Propensity Scores in Health Disparities Research: The Example of Cognitive Aging and the Hispanic Paradox.

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.

History of Low Hourly Wage and All-Cause Mortality Among Middle-aged Workers.

Importance: Earning a low wage is an increasingly recognized public health concern, yet little research exists on the long-term health consequences of sustained low-wage earning. Objective: To examine the association of sustained low-wage earning and mortality in a sample of workers with hourly wage reported biennially during peak midlife earning years. Design, Setting, and Participants: This longitudinal study included 4002 US participants, aged 50 years or older, from 2 subcohorts of the Health and Retirement Study (1992-2018) who worked for pay and reported earning hourly wages at 3 or more time points during a 12-year period during their midlife (1992-2004 or 1998-2010). Outcome follow-up occurred from the end of the respective exposure periods until 2018. Exposures: Low-wage-less than the hourly wage for full-time, full-year work at the federal poverty line-earning history was categorized as never earning a low wage, intermittently earning a low wage, and sustained earning a low wage. Main Outcomes and Measures: Cox proportional hazards and additive hazards regression models sequentially adjusted for sociodemographics, and economic and health covariates were used to estimate associations between low-wage history and all-cause mortality. We examined interaction with sex or employment stability on multiplicative and additive scales. Results: Of the 4002 workers (aged 50-57 years at the beginning of exposure period and 61-69 years at the end), 1854 (46.3%) were female; 718 (17.9%) experienced employment instability; 366 (9.1%) had a history of sustained low-wage earning; 1288 (32.2%) had intermittent low-wage earning periods; and 2348 (58.7%) had never earned a low wage. In unadjusted analyses, those who had never earned low wages experienced 199 deaths per 10 000 person-years, those with intermittent low wages, 208 deaths per 10 000 person-years, and those with sustained low wages, 275 deaths per 10 000 person-years. In models adjusted for key sociodemographic variables, sustained low-wage earning was associated with mortality (hazard ratio [HR], 1.35; 95% CI, 1.07-1.71) and excess deaths (66; 95% CI, 6.6-125); these findings were attenuated with additional adjustments for economic and health covariates. Significant excess death and elevated mortality risk were observed for workers with sustained low-wage exposure and employment fluctuations (eg, for sustained low-wage × employment fluctuated, HR, 2.18; 95% CI, 1.35-3.53; for sustained low-wage × stable employment, HR, 1.17; 95% CI, 0.89,-1.54; P for interaction = .003). Conclusions and Relevance: Sustained low-wage earning may be associated with elevated mortality risk and excess deaths, especially when experienced alongside unstable employment. If causal, our findings suggest that social and economic policies that improve the financial standing of low-wage workers (eg, minimum wage laws) could improve mortality outcomes.

Association of Low Hourly Wages in Middle Age With Faster Memory Decline in Older Age: Evidence From the Health and Retirement Study.

Little research has investigated the long-term relationship between low wages and memory decline, despite the growing share of low-wage workers in the US labor market. Here, we examined whether cumulative exposure to low wages over 12 years in midlife is associated with memory decline in later life. Using 1992-2016 data from the Health and Retirement Study, we analyzed data from 2,879 individuals born in 1936-1941 using confounder-adjusted linear mixed-effects models. Low-wage work was defined as an hourly wage lower than two-thirds of the federal median wage for the corresponding year and was categorized into "never," "intermittent," and "sustained" based on wages earned from 1992 to 2004. Memory function was measured at each study visit from 2004 to 2016 via a memory composite score. The confounder-adjusted annual rate of memory decline among "never" low-wage earners was -0.12 standard units (95% confidence interval: -0.13, -0.10). Compared with this, memory decline among workers with sustained earning of low midlife wages was significantly faster (ßtime×sustained = -0.014, 95% confidence interval: -0.02, -0.01), corresponding to an annual rate of -0.13 standard units for this group. Sustained low-wage earning in midlife was significantly associated with a downward trajectory of memory performance in older age. Enhancing social policies to protect low-wage workers may be especially beneficial for their cognitive health.

Comparing different operationalizations of allostatic load measured in mid-life and their patterning by race and cumulative life course socioeconomic status.

Since its conceptualization, there has been a lack of consensus on the best way to operationalize allostatic load (AL). As a marker of the cumulative, physiological wear and tear on the body resulting from chronic exposure to stressors, it follows that AL should be higher among people who have faced more stressful life experiences. Thus, the purpose of this study was to construct AL scores using different operationalizations and, as a measure of construct validity, compare whether each construction produced expected disparities in AL by race and a composite socioeconomic status (SES) variable which accounts for measures over the life course; we also explored differences by sex. We conducted the study in a sample of 45-52-year-old offspring from the Child Health and Development Studies, a longitudinal birth cohort established in the early 1960s. AL scores were constructed in 6 different ways and included 10 biomarkers from inflammatory, neuroendocrine, cardiovascular, and metabolic systems. Our main approach to constructing AL was to sum across high-risk biomarker quartiles, correct for medication use, and use sex-specific high-risk quartiles for specific biomarkers. Alternative constructions did not use sex-specific quartiles and/or weighted biomarkers within subsystems and/or did not correct for medication use. We estimated differences in AL scores by race, SES, sex and their pairwise interactions. All constructions of AL, including the main approach, produced expected disparities by race (higher scores for Black vs. non-Black participants) and life course SES (higher scores for low vs. high SES participants). However, disparities by sex only emerged when the AL score was constructed via approaches that did not use sex-specific high-risk quartiles; for these alternative constructions, overall, female participants had higher AL scores than male participants and Black female participants had the highest AL scores in the sample. For most constructions, the pairwise interaction between sex and SES, showed a stronger disparity in AL scores between low and high-SES female compared with low- and high-SES male participants; this suggests that, in terms of lowering AL, high life course SES may be more important for female than male participants. In conclusion, our results suggest that the basic AL concept is consistently expressed in different operationalizations, making it an especially useful and robust tool for understanding disparities by race and SES.

Could maternal thyroid function during pregnancy affect daughters' age at menarche through child growth? A mediation analysis.

Early menarche is associated with adverse health outcomes during adolescence as well as breast and other reproductive cancers later in adulthood. However, the causes of early menarche and the pathways through which they operate are not fully understood. Though maternal thyroid function during pregnancy affects child growth, and rapid childhood growth is associated with a decreased age at menarche, the relationship between prenatal maternal thyroid function and daughters' age at menarche has not been examined. We conducted a mediation analysis in a historical cohort of 260 mother-child pairs to estimate the total and indirect effects of maternal prenatal thyroid function on daughters' age at menarche. No association was observed between thyroid stimulating hormone (TSH) or anti-thyroid peroxidase antibodies (ATPO) and daughters' age at menarche. Using a sample-specific, a-priori cutoff at the 10th percentile, low levels of maternal free thyroxine (FT4) were associated with earlier daughter age at menarche, with a hazard ratio (95 % CI) of 1.70 (1.02, 2.84) comparing the bottom 10th percentile with the top 90th percentile of exposure levels. Higher maternal FT4 was associated with rapid child weight gain from ages 5-9, and rapid child weight gain from ages 5-9 was associated with earlier age at menarche; the estimated indirect effect of this pathway was null. While maternal FT4 is associated with earlier age at menarche in daughters, this is not mediated by rapid weight gain in our study population, suggesting that maternal FT4 is operating through a different pathway.

Is the Way Forward to Step Back? Documenting the Frequency With Which Study Goals Are Misaligned With Study Methods and Interpretations in the Epidemiologic Literature.

In any research study, there is an underlying process that should begin with a clear articulation of the study's goal. The study's goal drives this process; it determines many study features, including the estimand of interest, the analytic approaches that can be used to estimate it, and which coefficients, if any, should be interpreted. Misalignment can occur in this process when analytic approaches and/or interpretations do not match the study's goal; misalignment is potentially more likely to arise when study goals are ambiguously framed. In this study, misalignment in the observational epidemiologic literature was documented and how the framing of study goals contributes to misalignment was explored. The following 2 misalignments were examined: use of an inappropriate variable selection approach for the goal (a "goal-methods" misalignment) and interpretation of coefficients of variables for which causal considerations were not made (e.g., Table 2 Fallacy, a "goal-interpretation" misalignment). A random sample of 100 articles published 2014-2018 in the top 5 general epidemiology journals were reviewed. Most reviewed studies were causal, with either explicitly stated (n = 13; 13%) or associational-framed (n = 71; 69%) aims. Full alignment of goal-methods-interpretations was infrequent (n = 9; 9%), although clearly causal studies (n = 5 of 13; 38%) were more often fully aligned than were seemingly causal ones (n = 3 of 71; 4%). Goal-methods misalignments were common (n = 34 of 103; 33%), but most frequently, methods were insufficiently reported to draw conclusions (n = 47; 46%). Goal-interpretations misalignments occurred in 31% (n = 32) of the studies and occurred less often when the methods were aligned (n = 2; 2%) compared with when the methods were misaligned (n = 13; 13%).

Explaining the Black-white depression paradox: Interrogating the Environmental Affordances Model.

The Environmental Affordances (EA) model posits that Black Americans' engagement with unhealthy behaviors (i.e. smoking, alcohol use, eating calorie-dense foods) to cope with stressor exposure may simultaneously account for their observed greater risk of chronic physical illness, and their observed equal or lesser prevalence of depression, relative to white Americans - the so-called "Black-white depression paradox." However, the specific mechanisms through which such effects might arise have been theorized and analyzed inconsistently across studies, raising concerns regarding the appropriateness of existing empirical tests of the model as well as the validity of the conclusions. We specify the two mechanisms most consistent with the EA model - 'Mediation-only' and 'Mediation and Modification' - and derive a priori predictions based on each. We systematically test these pathways using a subset of 559 participants of the Child Health and Development Study who were included in an adult follow-up study between 2010 and 2012 and self-identified as Black or white. Results failed to support either of the two mechanisms derived from the EA model, challenging the validity and utility of the model for explaining racial differences in depression; efforts to develop alternative hypotheses to explain the paradox are needed.

Tracking of Obesity in Childhood into Adulthood: Effects on Body Mass Index and Fat Mass Index at Age 50.

Background: Obesity is present in 17% of US youth, age 2-19 years, but the extent to which obesity in childhood is associated with higher BMI and fat mass in middle age is unclear. In this study, links between childhood body size and BMI and body composition at age ∼50 were assessed. Methods: Child Health and Development Studies participants, born between 1960 and 1963 in Alameda County, and still living in California, from whom anthropometric data were collected at age 5, 9-11, and/or 15-17 years were followed-up at age ∼50 for anthropometric outcomes (251 women; 249 men). Linear regression analyses assessed whether overweight (85th to <95th BMI percentile) or obesity (≥95th BMI percentile) at age 5 were associated with BMI, fat mass index (FMI), and lean mass index (LMI) at age ∼50. Results: At age 50, participants with obesity at age 5 had BMI scores that were 6.51 units higher [95% confidence interval (CI) = 3.67-9.35] than participants who were normal weight at age 5; FMI and LMI scores were 4.15 (95% CI = 1.98-6.32) and 2.36 (95% CI = 1.45-3.26) units higher, respectively. However, obesity experienced at age 5 had only a modest positive predictive value for predicting the presence of obesity at age 50 (67%), whereas obesity present at age 15-17 had a higher positive predictive value (86%). Conclusions: The experience of obesity at age 5 for members of this birth cohort was associated with significantly higher BMI, FMI, and LMI at age ∼50.

Who is in this study, anyway? Guidelines for a useful Table 1.

OBJECTIVE: Epidemiologic and clinical research papers often describe the study sample in the first table. If well-executed, this "Table 1" can illuminate potential threats to internal and external validity. However, little guidance exists on best practices for designing a Table 1, especially for complex study designs and analyses. We aimed to summarize and extend the literature related to reporting descriptive statistics. STUDY DESIGN AND SETTING: In consultation with existing guidelines, we synthesized and developed reporting recommendations driven by study design and focused on transparency related to potential threats to internal and external validity. RESULTS: We describe a basic structure for Table 1 and discuss simple modifications in terms of columns, rows, and cells to enhance a reader's ability to judge both internal and external validity. We further highlight several analytic complexities common in epidemiologic research (missing data, sample weights, clustered data, and interaction) and describe possible variations to Table 1 to maintain and add clarity about study validity in light of these issues. We discuss considerations and tradeoffs in Table 1 related to breadth and comprehensiveness vs. parsimony and reader-friendliness. CONCLUSION: We anticipate that our work will guide authors considering layouts for Table 1, with attention to the reader's perspective.

Body mass index across the life course: emergence of race-by-sex disparities in early childhood.

PURPOSE: The purpose of this study was to assess when in the life-course race-by-sex disparities in body mass index (BMI) emerge. METHODS: Child Health and Development Studies participants, from whom height and weight data were collected at ages 5, 9-11, and 15-17 years, were followed up at the age of 50 years for anthropometric outcomes. Follow-up was completed for 605 subjects, 460 of whom were assessed for height and weight at the age of 50 years, had at least one available childhood BMI measure, and self-identified as either non-Hispanic black or non-Hispanic white. Linear regression analyses were conducted to determine whether interactions existed between race (black vs. white) and sex for predicting BMI at ages 5, 9-11, 15-17, and 50 years. RESULTS: At age 5 years, BMI was independent of sex for both blacks and whites, but by the age of 9-11 years, BMI was sex-dependent in blacks, with higher BMI observed among black females. This sex dependence for BMI among blacks persisted at ages 15-17 years and age 50 years. The race-by-sex interaction was significant at ages 9-11, 15-17, and 50 years (P for interaction = 0.001, 0.002, and 0.01, respectively). CONCLUSIONS: Race-by-sex disparities in body size were observed by the age of 9-11 years and persisted until the age of 50 years.

Disparities in self-rated health across generations and through the life course.

Extensive evidence leads us to expect that health disparities by race and socioeconomic status found in one generation might be reproduced in the next. To the extent that this occurs it is important to assess life course processes responsible for the reproduction. Prospective evidence concerning such life course processes is hard to come by as it requires long-term follow-up of individuals from childhood through adult life. We present data from the Child Health and Development Disparities study that provides evidence relevant to this issue with respect to self-rated health. Mothers and offspring recruited in California's Bay Area between 1959 and 1967 were assessed during pregnancy with follow-up exams of offspring along with in-person interviews with mothers (at offspring ages 5, 9-11, 15-17) and offspring (at ages 15-17, ∼50). Available data allow us to assess the importance of three potential life course pathways in the reproduction of inequalities in self-rated health - socioeconomic pathways, cognitive pathways and pathways involving emerging health itself. As expected we found that race and SES disparities in SRH are reproduced across generations. They are evident in mothers, not strong or significant in offspring at 15-17, but present once again in offspring at age ∼50. Concerning potential pathways, we found that indicators of child health were related to adult SRH and played some role in accounting for race but not SES disparities in adult SRH. Cognitive abilities were unrelated to adult SRH with childhood SES controlled. Childhood SES was associated with adult SRH independent of other childhood factors and is reduced to non-significance only when offspring college attainment is controlled. Race and SES disparities in self-reported health in one generation are re-expressed in the next with strongest support for SES pathways in this transmission.

Dichlorodiphenyltrichloroethane (DDT), DDT metabolites and pregnancy outcomes.

Organochlorine pesticides (OCPs) are persistent endocrine disruptors. OCPs cross the placenta; this prenatal exposure has been associated with adverse pregnancy outcomes. We investigated associations between prenatal exposure to OCPs and gestational age and birth weight in 600 infants born between 1960 and 1963. The primary OCP was 1,1,1-trichloro-2,2-bis(p-chlorophenyl)ethane (p,p'-DDT), its primary metabolite, 1,1'-dichloro-2,2'-bis(p-chlorophenyl)ethylene (p,p'-DDE) and the contaminant, 1,1,1-trichloro-2-(p-chlorophenyl)-2-(o-chlorophenyl)-ethane (o,p'-DDT). Regression analysis indicated that for each natural log unit increase in p,p'-DDT, birth weight increased by 274 g (95% CI: 122, 425) when controlling for p,p'-DDE and o,p'-DDT. At a given level of p,p'-DDT exposure, o,p'-DDT and p,p'-DDE were associated with decreased birth weight. p,p'-DDE was negatively associated with length of gestation, controlling for p,p'-DDT and o,p'-DDT. These findings suggest opposing associations between exposure to p,p'-DDT and p,p'-DDE and birth weight. We did not find evidence to support mediation by maternal thyroid hormone status nor that the association differed by sex.

Prenatal polychlorinated biphenyl exposure is associated with decreased gestational length but not birth weight: archived samples from the Child Health and Development Studies pregnancy cohort.

BACKGROUND: Polychlorinated biphenyls (PCBs), known endocrine disruptors, were banned in 1979 but persist in the environment. Previous studies are inconsistent regarding prenatal exposure to PCBs and pregnancy outcomes. We investigated associations between prenatal exposure to PCBs and gestational length and birth weight. METHODS: In a sample of 600 infants (born between 1960 and 1963) randomly selected from Child Health and Development Studies participants followed through adolescence we measured 11 PCB congeners in maternal post partum sera (within three days of delivery). Length of gestation was computed from the reported first day of the last menstrual period (LMP) and delivery date. Linear regression was used to estimate associations between PCB exposure and gestational age and birth weight, adjusting for potential confounders. PCBs were grouped according to hypothesized biological action (1b (sum of weak phenobarbital inducers), 2b (sum of limited dioxin activity), and 3 (sum of CYP1A and CYP2b inducers)) or degree of ortho- substitution (mono, di, tri). Secondary analyses examined associations between total PCB exposure and exposure to individual congeners. RESULTS: Each unit increase in mono-ortho substituted PCBs was associated with a 0.30 week decrease (95% confidence interval (CI) -0.59, -0.016), corresponding to a 2.1 (95% CI -4.13, -0.11) day decrease in length of gestation. Similar associations were estimated for di-ortho substituted PCBs, (1.4 day decrease; (95% CI -2.9, 0.1)) and group 3 PCBs (0.84 day decrease; (95% CI -1.8, 0.11). We found similar associations in congener specific analyses and for the sum of congeners. CONCLUSIONS: Our study provides new evidence that PCB exposure shortens length of gestation in humans. This may have public health implications for population exposures.