Search

Most prior work in quantitative approaches to generalizability and transportability emphasizes extending causal effect estimates from randomized trials to target populations. Extending findings from observational studies is also of scientific interest, and identifiability assumptions and estimation methods differ from randomized settings when there is selection on both the exposure and exposure-outcome mediators in combination with exposure-outcome confounders (and both confounders and mediators can modify exposure-outcome effects). We argue that this causal structure is common in observational studies, particularly in the field of life course epidemiology, for example, when extending estimates of the effect of an early-life exposure on a later-life outcome from a cohort enrolled in midlife or late life. We describe identifiability assumptions and identification using observed data in such settings, highlighting differences from work extending findings from randomized trials. We describe statistical methods, including weighting, outcome modeling, and doubly robust approaches, to estimate potential outcome means and average treatment effects in the target population and illustrate performance of the methods in a simulation study. We show that in the presence of selection into the study sample on both exposure and confounders, estimators must be able to address confounding in the target population. When there is also selection on mediators of the exposure-outcome relationship, estimators need to be able to use different sets of variables to account for selection (including the mediator), and confounding. We discuss conceptual implications of our results as well as highlight unresolved practical questions for applied work to extend findings from observational studies to target populations., Competing Interests: Disclosure: The authors report no conflicts of interest., (Copyright © 2024 Wolters Kluwer Health, Inc. All rights reserved.)

Background: Observational studies link high midlife systolic blood pressure to increased dementia risk. However, synthesis of evidence from randomized controlled trials has not definitively demonstrated that antihypertensive medication use reduces dementia risk. Here, we emulate target trials of antihypertensive medication initiation on incident dementia using three cohort studies, with attention to potential violations of necessary assumptions., Methods: We emulated trials of antihypertensive medication initiation on incident dementia using data from the Atherosclerosis Risk in Communities (ARIC) study, Cardiovascular Health Study (CHS), and Health and Retirement Study (HRS). We used data-driven methods to restrict participants to initiators and non-initiators with overlap in propensity scores and positive control outcomes to look for violations of positivity and exchangeability assumptions., Results: Analyses were limited by the small number of cohort participants who met eligibility criteria. Associations between antihypertensive medication initiation and incident dementia were inconsistent and imprecise (ARIC: HR = 0.30 [0.05, 1.93]; CHS: HR = 0.66 [0.27, 1.64]; HRS: HR = 1.09 [0.75, 1.59]). More stringent propensity score restriction had little effect on findings. Sensitivity analyses using a positive control outcome unexpectedly suggested antihypertensive medication initiation increased risk of coronary heart disease in all three samples., Conclusions: Positive control outcome analyses suggested substantial residual confounding in effect estimates from our target trials, precluding conclusions about the impact of antihypertensive medication initiation on dementia risk through target trial emulation. Formalized processes for identifying violations of necessary assumptions will strengthen confidence in target trial emulation and avoid inappropriate confidence in emulated trial results., Competing Interests: Conflicts of Interest:, (Copyright © 2024 Wolters Kluwer Health, Inc. All rights reserved.)

Background: Education is strongly associated with cognitive outcomes at older ages, yet the extent to which these associations reflect causal effects remains uncertain due to potential confounding., Methods: Leveraging changes in historical measures of state-level education policies as natural experiments, we estimated the effects of educational attainment on cognitive performance over 10 years in 20,248 non-Hispanic Black and non-Hispanic White participants, aged 45+ in the REasons for Geographic and Racial Disparities in Stroke (REGARDS) cohort (2003-2020) by (1) using state- and year- specific compulsory schooling laws, school-term length, attendance rate, and student-teacher ratio policies to predict educational attainment for US Census microsample data from 1980 and 1990, and (2) applying policy-predicted years of education (PPYEd) to predict memory, verbal fluency, and a cognitive composite. We estimated overall and race- and sex-specific effects of PPYEd on level and change in each cognitive outcome using random intercept and slope models, adjusting for age, year of first cognitive assessment, and indicators for state of residence at age 6., Results: Each year of PPYEd was associated with higher baseline cognition (0.11 standard deviation [SD] increase in composite measure for each year of PPYEd, 95% confidence interval [CI]: 0.07, 0.15). Subanalyses focusing on individual cognitive domains estimate the largest effects of PPYEd on memory. PPYEd was not associated with rate of change in cognitive scores. Estimates were similar across Black and White participants and across sex., Conclusions: Historical policies shaping educational attainment are associated with better later life memory, a major determinant of dementia risk., Competing Interests: Conflict of Interest: None declared, (Copyright © 2024 Wolters Kluwer Health, Inc. All rights reserved.)

Life course epidemiology is hampered by the absence of large studies with exposures and outcomes measured at different life stages in the same individuals. We describe when the effect of an exposure ( A ) on an outcome ( Y ) in a target population is identifiable in a combined ("synthetic") cohort created by pooling an early-life cohort including measures of A with a late-life cohort including measures of Y . We enumerate causal assumptions needed for unbiased effect estimation in the synthetic cohort and illustrate by simulating target populations under four causal models. From each target population, we randomly sampled early- and late-life cohorts and created a synthetic cohort by matching individuals from the two cohorts based on mediators and confounders. We estimated the effect of A on Y in the synthetic cohort, varying matching variables, the match ratio, and the strength of association between matching variables and A . Finally, we compared bias in the synthetic cohort estimates when matching variables did not d-separate A and Y to the bias expected in the original cohort. When the set of matching variables includes all variables d-connecting exposure and outcome (i.e., variables blocking all backdoor and front-door pathways), the synthetic cohort yields unbiased effect estimates. Even when matching variables did not fully account for confounders, the synthetic cohort estimate was sometimes less biased than comparable estimates in the original cohort. Methods based on merging cohorts may hasten the evaluation of early- and mid-life determinants of late-life health but rely on available measures of both confounders and mediators., Competing Interests: Disclosure: The authors report no conflicts of interest., (Copyright © 2024 Wolters Kluwer Health, Inc. All rights reserved.)

Background: We evaluated whether participants in the landmark Action to Control Cardiovascular Risk in Diabetes (ACCORD) trial represent US adults aged ≥40 with diabetes., Methods: Using the nationally representative 2017-2020 prepandemic National Health and Nutrition Examination Survey data, we made operational definitions of ACCORD eligibility criteria. We calculated the percentage of individuals aged ≥40 with diabetes and HbA1c ≥ 6.0% or ≥ 7.5% who met operational ACCORD eligibility criteria., Results: Applying survey sampling weights to 715 National Health and Nutrition Examination Survey participants aged ≥40 with diabetes and HbA1c ≥ 6.0% (representing 29,717,406 individuals), 12% (95% confidence interval [CI] = 8%, 18%) met the operational ACCORD eligibility criteria. Restricting to HbA1c ≥ 7.5%, 39% (95% CI = 28%, 51%) of respondents met the operational ACCORD eligibility criteria., Conclusions: ACCORD represented a minority of US middle-aged and older adults with diabetes. Given the differential risk profile between ACCORD participants and the general population with diabetes, extrapolating the trial findings may not be appropriate., Competing Interests: The authors report no conflicts of interest., (Copyright © 2024 Wolters Kluwer Health, Inc. All rights reserved.)

Background: Individuals of Mexican ancestry in the United States experience substantial socioeconomic disadvantages compared with non-Hispanic white individuals; however, some studies show these groups have similar dementia risk. Evaluating whether migration selection factors (e.g., education) associated with risk of Alzheimer disease and related dementia (ADRD) explain this paradoxical finding presents statistical challenges. Intercorrelation of risk factors, common with social determinants, could make certain covariate patterns very likely or unlikely to occur for particular groups, which complicates their comparison. Propensity score (PS) methods could be leveraged here to diagnose nonoverlap and help balance exposure groups., Methods: We compare conventional and PS-based methods to examine differences in cognitive trajectories between foreign-born Mexican American, US-born Mexican American, and US-born non-Hispanic white individuals in the Health and Retirement Study (1994-2018). We examined cognition using a global measure. We estimated trajectories of cognitive decline from linear mixed models adjusted for migration selection factors also associated with ADRD risk conventionally or with inverse probability weighting. We also employed PS trimming and match weighting., Results: In the full sample, where PS overlap was poor, unadjusted analyses showed both Mexican ancestry groups had worse baseline cognitive scores but similar or slower rates of decline compared with non-Hispanic white adults; adjusted findings were similar, regardless of method. Focusing analyses on populations where PS overlap was improved (PS trimming and match weighting) did not alter conclusions., Conclusions: Attempting to equalize groups on migration selection and ADRD risk factors did not explain paradoxical findings for Mexican ancestry groups in our study., Competing Interests: Disclosure: The authors report no conflicts of interest., (Copyright © 2023 Wolters Kluwer Health, Inc. All rights reserved.)

Background: Cannabis outlets may affect health and health disparities. Local governments can regulate outlets, but little is known about the effectiveness of local policies in limiting outlet densities and discouraging disproportionate placement of outlets in vulnerable neighborhoods., Methods: For 241 localities in California, we measured seven policies pertaining to density or location of recreational cannabis outlets. We geocoded outlets using web-scraped data from the online finder Weedmaps between 2018 and 2020. We applied Bayesian spatiotemporal models to evaluate associations of local cannabis policies with Census block group-level outlet counts, accounting for confounders and spatial autocorrelation. We assessed whether associations differed by block group median income or racial-ethnic composition., Results: Seventy-six percent of localities banned recreational cannabis outlets. Bans were associated with fewer outlets, particularly in block groups with higher median income, fewer Hispanic residents, and more White and Asian residents. Outlets were disproportionately located in block groups with lower median income [posterior RR (95% credible interval): 0.76 (0.70, 0.82) per $10,000], more Hispanic residents [1.05 (1.02, 1.09) per 5%], and fewer Black residents [0.91 (0.83, 0.98) per 5%]. For the six policies in jurisdictions permitting outlets, two policies were associated with fewer outlets and two with more; two policy associations were uninformative. For these policies, we observed no consistent heterogeneity in associations by median income or racial-ethnic composition., Conclusions: Some local cannabis policies in California are associated with lower cannabis outlet densities, but are unlikely to deter disproportionate placement of outlets in racial-ethnic minority and low-income neighborhoods., Competing Interests: The authors report no conflicts of interest., (Copyright © 2022 The Author(s). Published by Wolters Kluwer Health, Inc.)

Background: Identifying determinants of cognitive decline is crucial for developing strategies to prevent Alzheimer's disease and related dementias. However, determinants of cognitive decline remain elusive, with inconsistent results across studies. One reason could be differential survival. Cognitive decline and many exposures of interest are associated with mortality making survival a collider. Not accounting for informative attrition can result in survival bias. Generalized estimating equations (GEE) and linear mixed-effects model (LME) are commonly used to estimate effects of exposures on cognitive decline, but both assume mortality is not informative. Joint models combine LME with Cox proportional hazards models to simultaneously estimate cognitive decline and the hazard of mortality., Methods: Using simulations, we compared estimates of the effect of a binary exposure on rate of cognitive decline from GEE, weighted GEE using inverse-probability-of-attrition weights, and LME to joint models under several causal structures of survival bias., Results: We found that joint models with correctly specified relationship between survival and cognition performed best, producing unbiased estimates and appropriate coverage. Even those with misspecified relationship between survival and cognition showed advantage under causal structures consistent with survival bias. We also compared these models in estimating the effect of education on cognitive decline after dementia diagnosis using Framingham Heart Study data. Estimates of the effect of education on cognitive decline from joint models were slightly attenuated with similar precision compared with LME., Conclusions: In our study, joint models were more robust than LME, GEE, and weighted GEE models when evaluating determinants of cognitive decline., Competing Interests: The authors report no conflicts of interest., (Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.)

Background: Efforts to explain the burden of cardiovascular disease (CVD) often focus on genetic factors or social determinants of health. There is little evidence on the comparative predictive value of each, which could guide clinical and public health investments in measuring genetic versus social information. We compared the variance in CVD-related outcomes explained by genetic versus socioeconomic predictors., Methods: Data were drawn from the Health and Retirement Study (N = 8,720). We examined self-reported diabetes, heart disease, depression, smoking, and body mass index, and objectively measured total and high-density lipoprotein cholesterol. For each outcome, we compared the variance explained by demographic characteristics, socioeconomic position (SEP), and genetic characteristics including a polygenic score for each outcome and principal components (PCs) for genetic ancestry. We used R-squared values derived from race-stratified multivariable linear regressions to evaluate the variance explained., Results: The variance explained by models including all predictors ranged from 3.7% to 14.3%. Demographic characteristics explained more than half this variance for most outcomes. SEP explained comparable or greater variance relative to the combination of the polygenic score and PCs for most conditions among both white and Black participants. The combination of SEP, polygenic score, and PCs performed substantially better, suggesting that each set of characteristics may independently contribute to the prediction of CVD-related outcomes. Philip R. Lee Institute for Health Policy Studies, Department of Family & Community Medicine, UCSF., Conclusions: Focusing on genetic inputs into personalized medicine predictive models, without considering measures of social context that have clear predictive value, needlessly ignores relevant information that is more feasible and affordable to collect on patients in clinical settings. See video abstract at, http://links.lww.com/EDE/B879., (Copyright © 2021 Wolters Kluwer Health, Inc. All rights reserved.)

Background: Integrating results from multiple samples is often desirable, but privacy restrictions may preclude full data pooling, and most datasets do not include fully harmonized variable sets. We propose a simulation-based method leveraging partial information across datasets to guide creation of synthetic data based on explicit assumptions about the underlying causal structure that permits pooled analyses that adjust for all desired confounders in the context of privacy restrictions., Methods: This proof-of-concept project uses data from the Health and Retirement Study (HRS) and Atherosclerosis Risk in Communities (ARIC) study. We specified an estimand of interest and a directed acyclic graph (DAG) summarizing the presumed causal structure for the effect of glycated hemoglobin (HbA1c) on cognitive change. We derived publicly reportable statistics to describe the joint distribution of each variable in our DAG. These summary estimates were used as data-generating rules to create synthetic datasets. After pooling, we imputed missing covariates in the synthetic datasets and used the synthetic data to estimate the pooled effect of HbA1c on cognitive change, adjusting for all desired covariates., Results: Distributions of covariates and model coefficients and associated standard errors for our model estimating the effect of HbA1c on cognitive change were similar across cohort-specific original and preimputation synthetic data. The estimate from the pooled synthetic incorporates control for confounders measured in either original dataset., Discussion: Our approach has advantages over meta-analysis or individual-level pooling/data harmonization when privacy concerns preclude data sharing and key confounders are not uniformly measured across datasets., Competing Interests: The authors report no conflicts of interest. None of the funders had any role in design and conduct of the study; management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; or decision to submit the manuscript for publication., (Copyright © 2021 Wolters Kluwer Health, Inc. All rights reserved.)

Background: Evaluating the long-term health consequences of migration requires longitudinal data on migrants and non-migrants to facilitate adjustment for time-varying confounder-mediators of the effect of migration on health., Methods: We merged harmonized data on subjects aged 50+ from the US-based Health and Retirement Study (HRS) and the Mexican Health and Aging Study (MHAS). Our exposed group includes MHAS-return migrants (n = 1555) and HRS Mexican-born migrants (n = 924). Our unexposed group includes MHAS-never migrants (n = 16,954). We constructed a lifecourse data set from birth (age 0) until either age at migration to the United States or age at study entry. To account for confounding via inverse probability of treatment weights (IPTW), we modeled the probability of migration at each year of life using time-varying pre-migration characteristics. We then evaluated the effect of migration on mortality hazard estimated with and without IPTW., Results: Mexico to the United States migration was predicted by time-varying factors that occurred before migration. Using measured covariates at time of enrollment to account for selective migration, we estimated that, for women, migrating reduces mortality risk by 13%, although this estimate was imprecise and results were compatible with either large protective or deleterious associations (hazard ratio [HR] =0.87, 95% confidence interval [CI]: 0.60, 1.27). When instead using IPTWs, the estimated effect on mortality was similarly imprecise (HR = 0.98, 95% CI: 0.77, 1.25). The relationship among men was similarly uncertain in both models., Conclusions: Although time-varying social factors predicted migration, IPTW weighting did not affect our estimates. Larger samples are needed to precisely estimate the health effects of migration.

Directed acyclic graphs (DAGs), a prominent tool for expressing assumptions in epidemiologic research, are most useful when the hypothetical data generating structure is correctly encoded. Understanding a study's data generating structure and translating that data structure into a DAG can be challenging, but these skills are often glossed over in training. Campbell and Stanley's framework for causal inference has been extraordinarily influential in social science training programs but has received less attention in epidemiology. Their work, along with subsequent revisions and enhancements based on practical experience conducting empirical studies, presents a catalog of 37 threats to validity describing reasons empirical studies may fail to deliver causal effects. We interpret most of these threats to study validity as suggestions for common causal structures. Threats are organized into issues of statistical conclusion validity, internal validity, construct validity, or external validity. To assist epidemiologists in drawing the correct DAG for their application, we map the correspondence between threats to validity and epidemiologic concepts that can be represented with DAGs. Representing these threats as DAGs makes them amenable to formal analysis with d-separation rules and breaks down cross-disciplinary language barriers in communicating methodologic issues.

Background: Migration of adult children may impact the health of aging parents who remain in low- and middle-income countries. Prior studies have uncovered mixed associations between adult child migration status and physical functioning of older parents; none to our knowledge has examined the impact on unmet caregiving needs., Methods: Data come from a population-based study of Mexican adults ≥50 years. We used longitudinal targeted maximum likelihood estimation to estimate associations between having an adult child US migrant and lower-body functional limitations, and both needs and unmet needs for assistance with basic or instrumental activities of daily living (ADLs/IADLs) for 11,806 respondents surveyed over an 11-year period., Results: For women, having an adult child US migrant at baseline and 2-year follow-up was associated with fewer lower-body functional limitations [marginal risk difference (RD) = -0.14, 95% confidence interval (CI) = -0.26, -0.01] and ADLs/IADLs (RD = -0.08, 95% CI = -0.16, -0.001) at 2-year follow-up. Having an adult child US migrant at all waves was associated with a higher prevalence of functional limitations at 11-year follow-up (RD = 0.04, 95% CI = 0.01, 0.06). Having an adult child US migrant was associated with a higher prevalence of unmet needs for assistance at 2 (RD = 0.13, 95% CI = 0.04, 0.21) and 11-year follow-up for women (RD = 0.07, 95% CI = -0.02, 0.15) and 11-year follow-up for men (RD = 0.08, 95% CI = 0.00, 0.16)., Conclusion: Having an adult child US migrant had mixed associations with physical functioning, but substantial adverse associations with unmet caregiving needs for a cohort of older adults in Mexico.

Background: Dementia ascertainment is time-consuming and costly. Several algorithms use existing data from the US-representative Health and Retirement Study (HRS) to algorithmically identify dementia. However, relative performance of these algorithms remains unknown., Methods: We compared performance across five algorithms (Herzog-Wallace, Langa-Kabeto-Weir, Crimmins, Hurd, Wu) overall and within sociodemographic subgroups in participants in HRS and Wave A of the Aging, Demographics, and Memory Study (ADAMS, 2000-2002), an HRS substudy including in-person dementia ascertainment. We then compared algorithmic performance in an internal (time-split) validation dataset including participants of HRS and ADAMS Waves B, C, and/or D (2002-2009)., Results: In the unweighted training data, sensitivity ranged from 53% to 90%, specificity ranged from 79% to 97%, and overall accuracy ranged from 81% to 87%. Though sensitivity was lower in the unweighted validation data (range: 18%-62%), overall accuracy was similar (range: 79%-88%) due to higher specificities (range: 82%-98%). In analyses weighted to represent the age-eligible US population, accuracy ranged from 91% to 94% in the training data and 87% to 94% in the validation data. Using a 0.5 probability cutoff, Crimmins maximized sensitivity, Herzog-Wallace maximized specificity, and Wu and Hurd maximized accuracy. Accuracy was higher among younger, highly-educated, and non-Hispanic white participants versus their complements in both weighted and unweighted analyses., Conclusion: Algorithmic diagnoses provide a cost-effective way to conduct dementia research. However, naïve use of existing algorithms in disparities or risk factor research may induce nonconservative bias. Algorithms with more comparable performance across relevant subgroups are needed.

Background: Evidence suggests that aspects of the neighborhood environment may influence risk of problematic drug use among adolescents. Our objective was to examine mediating roles of aspects of the school and peer environments on the effect of receiving a Section 8 housing voucher and using it to move out of public housing on adolescent substance use outcomes., Methods: We used data from the Moving to Opportunity (MTO) experiment that randomized receipt of a Section 8 housing voucher. Hypothesized mediators included school climate, safety, peer drug use, and participation in an after-school sport or club. We applied a doubly robust, semiparametric estimator to longitudinal MTO data to estimate stochastic direct and indirect effects of randomization on cigarette use, marijuana use, and problematic drug use. Stochastic direct and indirect effects differ from natural direct and indirect effects in that they do not require assuming no posttreatment confounder of the mediator-outcome relationship. Such an assumption would be at odds with any causal model that reflects an intervention affecting a mediator and outcome through adherence to treatment assignment., Results: Having friends who use drugs and involvement in after-school sports or clubs partially mediated the effect of housing voucher receipt on adolescent substance use (e.g., stochastic indirect effect 0.45% [95% confidence interval: 0.12%, 0.79%] for having friends who use drugs and 0.04% [95% confidence interval: -0.02%, 0.10%] for involvement in after-school sports or clubs mediating the relationship between housing voucher receipt and marijuana use among boys). However, these mediating effects were small, contributing only fractions of a percent to the effect of voucher receipt on probability of substance use. No school environment variables were mediators., Conclusions: Measured school- and peer-environment variables played little role in mediating the effect of housing voucher receipt on subsequent adolescent substance use.

Background: In middle age, stroke incidence is higher among black than white Americans. For unknown reasons, this inequality decreases and reverses with age. We conducted simulations to evaluate whether selective survival could account for observed age patterning of black-white stroke inequalities., Methods: We simulated birth cohorts of 20,000 blacks and 20,000 whites with survival distributions based on US life tables for the 1919-1921 birth cohort. We generated stroke incidence rates for ages 45-94 years using Reasons for Geographic and Racial Disparities in Stroke (REGARDS) study rates for whites and setting the effect of black race on stroke to incidence rate difference (IRD) = 20/10,000 person-years at all ages, the inequality observed at younger ages in REGARDS. We compared observed age-specific stroke incidence across scenarios, varying effects of U, representing unobserved factors influencing mortality and stroke risk., Results: Despite a constant adverse effect of black race on stroke risk, the observed black-white inequality in stroke incidence attenuated at older age. When the hazard ratio for U on stroke was 1.5 for both blacks and whites, but U only directly influenced mortality for blacks (hazard ratio for U on mortality =1.5 for blacks; 1.0 for whites), stroke incidence rates in late life were lower among blacks (average observed IRD = -43/10,000 person-years at ages 85-94 years versus causal IRD = 20/10,000 person-years) and mirrored patterns observed in REGARDS., Conclusions: A relatively moderate unmeasured common cause of stroke and survival could fully account for observed age attenuation of racial inequalities in stroke.

Instrumental variables are routinely used to recover a consistent estimator of an exposure causal effect in the presence of unmeasured confounding. Instrumental variable approaches to account for nonignorable missing data also exist but are less familiar to epidemiologists. Like instrumental variables for exposure causal effects, instrumental variables for missing data rely on exclusion restriction and instrumental variable relevance assumptions. Yet these two conditions alone are insufficient for point identification. For estimation, researchers have invoked a third assumption, typically involving fairly restrictive parametric constraints. Inferences can be sensitive to these parametric assumptions, which are typically not empirically testable. The purpose of our article is to discuss another approach for leveraging a valid instrumental variable. Although the approach is insufficient for nonparametric identification, it can nonetheless provide informative inferences about the presence, direction, and magnitude of selection bias, without invoking a third untestable parametric assumption. An important contribution of this article is an Excel spreadsheet tool that can be used to obtain empirical evidence of selection bias and calculate bounds and corresponding Bayesian 95% credible intervals for a nonidentifiable population proportion. For illustrative purposes, we used the spreadsheet tool to analyze HIV prevalence data collected by the 2007 Zambia Demographic and Health Survey (DHS).

Background: The Moving To Opportunity (MTO) experiment manipulated neighborhood context by randomly assigning housing vouchers to volunteers living in public housing to use to move to lower poverty neighborhoods in five US cities. This random assignment overcomes confounding limitations that challenge other neighborhood studies. However, differences in MTO's effects across the five cities have been largely ignored. Such differences could be due to population composition (e.g., differences in the racial/ethnic distribution) or to context (e.g., differences in the economy)., Methods: Using a nonparametric omnibus test and a multiply robust, semiparametric estimator for transportability, we assessed the extent to which differences in individual-level compositional characteristics that may act as effect modifiers can account for differences in MTO's effects across sites. We examined MTO's effects on marijuana use, behavioral problems, major depressive disorder, and generalized anxiety disorder among black and Latino adolescent males, where housing voucher receipt was harmful for health in some sites but beneficial in others., Results: Comparing point estimates, differences in composition partially explained site differences in MTO effects on marijuana use and behavioral problems but did not explain site differences for major depressive disorder or generalized anxiety disorder., Conclusions: Our findings provide quantitative, rigorous evidence for the importance of context or unmeasured individual-level compositional variables in modifying MTO's effects.

Background: Inverse odds ratio weighting, a newly proposed tool to evaluate mediation in exposure-disease associations, may be valuable for a host of research questions, but little is known about its performance in real data. We compare this approach to a more conventional Baron and Kenny type of decomposition on an additive hazards scale to estimate total, direct, and indirect effects using the example of the role of literacy in mediating the effects of education on mortality., Methods: Health and Retirement Study participants born in the United States between 1900 and 1947 were interviewed biennially for up to 12 years (N = 17,054). Literacy was measured with a brief vocabulary assessment. Decomposition estimates were derived based on Aalen additive hazards models., Results: A 1 standard deviation difference in educational attainment (3 years) was associated with 6.7 fewer deaths per 1000 person-years (β = -6.7, 95% confidence interval [CI]: -7.9, -5.4). Of this decrease, 1.3 fewer deaths (β = -1.3, 95% CI: -4.0, 1.2) were attributed to the literacy pathway (natural indirect), representing 19% of the total effect. Baron and Kenny estimates were consistent with inverse odds ratio weighting estimates but were less variable (natural indirect effect: -1.2 [95% CI: -1.7, -0.69], representing 18% of total effect)., Conclusion: In a cohort of older Americans, literacy partially mediated the effect of education on mortality. See Video Abstract at http://links.lww.com/EDE/B78., Competing Interests: Conflicts of Interest: None declared.

Background: We describe bias resulting from individualized treatment selection, which occurs when treatment has heterogeneous effects and individuals selectively choose treatments of greatest benefit to themselves. This pernicious bias may confound estimates from observational studies and lead to important misinterpretation of intent-to-treat analyses of randomized trials. Despite the potentially serious threat to inferences, individualized treatment selection has rarely been formally described or assessed., Methods: The Moving To Opportunity trial randomly assigned subsidized rental vouchers to low-income families in high-poverty public housing. We assessed the Kessler-6 psychological distress and Behavior Problems Index outcomes for 2,829 adolescents 4-7 years after randomization. Among families randomly assigned to receive vouchers, we estimated probability of moving (treatment), predicted by prerandomization characteristics (c statistic = 0.63). We categorized families into tertiles of this estimated probability of moving, and compared instrumental variable effect estimates for moving on behavior problems index and Kessler-6 across tertiles., Results: Instrumental variable estimated effects of moving on behavioral problems index were most adverse for boys least likely to move (b = 0.93; 95% confidence interval: 0.33, 1.53) compared with boys most likely to move (b = 0.14; 95% confidence interval: -0.15, 0.44; P = 0.02 for treatment × tertile interaction). Effects on Kessler-6 were more beneficial for girls least likely to move compared with girls most likely to move (-0.62 vs. 0.02; interaction; P = 0.03)., Conclusions: Evidence of individualized treatment selection differed by child gender and outcome and should be evaluated in randomized trial reports, especially when heterogeneous treatment effects are likely and nonadherence is common., Competing Interests: The authors do not have any conflicts of interest to report.

Background: Health behaviors may contribute to socioeconomic inequalities in mortality, although the extent of such contribution remains unclear. We assessed the extent to which smoking, alcohol consumption, and physical inactivity have mediated the association between socioeconomic status (SES) and all-cause mortality in a representative sample of US adults., Methods: Initiated in 1992, the Health and Retirement Study is a longitudinal, biennial survey of a national sample of US adults born between 1931 and 1941. Our analyses are based on a sample of 8037 participants enrolled in 1992 and followed for all-cause mortality from 1998 through 2008. We used exploratory and confirmatory factor analysis to derive a measure of adult SES based on respondents' education, occupation, labor force status, household income, and household wealth. Potential mediators (smoking, alcohol consumption, and physical inactivity) were assessed biennially. We used inverse probability-weighted mediation models to account for time-varying covariates., Results: During the 10-year mortality follow-up, 859 (10%) participants died. After accounting for age, sex, and baseline confounders, being in the most-disadvantaged quartile of SES compared with the least disadvantaged was associated with a mortality risk ratio of 2.84 (95% confidence interval = 2.25-3.60). Together, smoking, alcohol consumption, and physical inactivity explained 68% (35-104%) of this association, leaving a risk ratio of 1.59 (1.03-2.45) for low SES., Conclusions: The distribution of health-damaging behaviors may explain a substantial proportion of excess mortality associated with low SES in the United States, suggesting the importance of social inequalities in unhealthy behaviors.

Background: Economic disadvantage is associated with depression and suicide. We sought to determine whether economic disadvantage reduces the effectiveness of depression treatments received in primary care., Methods: We conducted differential-effects analyses of the Prevention of Suicide in Primary Care Elderly: Collaborative Trial, a primary-care-based randomized, controlled trial for late-life depression and suicidal ideation conducted between 1999 and 2001, which included 514 patients with major depression or clinically significant minor depression., Results: The intervention effect, defined as change in depressive symptoms from baseline, was stronger among persons reporting financial strain at baseline (differential effect size = -4.5 Hamilton Depression Rating Scale points across the study period [95% confidence interval = -8.6 to -0.3]). We found similar evidence for effect modification by neighborhood poverty, although the intervention effect weakened after the initial 4 months of the trial for participants residing in poor neighborhoods. There was no evidence of substantial differences in the effectiveness of the intervention on suicidal ideation and depression remission by economic disadvantage., Conclusions: Economic conditions moderated the effectiveness of primary-care-based treatment for late-life depression. Financially strained individuals benefited more from the intervention; we speculate this was because of the enhanced treatment management protocol, which led to a greater improvement in the care received by these persons. People living in poor neighborhoods experienced only temporary benefit from the intervention. Thus, multiple aspects of economic disadvantage affect depression treatment outcomes; additional work is needed to understand the underlying mechanisms.

Background: Concerns have been raised that education may have greater benefits for persons at high risk of coronary heart disease (CHD) than for those at low risk., Methods: We estimated the association of education (less than high school, high school, or college graduates) with 10-year CHD risk and body mass index (BMI), using linear and quantile regression models, in the following two nationally representative datasets: the 2006 wave of the Health and Retirement Survey and the 2003-2008 National Health and Nutrition Examination Survey (NHANES)., Results: Higher educational attainment was associated with lower 10-year CHD risk for all groups. However, the magnitude of this association varied considerably across quantiles for some subgroups. For example, among women in NHANES, a high school degree was associated with 4% (95% confidence interval = -9% to 1%) and 17% (-24% to -8%) lower CHD risk in the 10th and 90th percentiles, respectively. For BMI, a college degree was associated with uniform decreases across the distribution for women, but with varying increases for men. Compared with those who had not completed high school, male college graduates in the NHANES sample had a BMI that was 6% greater (2% to 11%) at the 10th percentile of the BMI distribution and 7% lower (-10% to -3%) at the 90th percentile (ie, overweight/obese). Estimates from the Health and Retirement Survey sample and the marginal quantile regression models showed similar patterns., Conclusion: Conventional regression methods may mask important variations in the associations between education and CHD risk.

Background: Early-life socioeconomic status (SES) is associated with adult chronic disease, but it is unclear whether this effect is mediated entirely via adult SES or whether there is a direct effect of adverse early-life SES on adult disease. Major challenges in evaluating these alternatives include imprecise measurement of early-life SES and bias in conventional regression methods to assess mediation. In particular, conventional regression approaches to direct effect estimation are biased when there is time-varying confounding of the association between adult SES and chronic disease by chronic disease risk factors., Methods: First-reported heart disease, diabetes, and stroke diagnoses were assessed in a national sample of 9760 Health and Retirement Study participants followed biennially from 1992 through 2006. Early-life and adult SES measures were derived using exploratory and confirmatory factor analysis. Early-life SES was measured by parental education, father's occupation, region of birth, and childhood rural residence. Adult SES was measured by respondent's education, occupation, labor force status, household income, and household wealth. Using marginal structural models, we estimated the direct effect of early-life SES on chronic disease onset that was not mediated by adult SES. Marginal structural models were estimated with stabilized inverse probability-weighted log-linear models to adjust for risk factors that may have confounded associations between adult SES and chronic disease., Results: During follow-up, 24%, 18%, and 9% of participants experienced first onset of heart disease, diabetes, and stroke, respectively. Comparing those in the most disadvantaged with the least disadvantaged quartile, early-life SES was associated with coronary heart disease (risk ratio = 1.30 [95% confidence interval = 1.12-1.51]) and diabetes (1.23 [1.02-1.48]) and marginally associated with stroke via pathways not mediated by adult SES., Conclusions: Our results suggest that early-life socioeconomic experiences directly influence adult chronic disease outcomes.

Background: Selective attrition may introduce bias into analyses of the determinants of cognitive decline. This is a concern especially for risk factors, such as smoking, that strongly influence mortality and dropout. Using inverse-probability-of-attrition weights, we examined the influence of selective attrition on the estimated association of current smoking (vs. never smoking) with cognitive decline., Methods: Chicago Health and Aging Project participants (n = 3713), aged 65-109 years, who were current smokers or never- smokers, underwent cognitive assessments up to 5 times at 3-year interval. We used pooled logistic regression to fit predictive models of attrition due to death or study dropout across the follow-up waves. With these models, we computed inverse-probability-of-attrition weights for each observation. We fit unweighted and weighted, multivariable-adjusted generalized-estimating-equation models, contrasting rates of change in cognitive scores in current versus never-smokers. Estimates are expressed as rates of change in z score per decade., Results: During the 12 years of follow-up, smokers had higher mortality than never-smokers (hazard ratio = 1.93 [95% confidence interval = 1.67 to 2.23]). Higher previous cognitive score was associated with increased likelihood of survival and continued participation. In unweighted analyses, current smokers' cognitive scores declined 0.11 standard units per decade more rapidly than never-smokers' (95% CI = -0.20 to -0.02). Weighting to account for attrition yielded estimates that were 56% to 86% larger, with smokers' estimated 10-year rate of decline up to 0.20 units faster than never-smokers' (95% CI = -0.36 to -0.04)., Conclusions: Estimates of smoking's effects on cognitive decline may be underestimated due to differential attrition. Analyses that weight for the inverse probability of attrition help compensate for this attrition.