease, however, in all scores, except for the SDS, was observed in the months that followed. Furthermore, the high prevalence of PTSD and MDD among the adolescent survivors also remarkably decreased by 3 years after the accident.

We consider that the intervention was successful not as a result of one specific treatment, but as a result of the functional combination of several treatments or care programs provided to the survivors. When the accident occurred there were only a few experts within the psychological field in Uwajima City and, moreover, the local staff did not have enough experience to be involved in psychological care for the survivors of a major disaster. Therefore, the traditional community care system, which had been already established for the mentally disabled (e.g. people with schizophrenia), was highly exploited. Regardless of the limited resources, the community-based intervention (day care, visiting services and family support programs) was considerably effective in improving the social withdrawal of the adolescent survivors, and eliciting the self-help function of the survivors and their families.

The results suggest that a traditional community care system can be effective even for survivors with severe post-traumatic responses. To achieve such a successful outcome, however, we believe that four essential factors are needed. First, some experts should substantially participate in the planning of the intervention. Second, good relationships should be established between such facilities as schools, hospitals, and political offices. Third, the intervention programs should be adequately modified for involving as many survivors as possible. Finally, the intervention programs should be provided continuously. This last speculation corresponds with Shalev's opinion that treatment should be provided in the context of continuity of care, and clinical decisions should be made on the basis of longitudinal observations.24

In contrast, there might be other factors, such as bonds in the community or school, exposure to mass media, or negotiations for compensation, that affected the outcome of the adolescent survivors. We were unable, however, to more carefully consider these factors due to two apparent limitations of the study. First, there was no comparison with a group that had the same demographic profile as the adolescent survivors. In addition, the sample size of the present adolescent survivors was small.

In spite of these limitations, the present results indicate that good outcomes can be achieved through not only specific treatment, such as cognitive psychotherapy, but also the combination of standard treatment programs. Because of a number of actual limitations at the time of a disaster, there is a great need for further studies to explore the effectiveness of traditional community care systems.


  1. Top of page
  2. Aims

We wish to acknowledge the great efforts of the staff of the Uwajima Central Health Center. In particular, we thank Tatuyuki Teramoto MD and Naoto Takenouchi MD for their leadership in the present intervention. Statistical advice and analysis were provided by Assoc. Professor Masahiro Haraguchi of the Department of Psychology, Kurume University. Finally, we gratefully acknowledge all survivors of the Ehime Maru accident who willingly participated in the present study.


  1. Top of page
  2. Aims
  • 1
    Lundin T. Transportation disasters: A review. J. Trauma. Stress 1995; 8: 381389.
  • 2
    Joseph SA, Yule W, Williams RM, Hodgkinson P. Correlates of post-traumatic stress at 30 months: The herald of free enterprise disaster. Behav. Res. Ther. 1994; 32: 521524.
  • 3
    Eriksson NG, Lundin T. Early traumatic stress reactions among Swedish survivors of the m/s Estonia disaster. Br. J. Psychiatry 1996; 169: 713716.
  • 4
    Yule W, Bolton D, Udwin O, Boyle S, O'Ryan D, Nurrish J. The long-term psychological effects of a disaster experienced in adolescence: I. The incidence and course of PTSD. J. Child Psychol. Psychiatry 2000; 41: 503511.
  • 5
    Maeda M, Maruoka T, Maeda H. Psychological consequences for students who survived the Ehime Maru accident: A 26-month follow-up study. In: KatoN, KawataM, PitmanRK (eds). PTSD: Brain Mechanisms and Clinical Implications. Springer-Verlag, New York, 2006; 193202.
  • 6
    Asukai N, Kato H, Kawamura N et al. Reliability and validity of the Japanese-language version of the impact of event scale-revised (IES-R-J): Four studies of different traumatic events. J. Nerv. Ment. Dis. 2002; 190: 175182.
  • 7
    Goldberg DP, Hillier VF. A scaled version of the General Health Questionnaire. Psychol. Med. 1979; 9: 139145.
  • 8
    Fukunishi I. Cut-off point of Japanese version of General Health Questionnaire. Shinri Rinsho 1990; 3: 228234 (in Japanese).
  • 9
    Zung WW. A self-rating depression scale. Arch. Gen. Psychiatry 1965; 12: 6370.
  • 10
    Fukuda K, Kobayashi S. A study on a self-rating depression scale. Psychiatr. Neurol. Jap. 1973; 75: 673679 (in Japanese).
  • 11
    Blake DD, Weathers FW, Nagy LM et al. The development of a Clinician-Administered PTSD Scale. J. Trauma. Stress 1995; 8: 7590.
  • 12
    Asukai N, Hirohata S, Kato H et al. Psychometric properties of the Japanese-language Version of the Clinician-Administered PTSD Scales for DSM-IV. J. Trauma. Stress 2003; 1: 4753.
  • 13
    Sheehan DV, Lecrubier Y, Sheehan KH et al. The Mini-International Neuropsychiatric Interview (M.I.N.I.): The development and validation of a structured diagnostic psychiatric interview for DSM-IV and ICD-10. J. Clin. Psychiatry 1998; 59: 2233.
  • 14
    Otsubo T, Tanaka K, Koda R et al. Reliability and validity of Japanese version of the Mini-International Neuropsychiatric Interview. Psychiatry Clin. Neurosci. 2005; 59: 517526.
  • 15
    Weathers FW, Keane TM, Davidson JR. Clinician-administered PTSD scale: A review of the first ten years of research. Depress. Anxiety 2001; 13: 132156.
  • 16
    Kessler RC, Bromet E, Hughes M, Hughes M, Nelson CB. Posttraumatic stress disorder in the National Comorbidity Survey. Arch. Gen. Psychiatry 1995; 52: 10481060.
  • 17
    Brady KT, Killeen TK, Brewerton T et al. Comorbidity of psychiatric disorders and posttraumatic stress disorder. J. Clin. Psychiatry 2000; 61: 2232.
  • 18
    Erikson EH. Identity and the Life Cycle. W. W. Norton, New York, 1980.
  • 19
    Van Der Kolk BA. Adolescent vulnerability to posttraumatic stress disorder. Psychiatry 1985; 48: 365370.
  • 20
    Perkonigg A, Pfister H, Stein MB et al. Longitudinal course of posttraumatic stress disorder and posttraumatic stress disorder symptoms in a community sample of adolescents and young adults. Am. J. Psychiatry 2005; 162: 13201327.
  • 21
    Kulka R, Schlenger WE, Fairbank JA et al. Trauma and the Vietnam War Generation: Report of the Findings from the National Vietnam Veterans Readjustment Study. Brunner and Mazel, New York, 1990.
  • 22
    McFarlane AC, Papay P. Multiple diagnosis in posttraumatic stress disorder in the victims of a natural disaster. J. Nerv. Ment. Dis. 1992; 180: 498504.
  • 23
    Grace MC, Green BL, Lindy JD, Leonard AC. The Buffalo Creek disaster: A 14-year follow-up. In: WilsonJP, RaphaelB (eds). International Handbook of Traumatic Stress. Plenum Press, New York, 1993; 441449.
  • 24
    Shalev AY. Treating survivors in the immediate aftermath of traumatic events. In: YehudaR (ed.). Treating Trauma Survivors with PTSD. American Psychiatric Publishers, New York, 2002; 157188.