Journal of Epidemiology最新文献

Background: Despite Japan's universal health insurance system, health disparities have increased since the 1990s. However, the impact of area deprivation on various aspects of child health remains understudied.

Methods: This population-based cohort study followed 38,554 children born in Japan (May 10-24, 2010) from birth to age 5.5 years. Using an outcome-wide approach, Bayesian three-level logistic regression models (individuals in municipalities within eight major regions) assessed associations between municipality-level Area Deprivation Index (ADI) at birth and multiple preschool health outcomes (hospitalizations for all causes; respiratory infections; gastrointestinal diseases; Kawasaki disease; medical visits for asthma, allergic rhinitis, atopic dermatitis, food allergy, injury, intussusception; prevalence of overweight/obesity), adjusting for individual-level factors.

Results: Higher ADI was associated with increased risk of all-cause hospitalization (adjusted odds ratio [aOR] per 1-standard-deviation increase in ADI, 1.04; 95% credible interval [CrI], 1.01-1.07), respiratory infections (aOR, 1.08; 95% CrI, 1.04-1.13), gastrointestinal diseases (aOR, 1.11; 95% CrI, 1.03-1.20), asthma (aOR, 1.10; 95% CrI, 1.01-1.19). Overweight/obesity at age 5.5 years also increased with higher ADI (aOR, 1.11; 95% CrI, 1.06-1.16). Higher ADI was inversely associated with Kawasaki disease (aOR, 0.86; 95% CrI, 0.77-0.96), though not robust in sensitivity analysis. Geographic clustering was observed for all outcomes, particularly at municipality level.

Conclusion: We found persistent municipal-level health inequalities across various childhood health outcomes in Japan, despite its universal health insurance system. These findings suggest that policymakers should address health inequalities through comprehensive strategies targeting broader social determinants beyond health care access.

Background: Older Japanese adults have the highest drowning mortality rates globally, likely due to in-home bathing customs. However, epidemiological evidence of preventive strategies based on national data is lacking. We aimed to describe the trends in bathtub drowning deaths (International Classification of Diseases, Tenth Revision code: W65) across Japan and explore factors that may reduce W65-coded deaths.

Methods: We collected the data of all W65-coded deaths that occurred at home from 1995 to 2020 using death certificates from the Ministry of Health, Labour and Welfare. The national age-adjusted mortality rates (AMRs) and prefecture-specific age-standardized mortality ratios (SMRs) were calculated. Data on demographic, socioeconomic, environmental factors, and nursing care services were obtained from the Japan Portal Site of Official Statistics. Mixed-effects analysis was used to examine the association between SMR and potential contributing factors at the prefecture level.

Results: We identified 99,930 W65-coded deaths at home, with the highest incidence among individuals aged 80-84 years, peaking in January. Since 2010, AMRs have consistently exceeded 3.0 per 100,000. An inverse association was found between SMR and the number of geriatric health service facilities and senior welfare centers per capita (coefficients per 1 standard deviation increase, -0.09; 95% confidence interval (CI), -0.13 to -0.05, P < 0.001 and -0.07; 95% CI, -0.11 to -0.02, P = 0.004, respectively), after adjusting for demographic, socioeconomic, and environmental factors.

Conclusion: Sustained high AMRs suggest that the rising death toll was not solely due to aging. Increased access to nursing care facilities may help prevent W65-coded deaths.

Background: Preventable behavioral risk factors account for approximately one third of mortality, morbidity, and disability worldwide. This study aims to quantify the interest in behavioral risk factors within major medical journals in 2022 and to derive trends over the past 30 years in the entire medical literature.

Methods: We analyzed the proportion of publications dealing with tobacco smoking, alcohol drinking, use of illicit drugs, excess body weight and physical activity among all the 1,128 articles published in the Journal of the American Medical Association, the British Medical Journal, the Lancet, and New England Journal of Medicine in 2022. A joinpoint analysis was conducted running in PubMed/MEDLINE specific search strings to evaluate trends over the last 30 years in the four journals and in the whole medical literature.

Results: In 2022, of all publications from the four considered medical journals, 2.8% dealt with tobacco smoking, 1.6% alcohol drinking, 1.1% use of illicit drugs, 3.8% excess body weight, 2.7% physical activity and 8.0% dealt with any behaviors. The joinpoint analysis on the whole medical literature showed that papers on modifiable risk factors significantly increased from 3.9% in 1993 to 6.2% in 2014 (annual percent change [APC]: between +1.83% and +4.09%), and subsequently decreased between 2014 and 2019 (APC -0.31%), with an acceleration thereafter (APC -2.41% in 2019-2022).

Conclusion: For the first time we quantified the volume of medical research focused on preventable behavioral risk factors. This appears to be limited and declining over the last decade. Research on primary prevention should be a priority to face the emergence of associated non-communicable diseases globally.

Background: In observational studies estimating the association between treatment and time-to-event outcomes, time-related biases can substantially impact results. Immortal time bias is one of such biases, and two types are known: misclassified immortal time bias and excluded immortal time bias. These biases often arise from incorrect time-zero definition, especially with non-user controls. This study aims to illustrate immortal time bias in non-user controls using formulas, simulations, and real-world data.

Methods: For our simulations, we considered two scenarios: one with no confounding and no treatment effect, and the other with time-dependent confounding. We compare three different settings of time-zero for treatment and control groups. Method 1: Both groups were followed from cohort entry date (CED). Method 2: The treatment group was followed from treatment initiation date (TID), while the non-user group was followed from CED. Method 3: The treatment group was followed from TID, and non-users were matched to treatment patients, followed from the corresponding TID of their matched patient.

Results: Our simulation showed that both Method 1 and Method 2 can exhibit large biases in the estimated treatment effect due to immortal time bias. The magnitude of the bias is greater for Method 1 than for Method 2. On the other hand, Method 3 showed almost no bias. Even in the presence of time-dependent confounding, Method 3 did not introduce bias.

Conclusions: To reduce time-related biases, it is crucial for researchers to carefully define an appropriate time-zero, especially when using a non-user control group.

Background: New long-term care insurance (LTCI) certifications and mortality are key outcomes in cohort studies involving older adults; however, the coronavirus disease 2019 (COVID-19)'s comprehensive impacts on these outcomes remain underexplored. We examined the pandemic's impact on new LTCI applications and all-cause mortality in a metropolitan cohort.

Methods: In 2016, 15,500 individuals aged 65-84 years were randomly selected through stratified sampling from Ota City, Tokyo. LTCI and death records were tracked through December 2023; the monthly LTCI applications and all-cause deaths per 10,000 people were calculated. The COVID-19 pandemic period was defined as beginning in March 2020, after the World Health Organization Director-General characterized the situation as a pandemic on March 11, 2020. Interrupted time-series segmented regression analysis was used to compare trends pre- (January 2018-February 2020) and post-pandemic onset (March 2020-December 2023).

Results: From January 2018 to December 2023, 4,083 new LTCI applications and 2,457 deaths were recorded. New monthly LTCI applications showed a modest upward trend pre-pandemic (0.4 per 10,000 people; 95% confidence interval [CI], 0.1-0.8), declined sharply at the pandemic's onset (-9.6 per 10,000 people; 95% CI, -16.0 to -3.2), and subsequently increased at a higher rate than pre-pandemic levels (0.8 per 10,000 people; 95% CI, 0.6-1.0). Monthly all-cause deaths remained stable before and immediately after the pandemic's onset but rose slightly in the post-pandemic period (0.3 per 10,000 people per month; 95% CI, 0.2-0.5).

Conclusion: The COVID-19 pandemic influenced both new LTCI applications and all-cause mortality in this study. These impacts should be carefully considered in cohort studies examining these outcomes.

Background: This study aimed to assess the clinical relevance of three-dimensional occupational stress (job stressor score [A score], psychological and physical stress response score [B score], and social support for workers score [C score]) of the Brief Job Stress Questionnaire (BJSQ) in the national stress check program in Japan to irregular menstruation.

Methods: The present retrospective cohort study included 2,078 female employees aged 19-45 years who had both annual health checkups and the BJSQ between April 2019 and March 2022 in a national university in Japan. The outcome was self-reported irregular menstruation measured at annual health checkups until March 2023. A dose-dependent association between BJSQ scores and incidence of irregular menstruation was examined using Cox proportional hazards models to calculate multivariable-adjusted hazard ratios (HRs) of four quantile (0-49% [Q_0-49], 50-74% [Q_50-74], 75-89% [Q_75-89], and 90-100% [Q_90-100]) of the BJSQ scores.

Results: During 2.0 years of the median observational period, 257 (12.4%) women reported irregular menstruation. B score, not A or C scores, was identified as a significant predictor of irregular menstruation (adjusted HR of A, B, and C scores per 1 standard deviation: 1.06 [95% confidence interval {CI}, 0.89-1.27], 1.35 [95% CI, 1.15-1.57], and 0.93 [95% CI, 0.80-1.08], respectively). Women with higher B score had a significantly higher risk of irregular menstruation in a dose-dependent manner (adjusted HR of Q_0-49, Q_50-74, Q_75-89, and Q_90-100: 1.00 [reference], 1.38 [95% CI, 1.00-1.90], 1.48 [95% CI, 1.00-2.18], and 2.18 [95% CI, 1.38-3.43], respectively).

Conclusion: Psychological and physical stress response predicted irregular menstruation.

Background: Long-term sickness absence (LTSA) is an important public health challenge, yet limited data exist on its incidence in Japan. We aimed to describe the incidence of all-cause and cause-specific LTSA by sex and age using 10-year data from a large Japanese working population, with a focus on sex differences.

Methods: The study participants were employees from 16 worksites in the Japan Epidemiology Collaboration on Occupational Health Study between April 2012 and March 2022. LTSA, defined as sickness absence from work lasting 30 days or more, was recorded at each worksite. The causes of LTSA were classified using the International Classification of Diseases, 10th Revision. Incidence rates for all-cause and cause-specific LTSA were calculated based on sex and age.

Results: During 730,391 and 161,513 person-years of follow-up, 6,518 and 1,866 spells of LTSA were recorded in males and females, respectively. Females had higher incidence rates of all-cause LTSA than males (115.5 vs 89.2 per 10,000 person-years), especially among females in their 20s and 30s. This was partly attributed to younger females experiencing higher LTSA incidence rates due to mental disorders, neoplasms, and pregnancy-related illnesses. In older age, females had higher LTSA incidence rates than males for musculoskeletal diseases and injuries/external causes, whereas LTSA incidence rates due to circulatory diseases were lower than those in males.

Conclusion: The incidence of total and cause-specific LTSA varied greatly by sex and age, highlighting the need to consider employees' characteristics in the prevention and management of LTSA.

Background: Socioeconomic inequalities in cancer screening participation remain a public health issue worldwide. We assessed trends in cancer screening participation according to socioeconomic status in Japan between 2013 and 2022, considering the potential impact of the coronavirus disease 2019 (COVID-19) pandemic.

Methods: Data from the nationally representative Comprehensive Survey of Living Conditions (2013-2022: approximately 500,000 persons per survey) were analyzed for age-standardized self-reported cancer screening rates for stomach, lung, colon, breast (aged 40-69 years), and cervical (aged 20-69 years) cancers, stratified by education levels. An age-adjusted Poisson model was used to assess the statistical significance of changes between the survey years.

Results: A clear socioeconomic gradient was observed, particularly in stomach cancer screening, where the 2022 rates ranged from 28.3% (low education) to 58.2% (high education) for men and 20.2% to 43.2% for women, depending on education level. Between 2019 and 2022, screening rates for stomach, lung, and colorectal cancers changed by -1.2%, -0.9%, and +0.6% for men and -1.0%, +0.1%, and +1.4% for women, respectively. Breast and cervical cancer screening rates declined by 0.5% and 0.4%, respectively. The COVID-19 pandemic worsened inequalities, with a 3.1% decline in breast cancer screening among individuals with low education level, compared to a 1.0% decline among those with higher education level.

Conclusion: The COVID-19 pandemic had a minor impact on screening rates (counteracting increasing trends of screening rates), except for colorectal cancer screening rates; however, the impact was relatively severe for individuals with lower socioeconomic status, especially for women.

Background: Policy decisions should be guided by measures that capture the impact of exposures on outcomes and that explicitly account for present-day exposure distribution. Both the preventable and attributable fractions have been used for this purpose; however, exposure effects can vary across subpopulations, and when this occurs, appropriate interpretation of these measures should be facilitated by a discussion of the contributions of different subpopulations.

Methods: We analyze preventable and attributable fractions in the presence of effect modification. In particular, we use potential outcomes to formally define these quantities and to clarify the weighting of different strata in the total population measures.

Results: Our derivations show that stratum-specific preventable and attributable fractions are weighted in proportion to the relative frequencies of effect modifiers among individuals with the outcome of interest. We also demonstrate that these weights are valid for the related quantities, preventable and attributable proportions. Finally, we present an example that illustrates how effect modification affects interpretation of these measures.

Conclusions: In sum, when effect modification is present, investigators should consider reporting these measures by the relevant population strata, and information that would allow quantification of their implicit weights in the total population estimate. Our study provides a formal justification for this approach.