Long-term Trends in Mental Health Disorders After the 2011 Great East Japan Earthquake and Tsunami

Key Points Question What is the persistence of mental illness symptoms among individuals affected by the 2011 Great East Japan earthquake and tsunami 5.5 years after the disaster? Findings In a cohort study of 2781 community-dwelling older adults, approximately half of the individuals who reported posttraumatic stress symptoms and depression after the disaster had recovered by 5.5 years. The overall prevalence of depression remained stable between predisaster and postdisaster data. Meaning In this study, although mental illness symptoms persisted for more than 5 years among half of disaster survivors, the community-wide prevalence of depression remained unchanged, suggesting that the community itself was resilient.


Introduction
It is well established that disasters are associated with an increase in many psychological problems, including depressive symptoms and posttraumatic stress symptoms (PTSS). 1 However, the long-term prognosis of mental disorders following disaster remains sparsely documented. Few studies have followed survivors for more than 3 years. 2,3 Even fewer studies have managed to capture the mental health status of survivors predating their disaster experiences, 2,3 so we cannot determine whether the disaster caused worsening of mental health or whether poor mental health already predated disaster exposure in affected populations.
Absent a predisaster assessment of mental health status, accurate estimates of the incidence of mental illness are difficult to obtain. 4 Asking survivors to retrospectively recall their mental health prior to disaster is obviously hampered by recall bias. In 1 of the few existing studies to document predisaster mental health, Fergusson et al 5

followed survivors of the Canterbury earthquakes in New
Zealand during 2010 and 2011. The researchers used the Christchurch Health and Development Study, a 35-year longitudinal birth cohort (with 635 men and 630 women), in which predisaster mental health information was available among survivors. The study found that the prevalence of some types of mental disorders (ie, major depression, posttraumatic stress disorder, other anxiety disorders, and nicotine dependence) was significantly elevated among people who experienced traumatic events during the earthquakes. However, after adjusting for confounders that were assessed prior to the earthquakes, the excess risk of mental disorders stemming from disasterrelated experiences was substantially attenuated, with the exception of nicotine dependence.
Turning to the issue of long-term prognosis of mental illness in the aftermath of a disaster, Morina et al 3 conducted a systematic review of remission rates from posttraumatic stress disorder in adults. Among 42 studies with a mean follow-up period of 40 months, they found that posttraumatic stress disorder following disaster-related trauma had the highest mean remission rate (60.0%) compared with remission of symptoms related to physical disease (31.4%). 3 More specifically, 5 natural disaster-related studies were found, with an mean period between the disaster event and follow-up assessment of 21 months (the range for the period between trauma and baseline assessment was 1.5-6.5 months, and the range for the period between baseline assessment and follow-up was 12-30 months.). 3 However, none of those studies assessed mental health status prior to the disaster. [6][7][8][9][10] According to the findings from a seminal review by Norris et al, 11 psychological symptoms peak in the first year after the disaster and tend to improve as time passes, while a delayed onset of the disorder is less common. Twelve years after Hurricane Katrina, Raker et al 12 found that 39% and 29% of people had recovered from PTSS and depressive symptoms, respectively, while 3% and 9% experienced delayed onset of PTSS and incident depressive symptoms. On the other hand, 14% and 28% had persistent PTSS and depressive symptoms even 12 years after the disaster. 12 The factors associated with persistence vs remission of mental illness symptoms remain poorly understood, especially with regard to risk and protective factors that predate the disaster experience. Iwanuma city in Miyagi prefecture was 1 of the localities directly affected by the tsunami on March 11, 2011. The Iwanuma Study is a community-based longitudinal cohort established in 2010 that has collected information every 3 years. A unique feature of the Iwanuma Study is that baseline information from participants was assessed 7 months before the disaster and the follow-up rate among survivors remains very high. In a series of previous reports, [13][14][15][16] we documented that levels of social capital predating the disaster were associated with a lower incidence of PTSS and depressive symptoms. However, our previous reports involved relatively short observation periods (approximately 2.5 years postdisaster) while most of the survivors were still housed in temporary accommodation. In the present study, we sought to examine long-term prognosis of PTSS and depressive symptoms among older individuals who experienced the 2011 disaster.

Study Population
The Iwanuma Study is a community-based longitudinal cohort established in 2010, when the city of Iwanuma in Miyagi prefecture (population 44 187 in 2010) was selected as 1 of 31 field sites for a nationwide cohort study of healthy aging in Japan, called the Japan Gerontological Evaluation Study (JAGES). 17 The baseline of the Iwanuma Study was collected in August 1 to 20, 2010, when a census was undertaken of all the city's residents aged 65 years or older. The names and addresses of residents were obtained from municipal records, and the baseline postal survey inquired about sociodemographic information, health habits, health status, and people's social connections to others. The response rate to the baseline survey was 59.0% (5058 of 8576 residents), which is comparable to similar surveys of community-dwelling residents. Comparison of the baseline characteristics of our sample with the general population of the city (assessed from the local government census) confirmed that the Iwanuma Study participants were representative with respect to sex and age distribution as well as employment status. 14 As summarized in the flowchart (Figure 1)

Depressive Symptoms
Depressive symptoms were measured by the Geriatric Depression Scale Short Form, which was developed to measure depressive symptoms among older adults. 19,20 The form has 15 items with binary answers (ie, yes/no) (Box Responses were summed (range, 0-15) to assess depressive symptoms, with higher scores indicating more severe depressive symptoms. This scale was validated to screen major depressive symptoms (ie, a score of 5 or more) with an area under the receiver operating characteristic curve of 0.94 (sensitivity, 92%; specificity, 87%) against the Structured Clinical Interview for the Diagnostic and Statistical Manual of Mental Disorders, Third Edition, Revised, as the criterion standard. 21 This measure has been validated to assess depressive symptoms among older Japanese populations. [13][14][15]

Trajectories of PTSS and Depressive Symptoms
Following Norris et al 11

Statistical Analysis
First, we examined the overall community prevalence of mental illness symptoms at each survey. Do you think that most people are better off than you are? a These questions were reverse coded, ie, a lower score indicated more severe depressive symptoms.

Results
The analytic samples for trajectories of PTSS (from 2013-2016) and depressive symptoms (from 2010-2016) included 2275 and 1735 respondents, respectively ( Figure 1). The Table exhibits        These findings are in line with our main findings.

Discussion
The mental health of survivors of a traumatic event, such as a natural disaster, has been extensively studied. 22 However, owing to the lack of predisaster information, very few studies have been able to tease out the consequences of disaster exposure from preexisting conditions. In this study, we examined the community-level prevalence of mental health disorders as well as the within-individual trajectories of depressive symptoms and PTSS following the 2011 Great East Japan earthquake and tsunami. We have previously reported 13,14 that personal exposure to trauma (loss of relatives, housing damage) was associated with new onset depressive symptoms and PTSS. In contrast to these individual experiences, from a community-wide perspective, we found little evidence in the present study of an epidemic of depressive symptoms after the disaster. Interestingly, in the absence of predisaster information, researchers might have been tempted to conclude that the high prevalence of depression in the community was attributable to the traumatic experiences associated with the disaster. The second reason why the community prevalence of depressive symptoms appeared to be stable over time is because even though 13.6% of residents experienced new onset of depression after the disaster, this was offset by individuals who were depressed before the disaster but whose symptoms had remitted by follow-up. A third possible explanation for the stability of the prevalence of depression may be due to posttraumatic growth that some individuals experience after a disaster. 25 For example, some reports suggest that people may become more socially active following disaster experiences and that increased social capital might promote improved mental resilience. 26 There are 2 caveats to this interpretation. First, we did not observe the immediate consequences of the disaster in the very short term (eg, within the first year), during which people were more likely to suffer from acute grief and distress given the fact that the first year after a disaster is the peak of psychological symptoms. 11 Second, there may have been some selective attrition of individuals who were depressed before the disaster. In a previous report, we found that individuals who were severely depressed on the day of the earthquake and tsunami were twice as likely to lose their life, possibly as a result of delayed evacuation. 27 The selective mortality of these individuals would tend to underestimate the true counterfactual prevalence of depression in the community after the disaster. On the other hand, the consequence of this bias is likely to be small, given that a total of 34 of the baseline sample of 4857 participants died (yielding a 0.69% mortality rate on the day of the disaster).
Regarding PTSS, we found that approximately 60% of people who reported PTSS in 2013 had recovered by 2016. This is consistent with the findings of the meta-analysis among natural disaster survivors by Morina et al. 3 When the figures for depressive symptoms and PTSS are combined, our data suggest that nearly 40% of survivors were still experiencing significant mental health symptoms (PTSS or depressive symptoms) more than 5 years after the disaster. While this figure presents a scenario of persistent, long-term mental health problems in the wake of the disaster, much of it cannot be attributed to the disaster trauma given that nearly one-third of the sample reported significant depressive symptoms preceding the March 11, 2011, event. In other words, we did not find evidence of a community-wide epidemic of mental illness following a major disaster.

Strengths and Limitations
The major strength of this study was that we collected the baseline mental health data before the disaster. As Galea et al 4 have argued, study designs featuring only postdisaster assessments are unable to distinguish between mental health consequences of disaster vs the persistence of predisaster conditions. Our 3-wave analysis also allowed us to examine the within-individual trajectories of mental illness symptoms during a longer period than in most previous studies.
There are some limitations of our study. First, the response rate at the baseline survey was 59.0%, which is quite similar compared with other community-dwelling surveys. Furthermore, the characteristics of the baseline sample were very similar to 2010 National Census population in the area. 14 The response rates at follow-up surveys were very high (82.1% and 84.6% for the 2 follow-up survey waves, respectively). Second, there might be a chance of selection bias due to loss to follow-up during 3 years. People who were most depressed or had severe PTSS were more likely to have dropped out, which could potentially inflate the recovery rates. However, our sensitivity analysis showed similar results as our main findings.

Conclusions
In this study, approximately 40% of survivors from the Great East Japan earthquake and tsunami still experienced PTSS or depressive symptoms or both 5.5 years after the disaster. However, the mental resilience at the community level was remarkably stable, suggesting that the community itself was resilient.