American journal of epidemiology最新文献

Target trial emulation (TTE) is a popular framework for observational studies based on electronic health records (EHR). A key component of this framework is determining the patient population eligible for inclusion in both a target trial of interest and its observational emulation. Missingness in variables that define eligibility criteria, however, presents a major challenge towards determining the eligible population when emulating a target trial with an observational study. In practice, patients with incomplete data are almost always excluded from analysis despite the possibility of selection bias, which can arise when subjects with observed eligibility data are fundamentally different than excluded subjects. Despite this, to the best of our knowledge, very little work has been done to mitigate this concern. In this paper, we propose a novel conceptual framework to address selection bias in TTE studies, tailored towards time-to-event endpoints, and describe estimation and inferential procedures via inverse probability weighting (IPW). Under an EHR-based simulation infrastructure, developed to reflect the complexity of EHR data, we characterize common settings under which missing eligibility data poses the threat of selection bias and investigate the ability of the proposed methods to address it. Finally, using EHR databases from Kaiser Permanente, we demonstrate the use of our method to evaluate the effect of bariatric surgery on microvascular outcomes among a cohort of severely obese patients with Type II diabetes mellitus (T2DM).

Rotavirus vaccine appears to perform sub-optimally in countries with higher rotavirus burden. We hypothesized that differences in the magnitude of rotavirus exposures may bias vaccine efficacy (VE) estimates, so true differences in country-specific rotavirus VE would be exaggerated without accommodating differences in exposure. We estimated VE against any-severity and severe rotavirus gastroenteritis (RVGE) using Poisson regression models fit to pooled individual-level data from Phase II and III monovalent rotavirus vaccine trials conducted between 2000 and 2012. The standard approach model included terms for vaccination, country, and a vaccination-country interaction. Other models used proxies for exposure magnitude like severe RVGE rate or age at severe RVGE instead of country. Country-specific proxies were calculated from placebo group data or extracted from an external meta-analysis. Analyses included 83,592 infants from 23 countries in the Americas, Europe, Africa, and Asia. Using the standard approach, VE against severe RVGE substantially varied (10-100%). Using the severe RVGE rate proxy brought VE from all but two countries between 80% and 86%. Heterogeneity for VE against any-severity RVGE was similarly attenuated. Adjusting for exposure proxies reduced heterogeneity in country-specific rotavirus VE estimates. This phenomenon may extend to other vaccines against partially immunizing pathogens with global disparities in burden.

Cross-national comparisons of dementia prevalence are essential for identifying unique determinants and cultural-specific risk factors, but methodological differences in dementia classification across countries hinder global comparisons. This study maps the 10/66 algorithm for dementia classification, widely used and validated in low- and middle-income countries (LMICs), to the U.S. Aging, Demographics, and Memory Study (ADAMS), the dementia sub-study of the Health and Retirement Study, and assesses its performance in ADAMS. We identified the subset of 10/66 algorithm items comparably measured in ADAMS, then used these items to re-train the 10/66 algorithm against the ADAMS clinical dementia diagnosis, employing k-fold cross-validation to assess performance. We compared the modified 10/66 algorithm to four other dementia classification algorithms previously validated in ADAMS, both for overall dementia estimation as well as for estimating education gradients. The modified 10/66 algorithm had higher sensitivity (87%) and specificity (93%) than the comparison algorithms. All of the algorithms over-estimated the education gradient in dementia, although the modest ADAMS sample size precludes precise comparisons of education gradient accuracy. Overall, we found that the modified 10/66 algorithm performs well in classifying dementia status in the U.S. Our results support the validity of risk factor comparisons between U.S. and 10/66 LMIC dementia datasets.

Case-control studies of sun exposure and ultraviolet radiation (UVR) have consistently reported inverse associations with non-Hodgkin lymphoma (NHL) risk, but prospective studies have yielded mixed results. Few studies have explored these exposures in relation to multiple myeloma (MM) risk. To further evaluate these associations with NHL and MM risk and identify etiologically relevant exposure timing, we pooled data on 566,693 individuals from 6 United States (U.S.) prospective cohort studies (11,636 incident NHL; 2,749 incident MM; median follow-up: 20 years) and used geographic information systems models to estimate residential ambient UVR levels at time points from birth to adulthood. Using Cox proportional hazards models, we calculated hazard ratios (HRs) and 95% confidence intervals (CI) for associations of residential ambient UVR levels with NHL overall, NHL subtypes, and MM, adjusted for study, age and other putative risk factors. No UVR measures were significantly associated with NHL or NHL subtypes. Higher residential UVR levels during cohort follow-up were inversely associated with MM overall and among females (longitudinally-updated HR per interquartile range increase: 0.74; 95% CI: 0.63, 0.86) but not males (1.08; 0.90, 1.29). Our results do not confirm an inverse association of adult ambient UVR levels with NHL risk. The MM findings require further investigation.

Attributable burden of disease estimates reported population-wide do not reflect social disparities in exposures and outcomes. This makes one of the influential scientific tools in public health decision-making insensitive to the distribution of health impacts between socioeconomic groups. Our aim was to use the often-overlooked distributive property of the population attributable fraction (PAF) to quantitatively partition the population burden attributed to know risk factors into subgroups defined by their socioeconomic position (SEP). To illustrate our approach, we focus on lung cancer risk in relation to smoking and exposure to three occupational carcinogens: asbestos, silica dust and diesel motor exhaust. We directly estimate PAFs from a large population-based case-control study using multiple unconditional logistic regression, mutually adjusting for available known risk factors. We partition the PAFs of occupational exposures and smoking according to different SEP indicators: occupational class, prestige and trajectory, and education. Our results show that workplace exposures, smoking and their population health impacts concentrate among lower-SEP groups, a long-known reality that had never been measured through a PAF approach. While attempting to quantify the avoidable burden of diseases, it is useful to partition population-wide into SEP-specific metrics, as the modifiable exposures (behavioural, work-related, environmental) are socially stratified.

Few physical performance batteries exist that appropriately evaluate physical limitations in middle-aged adults. We aimed to develop a physical performance summary score that is appropriate for use in epidemiology studies of middle-aged adults using data from the Coronary Artery Risk Development in Young Adults (CARDIA) Function study, which assessed self-reported function (PROMIS-SF20a) and physical performance measures (gait speed, balance, lower-body strength, grip strength, and cardiovascular endurance). The CARDIA Physical Performance (CAPP) score was developed using sex-specific quartiles, assigning points based on these quartiles (0 for not attempted, 1-4 for each quartile), and summing points across all performance measures (0-20, higher scores reflect greater performance). We also examined the relationship between CAPP score and other function-related measures (physical activity, quality of life, sedentary behavior, body mass index, and waist circumference). Among 2,021 CARDIA Function participants [mean age: 60.0±3.6 years; 58% female; 44% Black] a 1-point higher CAPP score (μ= 12.3±4.1) was associated with a 0.85 higher PROMIS-SF20a score (β= 0.85, p< 0.001). CAPP score had a canonical correlation coefficient of 0.63 (p< 0.0001) suggesting a strong correlation with other function-related measures. CAPP score captured a wide range of physical performance and was correlated with self-reported function and other function-related measures.

Multimorbidity data is typically analysed by tallying disease counts, which overlooks nuanced relationships among conditions. We identified clusters of multimorbidity and subpopulations with varying risks and examined their association with all-cause mortality using a data-driven approach. We analysed 8-year follow-up data of people ≥35 years who were part of the CRONICAS Cohort Study, a multisite cohort from Peru. First, we used Partitioning Around Medoids and multidimensional scaling to identify multimorbidity clusters. We then estimated the association between multimorbidity clusters and all-cause mortality. Second, we identified subpopulations using finite mixture modelling. Our analysis revealed three clusters of chronic conditions: respiratory (cluster 1: bronchitis, COPD and asthma), lifestyle, hypertension, depression and diabetes (cluster 2), and circulatory (cluster 3: heart disease, stroke and peripheral artery disease). While only the cluster comprising circulatory diseases showed a significant association with all-cause mortality in the overall population, we identified two latent subpopulations (named I and II) exhibiting differential mortality risks associated with specific multimorbidity clusters. These findings underscore the importance of considering multimorbidity clusters and sociodemographic characteristics in understanding mortality risks. They also highlight the need for tailored interventions to address the unique needs of different subpopulations living with multimorbidity to reduce mortality risks effectively.

From 2001-2021, the age- and sex-adjusted Veteran suicide rate increased 76.3%. Surveillance of suicidal ideation (SI) and non-fatal suicidal self-directed violence (NF-SSDV) is a critical component of public health-oriented suicide prevention efforts. To facilitate national NF-SSDV surveillance, a biennial, population-based survey was initiated: Assessing Social and Community Environments with National Data (ASCEND) for Veteran Suicide Prevention. 17,396 Veterans participated in the first large-scale ASCEND survey (2022). This manuscript reports on SI and NF-SSDV prevalence among Veterans residing in the continental U.S., Hawaii, and Puerto Rico. Lifetime SI was reported by 31.98% (95%CI=30.97-32.99), post-military SI by 25.88% (95%CI=24.91-26.85) and past-year SI by 12.69% (95%CI=11.90-13.47). The most commonly considered SI method among those with past-year SI was gunshot. Additionally, 34.42% (95%CI=33.07-35.78) of Veterans with lifetime SI reported lifetime preparatory behaviors. Moreover, 9.13% (95%CI=8.43-9.82) of Veterans reported lifetime interrupted attempts. Lifetime suicide attempts (SA) were reported by 6.99% (95%CI=6.41-7.56) of Veterans, with 4.88% (95%CI=4.39-5.36) reporting post-military SA. The most common method in prior attempts was medication overdose. ASCEND provides a novel opportunity to elucidate the prevalence of SI and different types of NF-SSDV in the Veteran population. Recurring administration will elucidate changes in SI and NF-SSDV prevalence in the Veteran population over time.

Observational studies have reported that hearing aid (HA) use is associated with a reduced risk of dementia diagnosis, suggesting a possible protective effect. However, extant observational studies do not explicitly model causal effects, while randomised controlled trials on the effect of HA on dementia exhibit short follow-up. Here we used self-report, hearing tests, and healthcare records in UK Biobank to design a hypothetical intervention for the effect of HA use on the risk of dementia diagnosis in people with incident hearing loss (HL). HA users exhibited a higher risk of dementia diagnosis than non-users (RR=1.43, 95%CI=1.08-1.88). Associations between HA use and dementia diagnosis were robust across sensitivity analyses (RRs: 1.34-1.59) but adjustment for primary healthcare utilisation (0.77, 0.44-1.33) or primary and secondary care utilisation (0.68, 0.39-1.18) substantially decreased the observed effect. The decrease in effect estimates upon adjustment for primary (1.30, 0.95-1.78) and primary and secondary healthcare utilisation (1.30, 0.94-1.78) was smaller when participants with relatively early diagnoses of HL were included in the sample compared to when they were not. While the findings are not conclusive, they suggest residual confounding by healthcare utilisation and dating of HL diagnosis in participants without primary care data in UK Biobank.