Rates of Clinically Confirmed Stress-related Psychiatric Disorders Among Swedish Tsunami Survivors: 9-year Follow-up

Abstract

To the Editor: Our register-based follow-up study of Swedish survivors of the 2004 Southeast Asian tsunami published in Lancet Psychiatry revealed an increased risk of stress-related psychiatric disorders among adult survivors, compared to matched unexposed Swedes, throughout the 5 years after the disaster.1 Very little is known about the temporal pattern of stress-related psychiatric disorders among survivors beyond the first years after the event2,3 and we therefore conducted an additional follow-up of adult survivors for up to 9 years after the tsunami. We used the Swedish Tsunami Cohort1 consisting of 8,762 adult survivors (age ≥ 18 years in 2004) with permanent residence in Sweden who were confirmed alive through the register of the state police services in the aftermath of the tsunami. The unexposed group comprises 864,088 adults matched on sex, birth year, cohabitation status, and socioeconomic status. We retrieved clinical diagnoses of stress-related disorders (ICD-10: F43) from the Swedish Patient Register, which includes diagnoses made in specialized healthcare. Using flexible parametric models, we estimated hazard ratios (HRs) and 95% confidence intervals (CI) of stress-related disorders over the 9-year follow-up time, adjusting for pretsunami psychiatric disorders. The study was approved by the Regional Ethics Committees of Karolinska Institutet and Uppsala University. After register-linkage on national identifiers, unique for every Swedish resident, data analyses were done on de-identified datasets. Over the 9-year follow up, 3.6% (n = 313) of the Swedish adult tsunami survivors and 2.2% (n = 19,080) of the unexposed group received their first registered stress-related disorder diagnosis, with a corresponding elevated risk of stress-related disorders (3.56 vs. 2.19 per 1000 person-years, adjusted HR 1.72, 95% CI: 1.54, 1.93), highest for posttraumatic stress disorder (0.69 vs. 0.16 per 1000 person-years; adjusted HR 5.05, 95% CI: 3.89, 6.55). The results were similar across sexes. To assess temporal patterns, we performed an additional analysis in which we estimated separate HRs for different follow-up periods. The relative risk for stress-related disorders was substantially elevated during the first year posttsunami, decreased in magnitude thereafter, and no increased risk could be detected from 5 years and onwards (adjusted HR 1.04, 95% CI: 0.86, 1.26) (Figure 1).FIGURE 1.: Hazard ratios and 95% CIs of stress-related disorders among adult tsunami survivors as compared with matched unexposed individuals, adjusted for pre-event psychiatric history. We estimated hazard ratios from flexible parametric survival models.Our study indicates that after a transient major disaster, the dramatic rise in incidence of severe stress-related psychiatric disorders among survivors returns to unity around 5–6 years after the event. The temporal pattern of stress-related disorders in our study seems similar to research in populations suffering other types of life stressors, such as a diagnosis of a life threatening illness.4 While these results suggest that new onsets of stress-related psychiatric disorders among disaster survivors return to baseline after 5 years, we note that our previous studies of a subset of the tsunami cohort suggest that 16% of survivors are experiencing self-reported PTSD symptoms in a clinically significant range beyond 5 years after the disaster.5 Indeed, many disaster survivors with psychiatric morbidity defer seeking out treatment6 and may, as time goes by, instead seek treatment for other comorbid psychiatric disorders, such as substance abuse, sleep disturbances, anxiety, and depression. Due to the socio-economic characteristics of the sample and the particularly benign post-disaster context, our results may reflect a conservative estimate of the increased incidence of severe stress-related disorders seen in specialized or inpatient healthcare only. ACKNOWLEDGMENTS Authors’ contributors: E.B.T., H.S., and U.A.V. designed the study. E.B.T. and U.A.V. drafted the article, H.S. conducted the statistical analyses, and H.S., E.B.T. and U.A.V. had full access to all data and ensured their integrity. U.A.V. supervised the study. All authors took part in data interpretation and critical revision of the article. With approval from all authors, E.B.T. and U.A.V. had the final decision to submit.