Estimating transmission potential and risk factors for emerging infections often relies on incomplete surveillance data. When infections are missed or their sources are misclassified, estimates of the effective reproduction number (Rs) and associations between case attributes and infection source may be biased. We developed an analytic framework to quantify and correct these biases, treating imperfect observation and entanglement of transmission chains as forms of misclassification that distort estimates of transmission parameters. The bias depends on the probability of case observation (Pobs) and the extent of concurrent introductions (Rp). We illustrate this framework using surveillance data for mpox in the Democratic Republic of the Congo (1981-1986, 2013-2017), showing that incomplete observation leads to underestimation of Rs, whereas overlapping introductions can cause overestimation. Both types of bias also attenuate odds ratios for traits that distinguish primary (spillover) from secondary (human-tohuman) infections. Accounting for these biases provides context for the global emergence of mpox in 2022 and for how current transmission patterns differ from historical trends. These results demonstrate how standard epidemiologic concepts of misclassification and missingness can clarify bias in transmission studies and improve interpretation of surveillance data for emerging pathogens, including zoonoses and vaccine-preventable infections.

Epidemiologists increasingly estimate associations between mixtures of correlated co-exposures and health outcomes. We propose an exploratory approach to help differentiate co-exposure confounding from co-exposure amplification of bias (CAB) for studies considering correlated co-exposures using a case example and simulations based on observed relationships. Our example uses associations observed in published studies between co-occurring per- and polyfluoroalkyl substance (PFAS) biomarkers and tetanus antibody concentrations but could apply to any context where statistical modeling is performed with and without inclusion of correlated co-exposures. Using directed acyclic graphs (DAGs), to define an assumed causal structure, and simulations of three PFAS predicting antibody titers, we show that if researchers can assume one component in the exposure mixture is correlated with the other components, but does not cause the outcome, then CAB can be detected by testing whether the estimate for the non-causal component in a multi-pollutant model is different from zero. In our simulation, the estimate for the non-causal component was zero (95%CI, -0.22, 0.25) when CAB was absent and -0.25 (95%CI, -0.48, 0.00) when present. These results illustrate a possible method to identify CAB and select statistical models that may produce less biased effect estimates if our assumed DAG is correct.

Increasing oily fish intake has been proposed as a strategy for Alzheimer's disease (AD) prevention, but evidence for long-term benefit remains limited. Using observational data from 2012 cognitively healthy adults in the Framingham Heart Study Offspring cohort, we emulated a target trial of seven isocaloric fish intake strategies to estimate 21-year AD risk. We modeled repeated measures of diet and adjusted for pre-baseline diet, baseline covariates, and time-varying confounding using the parametric g-formula. The estimated 21-year AD risk ranged from 10.33% (95% confidence interval [CI]: 8.41%, 12.92%) under no fish intake to 9.98% (95% CI, 6.64%, 13.31%) under ≥ 3 servings/week. Risk ratios relative to no intervention were close to one and risk differences were close to zero across all strategies. Substituting fish for meat yielded similar results. We observed no clear long-term protective effect of sustained fish intake on AD risk, though exploratory analyses suggested a possible protective effect among older participants during earlier follow-up, warranting further investigation. This study illustrates the use of target trial emulation to estimate long-term dietary effects while accounting for time-varying confounding and competing risks in nutritional epidemiology.

Major disease diagnoses can create shocks that ripple through households. Population-level evidence in the U.S. on how spouses and partners fare after a partner's major diagnosis remains sparse and is often limited to caregivers or employment-related outcomes. Using the U.S. Health and Retirement Study (1992-2020), we evaluated population-level health spillovers to spouses after a partner's first diagnosis of cancer, stroke, or heart disease, using a stacked event-study design with matching (n = 29 166). Outcomes included subjective health, doctor-diagnosed conditions, limitations in basic and instrumental activities of daily living (ADL/IADL), and health behaviors and preventive care use. We evaluated heterogeneity by sociodemographic and economic conditions. Partners of newly diagnosed patients experienced modest but persistent deterioration: more frequent reports of poor health ($\boldsymbol{\beta}$=0.02, 95% confidence interval [CI]: 0.00, 0.03), more doctor-diagnosed conditions ($\boldsymbol{\beta}$=0.05, 95% CI: 0.01, 0.08) and higher depressive symptoms ($\boldsymbol{\beta}$=0.08, 95% CI: 0.01, 0.16). Effects were larger for women partnered to men with a diagnosis and for spouses with lower education or fewer assets. Health and preventive-care behaviors were largely unchanged. Findings motivate early, spouse-focused support and underscore the need for targeted interventions for spouses at highest risk.

Point-in-time measures of neighborhood poverty exposure suggest a link with physiological aging, but it remains unclear how histories and the timing of residential disadvantage connect to weathering. In this issue, D'Alessio et al. (Am J of Epidemiol. 2026) use 25 years of prospective residential data from the National Longitudinal Study of Adolescent to Adult Health (Add Health) to link neighborhood poverty across adolescence, early adulthood, and early midlife to three epigenetic clocks. They identify selective dose-response relationships. Each additional life stage in a high-poverty tract is associated with accelerated biological aging, and exposure is distributed with sharp racial inequality. The per-stage effect is modest in individual terms but consequential at population scale. Its unequal allocation makes residential segregation a plausible epigenetic mechanism of population-level racial disparities in aging. We argue that-despite its considerable strengths-the study leaves two questions unresolved. First, does health-selective residential mobility drive part of the association? Second, does the prominence of mid-life exposure reflect a true sensitive period, or the fact that mid-life poverty is measured in close proximity to biospecimen collection? We outline potential follow-up work that could address these questions via structured life course methods and better-powered cohorts with tract-linked data, including Add Health.

Heatwaves pose a growing global health threat, yet the mental health impacts of different heatwave types under climate change remain unclear. We quantified current and future mental disorder mortality attributable to daytime, nighttime, and compound heatwaves in Shanghai, China. Baseline associations between heatwaves and daily mental disorder deaths (2016-2019) were estimated using a time-stratified case-crossover design. Heatwaves were defined using a health-based excess heat factor, and attributable fractions in the 2010s were projected to the 2030s, 2060s, and 2090s under three Shared Socioeconomic Pathways and three population adaptation scenarios. All heatwave types were associated with increased mental disorder mortality, with the strongest effects observed for daytime heatwaves, followed by compound and nighttime heatwaves. Without adaptation, excess deaths attributable to daytime heatwaves were projected to increase by 8.3%-10.7% in the 2060s and 13.6%-15.3% in the 2090s, while the full exposure-response function projected smaller increases in attributable deaths for compound and nighttime heatwaves. Schizophrenia exhibited the greatest vulnerability, particularly to daytime heatwaves, with attributable deaths rising by 27.9%-34.9% by the 2090s. Suicide and dementia were most affected by compound and nighttime heatwaves, respectively. Partial and full adaptation substantially reduced future burdens, especially for nighttime heatwaves. Enhanced population adaptation could mitigate the death burden of mental disorders from heatwaves.

We sought to characterize racial identity response transitions among multiparous birthing people who ever identified as American Indian or Alaska Native (AIAN) and assess whether this fluidity alters year-specific statewide AIAN maternal health and birth outcome prevalence estimates. To do so, we used Michigan birth certificates (2014-2022) to identify individuals with $\boldsymbol{\ge}$2 live births who self-reported as AIAN for $\boldsymbol{\ge}\mathbf{1}$ birth (n = 4823). We categorized individuals as having changed or not changed their reported racial identity across births and recalculated AIAN-specific outcome estimates accordingly. Only 23.9% (1155) consistently identified as AIAN across all births, while 76.1% (3668) used different racial identity responses. Among those whose racial identity response changed, transitioning from identifying as AIAN to no longer identifying as AIAN was the most common pattern. Accounting for these transitions had minimal effect on prevalence estimates for preterm birth and gestational diabetes, while there were some prevalence changes for the prevalence of smoking before/during pregnancy. AIAN birthing people in Michigan often shift racial identity responses over time. Although this fluidity had little impact on surveillance of perinatal outcomes, reliance on point-in-time race data may inadequately reflect AIAN identity in public health data.

Neighborhood socioeconomic disadvantage is linked to accelerated biological aging using DNA methylation; however, limited knowledge exists of the relationship between duration or timing of exposure to neighborhood poverty across life stages and biological aging, measured by epigenetic clocks. Existing research has relied on samples with limited racial and gender diversity. We examined the relationship between multiple dimensions of neighborhood poverty (duration, life stage of first exposure, life stage exposure trajectories, and severity) and biological aging using data from 3637 participants followed for 25 years in the National Longitudinal Study of Adolescent to Adult Health. We employed survey linear regression models to estimate associations for three epigenetic clocks: GrimAge2, PhenoAge, DunedinPACE. In the total sample, only 8% of participants experienced neighborhood poverty across all life stages, which differed starkly by race (Non-Hispanic Blacks: 26% vs. Non-Hispanic Whites: 4%). We found that each additional life stage of exposure to neighborhood poverty was associated with significant biological age acceleration. Experiencing neighborhood poverty in mid-adulthood was consistently associated with accelerated biological aging. Our results suggest a relationship between duration and timing of exposure to neighborhood poverty and accelerated biological aging across key biological clocks, indicating possible lasting cellular level effects of residential context.

Patients with co-occurring diabetes and depression often have unmet healthcare needs. We evaluated the effect of the Community-based Integrated Care for Patients with Diabetes and Depression (CIC-PDD) model on healthcare utilization and costs. In this cluster-randomized trial conducted in community health centers in China, centers were assigned to enhanced usual care or integrated care. Adults with type 2 diabetes and depression (PHQ-9 ≥ 10) were recruited. A comprehensive care plan was implemented in the integrated care group. Healthcare utilization, costs, and self-management behaviors were assessed at 6 and 12 months using claims data and questionnaires. Outcomes included outpatient, inpatient, and self-treatment, including out-of-pocket (OOP) expenses. A total of 275 participants were assigned to the integrated care and 355 to the usual care. Compared with the control group, integrated care participants had more outpatient visits at 12 months (mean difference: 1.35; 95% CI, 0.45-2.24; P = 0.003) and higher self-treatment use (2.57; 95% CI, 0.02-5.13; P = 0.049). They also had higher self-treatment costs ($67.16; 95% CI, 9.45-124.87; P = 0.023) and OOP expenses ($127.21; 95% CI, 41.79-212.63; P = 0.004). These increases may reflect greater engagement with care and self-management, suggesting that CIC-PDD helped address unmet healthcare needs.

Disinvestment in segregated, low-income Black communities is rooted in structural racism and racial capitalism and contributes to racial health inequities. Virtual audits have been used to capture visible signs of neighborhood disinvestment, however, these methods have been primarily used and validated in dense urban cities in the Northeast and Western U.S. We used Google Street View to conduct virtual audits in the Jackson, Mississippi tri-county metropolitan area-a mid-sized, urban area in the U.S. South. Our approach engaged residents and community organizers in conversations to inform our interpretation of the local context and utilized exploratory and confirmatory factor analyses to develop a total disinvestment score for each observed street segment. In select neighborhoods, we also assessed neighborhood disinvestment longitudinally. We found that disinvestment scores were higher in historically Black neighborhoods and within Jackson city limits compared to surrounding, majority-white counties and increased over time in some neighborhoods. Our score was positively associated with housing vacancy, unemployment, residential segregation by race and income, and negatively associated with owner-occupied housing. Virtual audits can be valid tools for measuring physical signs of disinvestment in low-density cities in the US South, and for measuring contributions of the built environment to population health inequities.

There is limited research into the impact of takeaway food on metabolic dysfunction-associated steatotic liver disease (MASLD). This study aimed to evaluate the association between takeaway food consumption and MASLD in Chinese TCLSIH cohort (n = 3186) and UK Biobank (n = 99 781), plus 334 Chinese adults with biopsy-proven MASLD in the PERSONS cohort. Takeaway food consumption was assessed via questionnaires. MASLD was identified through liver ultrasonography and cardiometabolic factors (TCLSIH) and hospital records/death registries (UK Biobank). Cox proportional hazards regression models were used to evaluate hazard ratio (HR) and 95% confidence interval (CI). In the TCLSIH cohort, 724 participants developed MASLD within 12 941 person-years; the fully adjusted HRs (95% CIs) for MASLD across takeaway food consumption frequency were 1.00 (reference) for < 1 time/week, 0.92 (0.67, 1.27) for 1-3 times/week, 1.30 (1.07, 1.61) for ≥ 4 times/week in males. In the UK Biobank, 724 participants developed MASLD during 1 033 546 person-years; males eating takeaway food had a fully adjusted HR (95% CI) of 1.66 (1.08, 2.55) compared with non-consumers. In the PERSONS cohort, takeaway food consumption showed positive association with MASLD severity in males. These findings indicated that takeaway food consumption was positively associated with risk and severity of MASLD in adults, particularly among males.

Extreme Risk Protection Orders (ERPO), known as Gun Violence Restraining Orders in California, are designed to reduce gun violence by allowing courts to temporarily remove firearms from individuals who pose a risk to themselves or others. Population-level evidence on their impacts remains limited. We used the Longitudinal Study of Handgun Ownership and Transfer, which links California voter, mortality, and firearm transaction records, to evaluate the impact of California's ERPO law on homicide risk among cohabitants of handgun owners who are not themselves handgun owners. Using a difference-in-differences design, we compared changes in 1-year homicide risks before (2012-2015) and after (2016-2021) policy implementation among adults living with handgun owners versus those in households without known handgun owners. We estimated a 1-year difference of -0.42 (95% CI, -1.86, 1.02) homicide deaths per 100 000 persons among cohabitants of handgun owners who are not themselves handgun owners, with null findings robust across sensitivity analyses. Though our findings did not suggest substantial population-level protective effects, ERPOs may still reduce risk of homicide among cohabitants of ERPO respondents in ways not detectable at the population level. At the population level, their effectiveness may depend on more frequent or better-targeted use.

This study examined how race/ethnicity, sex/gender, and sexual orientation intersect under interlocking systems of oppression to socially pattern depression among US adults. With cross-sectional data from the 2015-2020 National Survey on Drug Use and Health (NSDUH; n=234 722), we conducted design-weighted multilevel analysis of individual heterogeneity and discriminatory accuracy (MAIHDA) under an intersectional framework to predict past-year and lifetime major depressive episode (MDE). With 42 intersectional groups constructed from seven race/ethnicity, two sex/gender, and three sexual orientation categories, we estimated age-standardized prevalence and excess/reduced prevalence attributable to two-way or higher interaction effects. Models revealed heterogeneity across groups, with prevalence ranging from 2.9-27.3% (past-year) and 5.9-43.1% (lifetime). Approximately 12.7% (past-year) and 12.5% (lifetime) of total individual variance were attributable to between-group differences, indicating key relevance of intersectional groups in describing the population distribution of depression. Main effects indicated, on average, people who were White, women, gay/lesbian, or bisexual had greater odds of MDE. Main effects explained most between-group variance. Interaction effects (past-year: 10.1%; lifetime: 16.5%) indicated a further source of heterogeneity around averages with groups experiencing excess/reduced prevalence compared to main effects expectations. We extend the MAIHDA framework to calculate nationally representative estimates from complex sample survey data using design-weighted, Bayesian methods.