Data analysis
The associations between the frequency of participants with a
high K10 score and demographic, lifestyle, caregiving status, work
environment, and work-related stress factors were analyzed. The
chi-square test was used for nominal scale data such as sex,
whereas the Cochran-Armitage test was used for ordinal scale data
such as self-rated health.
Furthermore, the simultaneous effect of factors on the frequency
of individuals with a high K10 score was analyzed using a linear
logistic model. The most appropriate model was selected on the
basis of the Akaike Information Criterion (AIC). The AIC is ameasure
of goodness of fit of a statistical model, and provides a means for
model selection. Starting from a model including sex, age, self-rated
health, quality of sleep, satisfaction with daily life, working hours
per day, conversation with supervisor and coworkers, support from
supervisor and coworkers, job overload, intra-/inter-group conflict,
job control, job satisfaction, and caring for elderly relatives as
covariates, the final model with a minimum AIC value was selected
as the most appropriate. The variables of age, self-rated health,
quality of sleep, conversation with supervisor and coworkers,
satisfaction with daily life, job overload, job satisfaction, and caring
for elderly relatives were included in the final model, whereas the
variables of sex, working hours per day, support from supervisor
and coworkers, intra-/inter-group conflict, and job control were
excluded. The maximum likelihood estimation of the final model
parameters was carried out, followed by calculation of the odds
ratio (OR) and its 95% confidence interval (CI) for each covariate in
the model.