Introduction Traumatic experiences and disordered sleep are strongly associated with drinking problems. We examined the effects of experiencing the Great East Japan Earthquake and subsequent nuclear power plant accident, and of sleep problems, on behavioral changes observed in nondrinkers. Methods This study examined cross-sectional data from the Mental Health and Lifestyle Survey conducted among residents in restricted areas of Fukushima in 2012. Participants were 21,454 evacuees aged 20 years or older at the time of disaster. People who did not drink before the disaster but became drinkers afterwards were compared with the rest of the cohort. We analyzed the association between behavioral changes in non-drinkers and potentially predictive variables, using logistic regression. Results The behavioral change of non-drinkers becoming drinkers (n=2,148) was significantly related to being male (OR=1.93, 95% CI : 1.74-2.15), being younger (21-49 yrs, OR=1.85, 95% CI : 1.60-2.13), having less educational attainment (up to high school graduate, OR=1.21, 95% CI : 1.091.35), smoking (OR=1.22, 95% CI : 1.08-1.38), losing family or relatives (OR=1.21, 95% CI : 1.071.37), change in employment (OR=1.19, 95% CI : 1.07-1.32), having severe sleep problems as measured by a Japanese version of the Athens Insomnia Scale (3-8, OR=1.45, 95% CI : 1.30-1.62), and severity of traumatic symptoms as measured by the PTSD Checklist Stressor-Specific (PCL-S) score (<44, OR=1.33, 95% CI : 1.17-1.51). Conclusion Having sleep problems and having more severe traumatic symptoms are significantly related to non-drinkers becoming drinkers.

We investigated the psychometric properties of the simplified Japanese version of the Athens Insomnia Scale (AIS‐SJ) using baseline data from the Fukushima Health Management Survey. Data from 22 878 men and 27 669 women aged 16 years and older were analysed (Mage = 52.9 ± 18.6). Participants lived in the Fukushima evacuation zone and experienced the Great East Japan Earthquake. The AIS‐SJ was used to assess participants’ insomnia symptoms, and its validity was examined by administering the Kessler 6‐item Psychological Distress Scale (K6) and assessing education, self‐rated health and disaster‐related experiences. A confirmatory factor analysis revealed that the two‐factor model was a better fit than the one‐factor model. The AIS‐SJ and its subscales had acceptable reliability (Cronbach's alpha, 0.81). Test of measurement invariance confirmed strict invariance across groups for the participants’ characteristics of gender and mental illness history, but not for participants’ age. AIS‐SJ scores exhibited a near‐normal distribution (skewness, 0.45; kurtosis, −0.89). There were significant age differences only among women, and gender differences in AIS‐SJ scores with small effect sizes. The AIS‐SJ scores had weak‐to‐moderate correlations with mental illness history, bereavement, experiencing the tsunami, experiencing the nuclear power plant incident, housing damage and losing one's job (polyserial correlations, 0.36, 0.17, 0.13, 0.18, 0.13, and 0.15, respectively), and strong correlations with self‐rated health (polyserial correlation, 0.51), psychological distress (rs, 0.60) and post‐traumatic stress disorder (rs, 0.60). The AIS‐SJ is a useful instrument for assessing community dwellers’ insomnia symptoms.