Journal of Epidemiology
Online ISSN : 1349-9092
Print ISSN : 0917-5040
ISSN-L : 0917-5040
Current issue
Showing 1-7 articles out of 7 articles from the selected issue
Commentary
Special Article
  • Tomohiro Shinozaki, Etsuji Suzuki
    2020 Volume 30 Issue 9 Pages 377-389
    Published: September 05, 2020
    Released: September 05, 2020
    [Advance publication] Released: July 18, 2020
    JOURNALS OPEN ACCESS
    Supplementary material

    Epidemiologists are increasingly encountering complex longitudinal data, in which exposures and their confounders vary during follow-up. When a prior exposure affects the confounders of the subsequent exposures, estimating the effects of the time-varying exposures requires special statistical techniques, possibly with structural (ie, counterfactual) models for targeted effects, even if all confounders are accurately measured. Among the methods used to estimate such effects, which can be cast as a marginal structural model in a straightforward way, one popular approach is inverse probability weighting. Despite the seemingly intuitive theory and easy-to-implement software, misunderstandings (or “pitfalls”) remain. For example, one may mistakenly equate marginal structural models with inverse probability weighting, failing to distinguish a marginal structural model encoding the causal parameters of interest from a nuisance model for exposure probability, and thereby failing to separate the problems of variable selection and model specification for these distinct models. Assuming the causal parameters of interest are identified given the study design and measurements, we provide a step-by-step illustration of generalized computation of standardization (called the g-formula) and inverse probability weighting, as well as the specification of marginal structural models, particularly for time-varying exposures. We use a novel hypothetical example, which allows us access to typically hidden potential outcomes. This illustration provides steppingstones (or “tips”) to understand more concretely the estimation of the effects of complex time-varying exposures.

    Download PDF (562K)
Original Article
  • Akiomi Inoue, Akizumi Tsutsumi, Yuko Kachi, Hisashi Eguchi, Akihito Sh ...
    2020 Volume 30 Issue 9 Pages 390-395
    Published: September 05, 2020
    Released: September 05, 2020
    [Advance publication] Released: July 13, 2019
    JOURNALS OPEN ACCESS

    Background: Using a 1-year prospective design, we examined the association of job dissatisfaction with long-term sickness absence lasting 1 month or more, before and after adjusting for psychosocial work environment (ie, quantitative job overload, job control, and workplace social support) in Japanese employees.

    Methods: We surveyed 14,687 employees (7,343 men and 7,344 women) aged 20–66 years, who had not taken long-term sickness absence in the past 3 years, from a financial service company in Japan. The Brief Job Stress Questionnaire, including scales on job satisfaction and psychosocial work environment, was administered, and information on demographic and occupational characteristics (ie, age, gender, length of service, job type, and employment position) was obtained from the personnel records of the surveyed company at baseline (July–August 2015). Subsequently, information on the start dates of long-term sickness absences was obtained during the follow-up period (until July 2016) from the personnel records. Cox’s proportional hazard regression analysis was conducted.

    Results: After adjusting for demographic and occupational characteristics, those who perceived job dissatisfaction had a significantly higher hazard ratio of long-term sickness absence than those who perceived job satisfaction (hazard ratio 2.91; 95% confidence interval, 1.74–4.87). After additionally adjusting for psychosocial work environment, this association was weakened and no longer significant (hazard ratio 1.55; 95% confidence interval, 0.86–2.80).

    Conclusions: Our findings suggest that the association of job dissatisfaction with long-term sickness absence is spurious and explained mainly via psychosocial work environment.

    Download PDF (955K)
  • Kaori Honjo, Hiroyasu Iso, Ai Ikeda, Kazumasa Yamagishi, Isao Saito, T ...
    2020 Volume 30 Issue 9 Pages 396-403
    Published: September 05, 2020
    Released: September 05, 2020
    [Advance publication] Released: July 27, 2019
    JOURNALS OPEN ACCESS
    Supplementary material

    Background: Few studies examining the impact for women of employment status on health have considered domestic duties and responsibilities as well as household socioeconomic conditions. Moreover, to our knowledge, no studies have explored the influence of work-family conflict on the association between employment status and health. This research aimed to investigate the cross-sectional associations between employment status (regular employee, non-regular employee, or self-employed) with self-rated health among Japanese middle-aged working women.

    Methods: Self-report data were obtained from 21,450 working women aged 40–59 years enrolled in the Japan Public Health Center-based Prospective Study for the Next Generation (JPHC-NEXT Study) in 2011–2016. Multivariate odds ratios (ORs) and 95% confidence intervals (CIs) were calculated for poor self-rated health (‘poor’ or ‘not very good’) by employment status. Sub-group analyses by household income and marital status, as well as mediation analysis for work-family conflict, were also conducted.

    Results: Adjusted ORs for the poor self-rated health of non-regular employees and self-employed workers were 0.90 (95% CI, 0.83–0.98) and 0.84 (95% CI, 0.75–0.94), respectively, compared with regular employees. The identified association of non-regular employment was explained by work-family conflict. Subgroup analysis indicated no statistically significant modifying effects by household income and marital status.

    Conclusion: Among middle-aged working Japanese women, employment status was associated with self-rated health; non-regular employees and self-employed workers were less likely to report poor self-rated health, compared with regular employees. Lowered OR of poor self-rated health among non-regular employees may be explained by their reduced work-family conflict.

    Download PDF (256K)
  • Takafumi Abe, Jun Kitayuguchi, Shinpei Okada, Kenta Okuyama, Tatsunosu ...
    2020 Volume 30 Issue 9 Pages 404-411
    Published: September 05, 2020
    Released: September 05, 2020
    [Advance publication] Released: August 10, 2019
    JOURNALS OPEN ACCESS
    Supplementary material

    Background: Although moderate-to-vigorous physical activity (MVPA) has multiple health benefits, current participation in recommended MVPA level and its determinants among Japanese children and adolescents remain unclear. Therefore, this cross-sectional study investigated the prevalence of meeting recommended MVPA level and its correlates among Japanese children and adolescents.

    Methods: Using the Japanese version of the World Health Organization (WHO) Health Behaviour in School-aged Children survey questionnaire, we confirmed the prevalence of meeting recommended MVPA level in all primary schools (PS) and junior high schools (JHS) in Unnan City, Japan. We evaluated its association with school grade, gender, body weight status, screen time, consumption of breakfast, physical activity (PA) preference, and population density using Poisson regression.

    Results: We found that 20.1% of the 1,794 students (9–15 years old) met the WHO recommendation. Meeting recommended MVPA level was significantly associated with being in the sixth grade of PS (prevalence ratio [PR] 0.57; 95% confidence interval [CI], 0.39–0.84) and first (PR 1.52; 95% CI, 1.16–1.99), second (PR 1.45; 95% CI, 1.10–1.90), and third grade of JHS (PR 0.40; 95% CI, 0.26–0.62) (vs fourth grade of PS); being a boy (PR 1.33; 95% CI, 1.12–1.59) (vs girl); liking PA (PR 3.72; 95% CI, 2.22–6.22) (vs dislike); and belonging to a medium-population-density (PR 0.73; 95% CI, 0.61–0.88) or low-population-density area (PR 0.67; 95% CI, 0.48–0.94) (vs high-population-density area).

    Conclusions: About 20% of Japanese children and adolescents engaged in the recommended MVPA level. MVPA was associated with grade, gender, preference for PA, and population density.

    Download PDF (516K)
  • Takashi Oshio
    2020 Volume 30 Issue 9 Pages 412-419
    Published: September 05, 2020
    Released: September 05, 2020
    [Advance publication] Released: August 10, 2019
    JOURNALS OPEN ACCESS

    Background: A growing amount of evidence demonstrates the adverse impacts of economic downturns on population health. However, the extent to which the macroeconomic conditions at labor market entry affect health outcomes in later life remains relatively understudied. This study focused on the health outcomes of the cohort who entered the labor market during the “employment ice age” (EIA; 1993–2004) in Japan, when young people had difficulty finding jobs after graduating from college or high school.

    Methods: We used repeated cross-sectional data (N = 3,054,782; 1,500,618 men and 1,554,164 women) obtained from an 11-wave population-based nationwide survey conducted every 3 years from 1986 through 2016. We considered three health outcomes: being in hospital, subjective symptoms, and self-rated health (SRH). We employed two types of statistical analyses: an age-period-cohort (APC) analysis, which controlled for age and period (wave) effects, and a difference-in-differences (DiD) analysis, in which the EIA experience was regarded as a treatment.

    Results: The APC analysis confirmed the relative disadvantage of the EIA cohort for all three outcomes; for instance, the odds ratio of poor SRH for the EIA cohort was 1.29 (95% confidence interval [CI], 1.21–1.38) for men and 1.25 (95% CI, 1.17–1.34) for women. The DiD analysis confirmed the robustness of these results, especially for men.

    Conclusions: The results underscored the lingering impact of the macroeconomic conditions at labor market entry on health outcomes in later life in Japan.

    Download PDF (1423K)
Statistical Data
feedback
Top