Descriptive characteristics
A total of 1,089 long-term residents of the 13 shelters in the disaster area (the shelter group) were interviewed, including 242 (55.4%) out of 437 in Iwate, 329 (55.6%) out of 592 in Miyagi, and 518 (44.0%) out of 1178 in Fukushima. The overall response rate for shelter respondents was 49.4%. The participation rate for the general population was 46.1% (852 out of 1,850). Respondents from the shelters were more likely to be female, older, widowed, have low levels of education, and report disaster-related experiences compared to the general population (Table 1). Shelter respondents had significantly lower rates of most pre-disaster disorders compared to the general population.
Table 1. Characteristics of the shelter group and the general population sample three year after the Great East Japan Earthquake in 2011
|
Shelter group (n=1,089)
|
General population in east Japan (n=852)
|
p for difference†
|
|
n
|
%
|
n
|
%
|
|
Sex (females)
|
667
|
61.2
|
438
|
51.4
|
<0.010
|
Age, years
|
|
|
|
|
|
20-39
|
100
|
9.2
|
233
|
27.3
|
<0.001
|
40-64
|
357
|
32.8
|
427
|
50.1
|
|
65+
|
632
|
58.0
|
192
|
22.5
|
|
Marital status
|
|
|
|
|
Married
|
628
|
57.7
|
615
|
72.2
|
<0.001
|
Divorced/separated
|
75
|
6.9
|
46
|
5.4
|
|
Widowed
|
266
|
24.4
|
39
|
4.6
|
|
Never married
|
120
|
11.0
|
152
|
17.8
|
|
Education (high school or higher)
|
653
|
60.0
|
788
|
92.5
|
<0.001
|
ADL (limited)
|
57
|
5.2
|
1
|
0.1
|
<0.001
|
Disaster experiences
|
|
|
|
Own injury (any)
|
53
|
4.9
|
1
|
0.1
|
<0.001
|
Bereavement during the disaster
|
|
None
|
450
|
41.3
|
835
|
98.0
|
<0.001
|
Family/relative
|
313
|
28.7
|
5
|
0.6
|
<0.001
|
Friend/acquaintance
|
128
|
11.8
|
12
|
1.4
|
<0.001
|
Both
|
198
|
18.2
|
-
|
-
|
<0.001
|
House damage (any)
|
877
|
80.5
|
23
|
2.7
|
<0.001
|
Radiation anxiety (high)
|
326
|
29.9
|
118
|
13.8
|
<0.001
|
Lifetime prevalence of mental disorders before the disaster
|
|
Any‡
|
79
|
7.3
|
157
|
18.4
|
<0.001
|
Major depressive episode
|
13
|
1.2
|
32
|
3.8
|
<0.001
|
Manic episode
|
1
|
0.1
|
2
|
0.2
|
0.587
|
GAD
|
6
|
0.6
|
16
|
1.9
|
0.008
|
Panic disorder
|
4
|
0.4
|
4
|
0.5
|
0.736
|
PTSD
|
4
|
0.4
|
3
|
0.4
|
1.000
|
Alcohol use disorder
|
64
|
5.9
|
123
|
14.4
|
<0.001
|
† Chi-square test for age and marital status; otherwise, Fisher exact test.
‡ Any of 6 mental disorders including major depressive episode, manic episode, generalized anxiety disorder (GAD), panic disorder, post-traumatic stress disorder (PTSD), and alcohol use disorder.
Incidence and remission of post-disaster mental disorders in the shelter group and the general population
In the shelter group and the general population, 3.3% and 0.7% of respondents, respectively, met criteria for ANY mental disorder in 2011, the year the disaster happened (Table 2). The three-year cumulative incidence until 2013 was 5.6% and 2.7%, respectively. MDE was the most common disorder in the both samples. The three-year cumulative incidence of ANY mental disorder was significantly higher in the shelter group than in the general population (p = 0.010 for crude analysis and p = 0.002 after adjusting for sex, age, and education). In terms of specific disorders, compared to the general population, the shelter group had significantly higher cumulative incidences of MDE, GAD, and PTSD (p < 0.01), but not MAN, PD, or AUD (p > 0.05).
Table 2. Cumulative incidence of mental disorders among respondents without pre-disaster mental disorders following the Great East Japan Earthquake 2011*
|
ANY
% (SE)
|
MDE
% (SE)
|
MAN
% (SE)
|
GAD
% (SE)
|
PD
% (SE)
|
PTSD
% (SE)
|
AUD
% (SE)
|
Shelter group (n=1010) †
|
|
|
|
|
|
|
|
Post disaster year 1
|
3.3(0.6)
|
2.9(0.5)
|
0.2(0.1)
|
1.4(0.4)
|
0.1(0.1)
|
0.6(0.2)
|
0.1(0.1)
|
Post disaster year 2
|
4.8(0.7)
|
4.1(0.6)
|
0.2(0.1)
|
1.9(0.4)
|
0.1(0.1)
|
1.0(0.3)
|
0.2(0.1)
|
Post disaster year 3
|
5.6(0.7)
|
4.3(0.6)
|
0.2(0.1)
|
2.3(0.5)
|
0.2(0.1)
|
1.0(0.3)
|
0.4(0.2)
|
General population (n=695) †
|
|
|
|
|
|
|
|
Post disaster year 1
|
0.7(0.3)
|
0.3(0.2)
|
0(-)
|
0.1(0.1)
|
0(-)
|
0(-)
|
0.4(0.2)
|
Post disaster year 2
|
1.3(0.4)
|
0.6(0.3)
|
0(-)
|
0.1(0.1)
|
0(-)
|
0(-)
|
0.7(0.3)
|
Post disaster year 3
|
2.7(0.6)
|
1.4(0.5)
|
0.1(0.1)
|
0.3(0.2)
|
0(-)
|
0(-)
|
1.2(0.4)
|
Log-rank p for group difference
|
0.010
|
0.006
|
0.521
|
0.001
|
0.240
|
0.006
|
0.072
|
Cox hazard model adjusted for sex, age, and educational attainment:
|
Adjusted hazard ratio
|
2.23
|
2.46
|
4.65
|
11.13
|
NC
|
NC
|
0.60
|
95%CI
|
(1.33- 3.75)
|
(1.27- 4.76)
|
(0.45- 48.09)
|
(2.55- 48.65)
|
|
|
(0.17- 2.08)
|
p
|
0.002
|
0.007
|
0.299
|
0.001
|
|
|
0.421
|
* Percentage (%) and the standard error (SE) were presented.
† Respondents with any disorder before the disaster were excluded. Any disorder includes any of 6 mental disorders including major depressive episode (MDE), manic episode (MAN), generalized anxiety disorder (GAD), panic disorder (PD), post-traumatic stress disorder (PTSD), and alcohol use disorder (AUD)
NC: not calculated.
The proportion of individuals who had not remitted from ANY post-disaster mood and anxiety disorders (excluding alcohol use disorder) was higher in the shelter group (n = 57) than in the general population (n = 14) (log-rank test, chi-square = 17.9, df = 1, p < 0.001) (Figure). The median durations of post-disaster ANY were 2.00 years (standard error [SE], 0.36) and 0.25 years (SE, 0.10) in the shelter group and the general population, respectively.
Seeking treatment for post-disaster mental disorders
The proportions of individuals who visited psychiatrists or other physicians among people with post-disaster ANY were greater in the shelter group (42.6%, 26 of 61 cases) than in the general population (17.4%, 4 of 23 cases) (Fisher exact test, p = 0.011) (see Appendix table 2). The proportions of those seeking treatment for MDE were 39.5% (17 of 43 cases) for the shelter group and 30.8% (4 of 13 cases) for the general population (Fisher exact test, p = 0.339). The proportion of treatment for PTSD (9.1%) was lower than for the other disorders in the shelter group, while no comparison was possible between the two groups because of no post-disaster cases with PTSD in the general population. None of respondent with post-disaster AUD (n = 5 for the disaster areas; n = 9 for the general population) had visited psychiatrists or other physicians.
Disaster risk factors associated with onset and remission of post-disaster mental disorders in the shelter group
In the shelter group, being divorced/separated (OR=3.00, 95%CI = 1.35 – 6.68, p=0.007), personal injury (OR = 6.10, 95% CI = 2.97 - 12.55, p<0.001), and bereavement of both family/relative and friend (OR = 2.73, 95% CI = 1.13, 6.60, p=0.026) were significantly associated with post-disaster onset of ANY disorder (Table 3). In the disorder-specific analyses (data available upon request), it was found that female sex (OR=2.41, 95%CI = 1.12 - 5.17, p=0.024), being never married (OR=4.05, 95%CI = 1.42 – 11.53, p=0.009), personal injury (OR= 8.84, 95%CI = 3.91 – 20.00, p<0.001), and bereavement of both family/relative and friends (OR = 3.41, 95% CI = 1.23 – 9.42, p=0.018) were significantly associated with the onset of MDE. Personal injury (OR = 8.65, 95% CI = 1.95 – 38.42, p=0.005) and bereavement of both family/relative and friend (OR = 22.53, 95% CI = 1.67 – 303.83, p=0.019) were significantly associated with PTSD. Personal injury was the only disaster risk factor significantly associated with the onset of GAD (OR = 5.98, 95% CI = 2.08 – 17.21, p=0.001).
Table 3. Factors associated with post-disaster mental disorders in the shelter group during three years after the Great East Japan Earthquake 2011†
Variables
|
OR‡
|
95%CI
(low to high)
|
p
|
Sex (female)
|
0.69
|
0.38
|
1.27
|
0.235
|
Age, years
|
|
|
|
|
20-39
|
1.00
|
|
|
|
40-64
|
1.56
|
0.57
|
4.31
|
0.388
|
65+
|
1.38
|
0.46
|
4.17
|
0.570
|
Education (high school or higher)
|
1.12
|
0.61
|
2.04
|
0.721
|
Marital status
|
|
|
|
|
Married
|
1.00
|
|
|
|
Divorced/separated
|
3.00
|
1.35
|
6.68
|
0.007
|
Widowed
|
0.90
|
0.45
|
1.82
|
0.778
|
Never married
|
1.90
|
0.75
|
4.82
|
0.174
|
Current ADL (limited)
|
0.55
|
0.12
|
2.39
|
0.423
|
Personal injury (yes)
|
6.10
|
2.97
|
12.55
|
<0.001
|
Bereavement during the disaster
|
|
|
|
|
None
|
1.00
|
|
|
|
Family/relative
|
0.94
|
0.45
|
1.95
|
0.867
|
Friend/acquaintance
|
1.87
|
0.80
|
4.37
|
0.150
|
Both
|
2.73
|
1.13
|
6.60
|
0.026
|
Home damage (yes)
|
2.12
|
0.90
|
5.00
|
0.085
|
Radiation risk perception (high)
|
1.46
|
0.80
|
2.68
|
0.217
|
Study area (prefecture)
|
|
|
|
|
Iwate
|
1.00
|
|
|
|
Miyagi
|
0.80
|
0.31
|
2.04
|
0.638
|
Fukushima
|
1.75
|
0.72
|
4.26
|
0.216
|
† Discrete proportional hazard model was used. Among 1,010 respondents, 61 developed any of 6 post disaster mental disorders (major depressive episode, manic episode, generalized anxiety disorder, panic disorder, PTSD, and alcohol use disorder).
‡ OR: odds ratio.
None of the demographic and disaster-related factors were significantly associated with the remission from post-disaster mood and anxiety disorders among 57 respondents who had any post-disaster mood and anxiety disorder (Kaplan-Meier method with the log-rank test, p>0.05). (data available upon request).
Post-hoc power analysis
For the comparison of the three-year cumulative incidences of post-disaster mental disorders (z-test), we observed an about two-fold difference between the shelter sample (5.6% among 1,010 respondents) and the general population (2.7% among 695 respondents).The statistical power to detect the difference with two-tailed p=0.05 was estimated as 0.85.For the comparison of the remission of post-disaster mental disorders (log-rank test), for instance, at one-year, we observed median survival durations in the shelter sample (2.00 year among 57 respondents) and the general population (0.25 years among 14 respondents). An estimation of the statistical power to detect the difference with two-tailed p=0.05 was 0.91.