American journal of epidemiology最新文献

Our objective was to examine the role of structural racism and economic disadvantage in perinatal health inequities using the Environmental influences on Child Health Outcomes Cohort. Participants' addresses were linked to area-level measures of life expectancy, education, unemployment, health insurance, jail rate, segregation, and housing cost burden. We created absolute measures to represent economic disadvantage and relative measures comparing values for Black or Latinx people to White people in the same area to represent structural racism. We used quantile G-computation to estimate the effects of a one-quartile increase in all exposures simultaneously on fetal growth and gestational age measures. A one-quartile increase in economic disadvantage was associated with a reduction in birthweight [(-25.65 grams, 95% CI (-45.83, -5.48)], but not gestational age [-0.02 weeks, 95% CI (-0.13, 0.09)]. With a one-quartile increase in Latinx-White structural racism, we observed reductions in birthweight [-80.83, 95% CI (-143.42, -18.23)] among Latinx participants. A one-quartile increase in Black-White structural racism was weakly associated with lower birthweight among Black participants [-15.70, 95% CI (-82.89, 51.48)] but was associated with higher birthweight among White participants [57.47, 95% CI (13.26, 101.67)]. Our findings suggest co-occurring forms of structural inequity likely influence racialized disparities in fetal growth outcomes.

Exposure to fine particulate matter (PM2.5) is associated with cardiometabolic risk among World Trade Center Health Program general responders, but current studies focus mainly on PM2.5 mass. We studied the associations between annual source-apportioned PM2.5 exposures, and self-reported diabetes or repeated blood glucose measurements among general responders enrolled between 2003 and 2019 (n = 34 764), residing in Tri-state area. We used non-negative matrix factorization to attribute PM2.5 component to sources (ie, biomass burning, oil combustion, metal industry, other industries, and motor vehicles). We used multivariable mixed-effect models to estimate the simultaneous associations of the source-apportioned PM2.5 exposures with the outcomes. We found (% change [95% CIs]) an interquartile range increase in PM2.5 attributed to metal industry sources (0.42 μg/m3) to be associated with an 8.35% (1.39%, 15.79%) higher risk of diabetes and a 1.31% (0.80%, 1.82%) increase in glucose levels. Source-specific associations with glucose were modified by sex, showing larger associations with biomass burning (1.08% [0.32%, 1.85%] per 0.44 μg/m3) and motor vehicle (1.34% [0.76%, 1.93%] per 0.92 μg/m3) pollution among women, and larger associations with oil-combustion (0.68% [0.03%, 1.34%] per 1.74 μg/m3) pollution among men. These findings can inform policies and interventions targeting emissions from these specific sources, particularly for general responders with a history of extreme air pollution exposures.

Health visiting is a complex public health intervention in which specialist nurses work with families to support the healthy development of children up to 5 years of age. Using routinely collected administrative health data, we emulated a target trial to estimate the effect of enhanced health visiting services on potentially avoidable hospital admissions for children born in 10 local areas in England between 2016 and 2019. We found that receiving additional support from the health visiting team in the early weeks of life was associated with an increased odds of a child experiencing a potentially avoidable hospitalization (OR, 1.28; 95% CI, 1.02-1.60). Health visiting may encourage families to seek secondary health care, for example by building confidence in public services or heightening parental anxiety about the risks of childhood health conditions. However, qualitative research and sensitivity analyses indicated that our effect estimate may have been subject to residual confounding, selection bias or both. An in-depth understanding of the intervention and the mechanisms through which treatments are assigned is essential for generating valid estimates of causal effects.

Knowledge of patterns in COVID-19 deaths by area-level income over time and the mediating role of vaccination in inequality patterns remains limited. We used data from a population-based retrospective cohort of 11 248 572 adults in Ontario, Canada. Cause-specific hazard models were used examine the relationship between income (2016 Census at the dissemination area level) and COVID-19 deaths between March 1, 2020 and January 30, 2022, stratified by wave. We used regression-based causal mediation analyses to examine the mediating role of vaccination in the relationship between income and COVID-19 deaths during waves 4 and 5. After accounting for demographics, baseline health, and other social determinants of health, inequalities in COVID-19 deaths by income persisted over time (HR [95% CI] comparing lowest income vs highest income quintiles were 1.37 [0.98-1.92] for wave 1, 1.21 [0.99-1.48] for wave 2, 1.55 [1.22-1.96] for wave 3, and 1.57 [1.15-2.15] for waves 4 and 5). By the start of wave 4, 7 534 259 (67.7%) of those alive were vaccinated, with lower odds of vaccination in the lowest income vs highest income quintiles (0.71 [0.70-0.71]). This inequality in vaccination accounted for 56.9% [22.5%-91.3%] of inequalities in COVID-19 deaths between individuals in the lowest income vs highest income quintiles. Efforts are needed to address vaccination gaps and residual heightened risks associated with lower income to improve health equity in COVID-19 outcomes.

Understanding the absolute risk of developing a second primary cancer is important to guide patient surveillance and education. We aimed to examine the cumulative incidence and factors associated with development of a second primary cancer (melanoma versus other) after diagnosis of a first primary melanoma (invasive or in situ). We analyzed a population-based study cohort of 154 695 people diagnosed with a first primary melanoma in New South Wales, Australia, between 1982 and 2019. The cohort was followed for future cancer incidence and vital status for a median of 7.0 years. We used Fine-Gray models to account for death as a competing risk. After a first primary melanoma, 23.7% developed a second primary cancer, including 12.7% who developed a second primary invasive or in situ melanoma (mean 5-year risk: 7.6%). The next most common second primary cancer types were prostate, breast and colon cancers, with mean 5-year risks after the initial melanoma diagnosis of 2.8% (male-specific incidence), 0.7% (2.8% female-specific incidence), and 0.6%, respectively. The most common second primary cancer among people with a first primary melanoma was another melanoma (invasive or in situ), requiring long-term careful surveillance of their skin even if the probability of recurrence from the first melanoma is low.

Our aim is to investigate the association between visit-to-visit variability of nine risk factors and incident cardiovascular disease (CVD) in a large multi-ethnic population cohort study. We used the Multi-Ethnic Study of Atherosclerosis cohort. We included individuals with no previous history of CVD, with at least three repeated measurements on each risk factor including total cholesterol, high-density lipoprotein cholesterol (HDL-C), low-density lipoprotein cholesterol (LDL-C), non-HDL-C, triglyceride, Chol/HDL-C ratio, diastolic blood pressure, systolic blood pressure (SBP), and body mass index (BMI). Visit-to-visit variability was estimated via the variability independent of the mean. A Cox proportional hazards model was used to estimate the association between visit-to-visit variability and the hazard of developing CVD. There was a statistically significant association between visit-to-visit variability in SBP, BMI, HDL-C, and the rate of incident CVD. This rate was higher in individuals with high visit-to-visit variability for SBP (HR, 1.28; 95% CI, 1.01-1.63; P = .04), BMI (HR, 1.58; 95% CI, 1.25-2.00; P < .001), and HDL-C (HR, 1.3; 95% CI, 1.03-1.65; P = .025), compared to those with low visit-to-visit variability. Our findings suggest that visit-to-visit variability in some CVD risk factors could be independently associated with incident CVD and may be useful to clinicians in risk stratification.

Time-to-event outcomes are widely used in clinical and epidemiological research. For instance, studies of medical product safety often involve comparative analyses of rare time-to-event outcomes. The effects of misclassified outcomes and error in survival times for time-to-event data have not been widely investigated. In this Monte Carlo simulation study, we compared the relative bias of absolute and relative measures of effect under varying degrees of outcome misclassification, outcome incidences, direction of error in survival times, and the time point of inference. Relative measures of effect were susceptible to considerable downward bias, which was larger when the outcome incidence and specificity were lower, error in survival times led to earlier times, time point of inference was earlier, and the estimation excluded samples for which an estimate could not be obtained. For absolute measures of effect, the pattern of bias was much simpler, greater downward bias was primarily a function of the degree of sensitivity. The results suggest when the outcome incidence is rare, specificity and sensitivity are high, absolute measures of effect may be preferable to relative measures of effect.

Both ambient air pollution exposure and biological aging are associated with incident liver diseases, but previous studies mainly focused on single-factor associations. This study aimed to assess the combined effects of air pollutant exposure and biological aging on liver diseases incidence and investigate the potential mediating role of biological aging. We analyzed 418 576 UK Biobank participants. Annual mean concentrations of PM2.5, PM10, PM2.5-10, NO2, and NO in 2010 were used to generate a weighted air pollution score. Biological ages were assessed using the Klemera-Doubal method biological age (KDM-BA) and phenotypic age (PhenoAge). Cox regression models and quantile g-computation were used to evaluate associations and joint effects. Mediation analyses explored the role of biological aging. Over a median follow-up of 13.57 years, 7991 (1.91%) participants developed liver diseases. Exposure to all pollutants and biological aging were associated with higher liver diseases risk. And NO2 contributed 42.31% to the mixture effect. Participants with higher levels of air pollutant exposure and biologically older status had a higher risk. Furthermore, the mediated proportion of accelerated biological aging was 1.9% to 7.7% for air pollution-associated liver diseases. Ambient air pollution exposure may increase liver diseases risk, with biological aging potentially involved in the mechanisms.

Purpose: To describe a novel data ecosystem and quantify improvements in documented COVID-19 vaccine uptake using closed claims data and state vaccine registries.

Methods: De-identified data from administrative claims were linked via tokenization to state vaccine registry data from two states with mandatory vaccine reporting. COVID-19 vaccine uptake in claims alone versus supplemented with registry data was measured separately for original wildtype, BA.4/5-adapted bivalent, and XBB.1.5-adapted formulations.

Results: There were over 9 million individuals included in each formulation distribution study period. Using claims data alone versus with additional state registry data, 20% versus 57% of individuals had documented receipt of a primary wildtype series (≥2 doses of two-dose or ≥ 1 of single-dose series), 7% versus 10% BA.4/5-adapted bivalent, and 6% versus 8% XBB.1.5-adapted COVID-19 vaccine. Most claims-based vaccination events were captured with drug, rather than procedure, codes. Individuals with Medicaid-based insurance were less likely to be captured as vaccinated in claims data.

Conclusions: The addition of state vaccine registries increased COVID-19 vaccine capture by 182% for primary wildtype series, 41% BA.4/5-adapted bivalent, and 38% XBB.1.5-adapted formulations. Tokenization of claims data to vaccine registries presents an opportunity for more accurate estimates of vaccine uptake and effectiveness than claims data alone.

Adolescents' menstrual cycle characteristics can be "vital signs" of health and impact quality of life. While endocrine-disrupting pesticides are commonly used in agriculture, limited research exists on how exposure might affect the adolescent menstrual cycle. We examined the association between prenatal residential proximity to 11 agricultural pesticides and menstrual cycle characteristics at 16 years of age among 273 Latina adolescents from the Center for the Health Assessment of Mothers and Children of Salinas study. We estimated prenatal pesticide exposure by linking maternal residential addresses to California's pesticide use reporting database. Menstrual characteristics, including cycle length irregularities, painful menstruation, and heavy bleeding, were evaluated through a questionnaire. We used generalized linear models to evaluate exposure-outcome associations 1 pesticide at a time. To adjust for co-exposure to pesticides, we used Bayesian hierarchical models to include all pesticide exposures in 1 model. In our single-exposure model, we observed increased odds of heavy bleeding (OR, 1.29; 95% CI, 1.01-1.64) for each doubling in prenatal methomyl exposure. This association persisted in our joint exposure model (OR, 1.09; CrI, 0.99-1.19). Our results suggest prenatal exposure to endocrine-disrupting pesticides may impact certain adolescent menstrual cycle characteristics.