Relationships Between Childhood Bullying/Domestic Violence Experience and Insomnia among Employees in Japan

Design and participants

We conducted a cross-sectional study using data obtained from a survey of the Tsukuba Science City Network in Tsukuba, a city situated about 70 km northeast of Tokyo. The Tsukuba Science City Network is an organization of the city’s research and academic institutes that aims to promote cooperation among its member institutes (MEXT, n.d.). The Tsukuba Science City Network conducts a mental health survey every 5 years of the researchers and engineers from the member institutes in Tsukuba city. From February to March 2017, we conducted the survey via an anonymous web questionnaire entitled The 7th Life Environment and Workplace Stress Survey. The total number of subjects in this study was 19,481 workers at 53 research institutes in Tsukuba City, all of whom were affiliated with the Tsukuba Science City Network. These workers were contacted via e-mail through the general affairs department of each institution and directed to a self-administered questionnaire form via a URL described in the e-mail. Participants were given the option to choose either the Japanese or English version of the questionnaire. All participants remained anonymous, and no data regarding their institute affiliation was obtained. The results of this survey have been published on the Tsukuba Science City Network website, and several studies using this survey have been reported (Hori et al., 2019; Takahashi et al., 2019; Ikeda et al., 2020). The total number of respondents was 7255, yielding a response rate of 37.2%. Sixty participants were excluded due to missing data or no response. The final sample consisted of 7175 participants (average age: 44 years; 4509 men and 2666 women) aged 20 to 65 years who answered the main response items. The detailed protocols are available at protocols.io (DOI: dx.doi.org/10.17504/protocols.io.14egn24nmg5d/v1).

Measures

Experiences of bullying/DV

We collected data about experience of bullying/DV by means of an original questionnaire. The experience of bullying victimization was defined by answering “yes” to the question “Have you ever been bullied by others?” Similarly, the experience of DV victimization was defined by answering “yes” to the question “Have you ever been a victim of violence by a family member?” We observed when the experienced occurred (elementary school, junior high school, high school, or university and after). The respondents were able to select overlapping periods of time when they experienced victimization. They were classified into five groups according to the number of experiences; never experienced at all, experienced one period, experienced two periods, experienced three periods, and experienced four periods. We defined the experience from early childhood through high school years as a childhood experience and the experience from university years and after as a current experience.

Therefore, childhood bullying or DV victimization experiences were divided into four groups; never experienced at all, experienced one period, experienced two periods, and experienced three periods.

Athens Insomnia Scale (AIS)

The outcome variable analyzed in this study was insomnia, which was measured using the Japanese version of the AIS (Okajima et al., 2013). This self-administered questionnaire assesses insomnia according to the 10th revision of the International Classification of Diseases criteria (Soldatos et al., 2000; Okajima et al., 2013). The Japanese version of the AIS has been validated (Okajima et al., 2013) and is commonly used in epidemiological studies. It consists of 8 items: the first 5 assess difficulties with sleep induction, waking during the night, early-morning waking, total sleep time, and overall sleep quality, while the remaining 3 measure the consequences of insomnia during the day, including problems with sense of well-being, overall functioning, and sleepiness. Respondents were asked to rate their experiences with these symptoms over the past month on a 4-point scale ranging from 0 (“not problematic at all”) to 3 (“extremely problematic”). The total score ranges from 0 to 24. A score of less than 4 indicates no problems, while a score of 4 to 5 suggests that consultation with a physician may be necessary (some suspicion of insomnia). A score greater than 6 indicates that consultation with a physician is necessary (suspected insomnia) (Soldatos et al., 2003). Respondents were then dichotomized into two groups: the insomnia group (AIS total score ≥6) and the no-insomnia group (AIS total score ≤5).

Occupational stress

Chronic occupational stress as perceived by the workers was assessed using the Brief Scales for Job Stress (BSJS) (Nishikido N, Kageyama T, Kobayasi T, 2000). The BSJS is a 20-item questionnaire developed by Nishikido and colleagues, based on the job demand-control-support model similar to the Job Content Questionnaire (Karasek et al., 1998). All participants were asked to “select the response that most closely matches your feelings about the descriptions of your current working circumstances”. Responses were rated on a 4-point scale (from 1=“disagree” to 4=“agree”), and the mean scores (range: 1.00–4.00) were calculated for six subscales: “workload,” “mental workload,” “interpersonal relationships,” “job control,” “reward from work,” and “support from colleagues and superiors.” These subscales have sufficient internal consistencies (Nishikido N, Kageyama T, and Kobayasi T, 2000). Workload, mental workload, and interpersonal relationships were defined as job-related stress, while job control, reward from work, and support from colleagues and superiors were defined as buffers against it.

Other covariates

The covariates were age; sex (man, woman); smoking history (current, former, non-smoker); education (high school, university, graduate university, other school); annual house income (less than 4 million yen, more than 4 million yen and less than 8 million yen, more than 8 million yen and less than 12 million yen, more than 12 million yen); exercise habits (a few times a month, once a week, twice a week, 3 or more times a week); marriage status (unmarried, married, divorced, bereaved); presence of children (yes, no); working years (less than 1 year, more than 1 year and less than 3 years, more than 3 years and less than 5 years, more than 5 years and less than 10 years, more than 10 years); regularly go to a hospital (yes, no); residence (Tsukuba science city, Ibaraki prefecture, Tokyo-metropolitan, other); occupation (research or education, office work, technical job, other); job type (full-time, fixed-term, full-time, part-time, permanent); and type of organization (national, agency, private, corporation).

Statistical analysis

The AIS rates 5 points or less as no insomnia symptoms, and 6 points or more, as insomnia symptoms. A crosstabulation table was prepared for those with and without insomnia symptoms by AIS, and chi-square and unpaired t-tests were performed.

We conducted a binomial logistic regression analysis with two groups as the objective variables. We created models with childhood bullying experience and domestic violence experience as explanatory variables and additional inputs of age, gender, smoking habits, family income, education, exercise habits, marital status, presence of children, hospital visits, years of work, place of residence, and BSJS as covariates. The BSJS was entered into a logistic regression model with the mean scores of each of the subscales as quantitative variables. Multicollinearity was checked in each model. Hosmer-Lemeshow tests were conducted to examine the quality of each model. Also, Nagelkerke pseudo R-squared measures was calculated.

All statistical tests were 2-sided, and probability values below 0.05 were considered to indicate significance. We also calculated the 95% confidence intervals.

EZR version 1.40 (Saitama Medical Center, Jichi Medical University, Saitama, Japan) (Kanda, 2013) were used for the statistical analysis. EZR is a graphical user interface for R (The R Foundation for Statistical Computing, Vienna, Austria). More precisely, it is a modified version of R commander designed to add statistical functions frequently used in biostatistics.

Ethical considerations

The web survey contained clear statements that participation was entirely voluntary, that it was an anonymous survey, that the privacy of the respondent would be respected, and that the data would be strictly controlled. In The 7th Life Environment and Workplace Stress Survey, the respondents were not told in advance that the data would be used for research purposes.

The consent of the participants was taken when The Tsukuba Science City Network conducted the study. We received permission from The Tsukuba Science City Network to use the data for our research. We published an opt-out notice on our laboratory’s homepage, stating that the data would be used for research purposes without any name on the participants’ names (http://occup-aerospace-psy.org/content/files/7tsukukyov3R1.pdf).

From the 8th survey, we will inform respondents in advance that the collected data will be used for research purposes. This research proposal was reviewed and approved by the ethics committee of the University of Tsukuba (approval #1374). All procedures were conducted in accordance with the ethical standards of the national research committee and the Helsinki Declaration.

Descriptive characteristics

Table 1 and Table 2 show the participants’ characteristics. The insomnia group consisted of 2997 patients (41.8%) and the non-insomnia group consisted of 4178 patients (58.2%). A significant difference was found in the number of individuals in the insomnia and non-insomnia groups between the bullying victim group and the DV victim group (chi-squared test, p<0.001). The insomnia group was also higher on all means of the BSJS subscales (unpaired t-test, all p<0.001).

Binomial logistic regression analysis

Table 3 and Table 4 show the results of a binomial logistic regression analysis. A Hosmer-Lemeshow test was performed on the model and the probability value was less than 0.05, indicating a poor fit (p<0.001). Nagelkerke pseudo R-squared measures was 0.151. Job type and occupation were excluded from the explanatory variables because multicollinearity was obtained (variance inflation factor>10).

Regarding childhood bullying experiences, there was a trend toward insomnia in the group that experienced one period of bullying and in the group that experienced three periods of bullying compared to the group that never experienced (OR 1.38, 95%CI 1.19-1.60, OR 1.83, 95%CI 1.19-2.84, respectively). On the other hand, regarding childhood experience of domestic violence, there was a trend toward insomnia in the groups that experienced two periods and three periods compared to the group that never experienced (OR 1.82, 95%CI 1.24-2.68, OR 2.06, 95%CI 1.24-3.43, respectively). In BSJS, workload (OR 1.15, 95%CI 1.06-1.24), mental workload (OR 1.42, 95%CI 1.31-1.55), and interpersonal relationship (OR 1.38, 95%CI 1.28-1.48) were positively associated with insomnia as stress load factors, respectively. Regarding mitigating factors, Support from colleagues and superiors (OR 0.86, 95%CI 0.78-0.94) and Reward from work (OR 0.82, 95%CI 0.76-0.88) were negatively associated with insomnia, but not significantly associated with job control (OR 0.92, 95%CI 0.85-1.00).