Search

Objectives Measuring cognition in an aging populabtion is a public health priority. A move towards survey measurement via the web (as opposed to phone or in-person) is cost-effective but challenging as it may induce bias in cognitive measures. We examine this possibility using an experiment embedded in the 2018 wave of data collection for the U.S. Health and Retirement Study (HRS). Methods We utilize techniques from multiple group item response theory to assess the effect of survey mode on performance on the HRS cognitive measure. We also study the problem of attrition by attempting to predict dropout and via approaches meant to minimize bias in subsequent inferences due to attrition. Results We find evidence of an increase in scores for HRS respondents who are randomly assigned to the web-based mode of data collection in 2018. Web-based respondents score higher in 2018 than experimentally matched phone-based respondents, and they show much larger gains relative to 2016 performance and subsequently larger declines in 2020. The differential in favor of web-based responding is observed across all items, but is most pronounced for the Serial 7 task and numeracy items. Due to the relative ease of the web-based mode, we suggest a cutscore of 12 being used to indicate CIND (cognitively impaired but not demented) status when using the web-based version rather than 11. Discussion The difference in mode may be nonignorable for many uses of the HRS cognitive measure. In particular, it may require reconsideration of some cutscore-based approaches to identify impairment.

The implications of selection bias due to volunteering (volunteer bias) for genetic association studies are poorly understood. Because of its large sample size and extensive phenotyping, the UK Biobank (UKB) is included in almost all large genomewide association studies (GWAS) to date, as it is one of the largest cohorts. Yet, it is known to be highly selected. We develop inverse probability weighted GWAS (WGWAS) to estimate GWAS summary statistics in the UKB that are corrected for volunteer bias. WGWAS decreases the effective sample size substantially compared to GWAS by an average of 61% (from 337,543 to 130,684) depending on the phenotype. The extent to which volunteer bias affects GWAS associations and downstream results is phenotype-specific. Through WGWAS we find 11 novel genomewide significant loci for type 1 diabetes and 3 for breast cancer. These loci were not identified previously in any prior GWAS. Further, genetic variant’s effect sizes and heritability estimates become more predictive in WGWAS for certain phenotypes (e.g., educational attainment, drinks per week, breast cancer and type 1 diabetes). WGWAS also alters biological annotation relations in gene-set analyses. This suggests that not accounting for volunteer-based selection can result in GWASs that suffer from bias, which in turn may drive spurious associations. GWAS consortia may therefore wish to provide population weights for their data sets or rely more on population-representative samples.

Measuring cognition in an aging population is a public health priority. A move towards survey measurement via the web (as opposed to phone or in-person) is cost effective but challenging as it may induce bias in cognitive measures. We examine this possibility using an experiment embedded in the 2018 wave of data collection for the US Health and Retirement Study (HRS). We find evidence of an increase in scores for HRS respondents who are randomly assigned to the web-based mode of data collection in 2018. Web-based respondents score higher in 2018 than do phone-based respondents, and they show much larger gains relative to 2016 performance and subsequently larger declines in 2020. The bias in favor of web-based responding is observed across all item types, but most pronounced for the serial 7 and items on financial literacy. These results imply a larger threshold for web-based respondents in applications requiring an indicator of cognitive impairment. Implications for both use of HRS data and future survey work on cognition are discussed.

As the population of the US ages, there is interest in assessing health conditions associated with age and longevity, such as age-related decline in cognitive functioning. As a result, there is an increased focus on measuring cognitive functioning in surveys of older populations. One challenge relates to conducting comparable measurement across survey modes (e.g., phone vs. web). Compounding this is that mode of survey administration is often not assigned randomly making inter-group comparison more difficult. This paper addresses these issues using a novel experiment embedded within the Health and Retirement Study (HRS). The HRS, a US-based cohort of people over 50, has measured cognition since its inception using both in-person and telephone modes. In 2018, a sample of approximately 3700 respondents was identified as web-eligible based on a prior report of internet access along with other selection criteria. Of these, 60% were randomly selected for the web sample with the remainder serving as controls, assigned to telephone mode for comparison purposes. We deploy techniques from item response theory (IRT) and differential item functioning (DIF) to estimate the difference in cognitive functioning between web and phone respondents in 2018 based on longitudinal cognition data collected prior to 2018. Second, we estimate the overall effect of taking the survey via the web as compared to the phone. Third, we examine item-level variation in the magnitude of the mode effect and suggest possible methods for adjustment to support longitudinal consistency. These results are important in guiding future research that utilizes web-based cognitive measures.

Cellular changes in the adaptive immune system accompanies the aging process and contributes to an age-related immune phenotype (ARIP) characterized by decrease in naïve T (TN) cells and increase in memory T (TM) cells. However, a population level marker of ARIP associated with biological aging and age-related chronic conditions has not been evaluated previously. We developed two ARIP measures based on well understood age related changes in T cell distribution: TN/ (TCM (Central Memory) + TEM (Effector Memory) + TEFF (Effector)) or TN/ TM in CD4+ and CD8+ T cells. We compared them with more commonly used ARIP measures such as CD4/CD8 ratio and CD8+ TN cells by evaluating associations with chronological age and phenotypic age using linear regression and association with multimorbidity using multinomial logistic regression. CD8+ TN and TN/TM had the strongest inverse association with chronological age (beta estimates: -3.41, -3.61; p-value< 0.0001). CD4+ TN/TM had the strongest inverse association with phenotypic age (beta estimate = -0.74; p-value < 0.0001) after adjustment for age, sex, race and CMV serostatus. CD4+ TN/TM was inversely associated with co-occurring chronic conditions (odds ratio for 2 conditions and 3 conditions vs. 0 conditions: 0.74 (95%CI: 0.63-0.86) and 0.75 (95% CI: 0.63-0.90), respectively) after adjustment for age, sex, race, CMV serostatus, smoking, and BMI. CD4+ TN/TM had a stronger association with phenotypic age and age-related morbidity compared to other ARIP measures. Future longitudinal studies can help us evaluate if CD4+ TN/TM can predict risk of aging related outcomes.

Immune cells distribution is shaped by numerous factors including environmental factors, age, and genetics. Cohabitation has been associated with similar microbiomes, possibly due to dietary patterns and exposure to similar pathogens but has not been studied in the context of adaptive and innate immune systems previously. We used immunophenotyping data of 2283 households with participants living in the same household and compared it to 2283 randomly generated pairs of participants from the Health and Retirement study. The adaptive immune cells (subsets of T-cells and B-cells), and innate immune cells (monocytes, natural killer cells, and neutrophils) were compressed to two coordinates using multidimensional scaling. The Euclidean distances between participants in the same household were compared to the distances between the random pairs of participants using two sample independent t-tests. The mean distances of the immune coordinate points for adaptive immune cells between participants in the same household were lower than the randomly paired participants (p-value < 0.0001) and the variability of intra-household distances was lower than the random pairs (IQR: 7.18 vs 8.99). For the innate immune cells, the mean distances between participants in the same household were slightly lower than the randomly paired participants (p-value = 0.03) but the variability of the intra-household distances was higher than the random pairs (IQR: 4.08 vs 3.65). Variability in the adaptive immune system among participants living in the same household were substantially lower indicating the influence of shared environmental conditions in determining the adaptive immune profiles.

Understanding the ``fit'' of models designed to predict binary outcomes has been a long-standing problem. We propose a flexible, portable, and intuitive metric for quantifying the change in accuracy between two predictive systems in the case of a binary outcome: the InterModel Vigorish (IMV). The IMV is based on an analogy to weighted coins, well-characterized physical systems with tractable probabilities. The IMV is always a statement about the change in fit relative to some baseline model---which can be as simple as the prevalence---whereas other metrics are stand-alone measures that need to be further manipulated to yield indices related to differences in fit across models. Moreover, the IMV is consistently interpretable independent of baseline prevalence. We contrast this metric with alternatives in numerous simulations. The IMV is more sensitive to estimation error than many alternatives and also shows distinctive sensitivity to prevalence. We then showcase its flexibility across examples spanning the social, biomedical, and physical sciences. We also demonstrate how it can be used to provide straightforward interpretation of logistic regression coefficients. The IMV allows for precise answers to questions about changes in model fit in a variety of settings in a manner that will be useful for furthering research with binary outcomes.

DNA methylation (DNAm) patterns related to age and aging phenotypes (i.e., epigenetic clocks) are of growing interest as indicators of biological age and risk of negative health outcomes. We investigated associations between the components of GrimAge, an epigenetic clock estimated from DNAm patterns for seven blood protein levels and smoking pack years, and 2-year mortality in the Health and Retirement Study (HRS) to determine if any of the DNAm subcomponents were driving observed associations. A representative subsample of individuals who participated in the HRS 2016 Venus Blood Study were included in this analysis (N=3430). DNAm was measured with the Infinium Methylation EPIC BeadChip. Deaths that occurred between 2016 and 2018 contributed to 2-year mortality estimates (N=159, 4.5% of the sample). Weighted logistic regression estimated the association first between GrimAge and 2-year mortality and second between the DNAm subcomponents and 2-year mortality. All models were adjusted for age, sex, race/ethnicity, education, current smoking status, smoking pack years and cell composition of the biological sample. The average GrimAge for participants with and without 2-year mortality was 77 years 68 years respectively. A one-year increase in GrimAge was associated with 17% higher odds of 2-year mortality (95% CI: 1.16, 1.17). Two of the seven DNAm blood protein subcomponents of GrimAge (TIMP metallopeptidase inhibitor 1, adrenomedullin) and DNAm smoking pack years were associated with 2-year mortality and DNAm smoking pack years appeared to drive the overall GrimAge association with 2-year mortality. GrimAge was a better predictor of 2-year mortality than the DNAm subcomponents individually.

DNA methylation (DNAm) is an increasingly popular biomarker of health and aging outcomes. Smoking behaviors have a significant and well documented correlation with methylation signatures within the epigenome and are important confounding variables to account for in epigenome-wide association studies (EWAS). However, the common classification of individuals as ‘current’, ‘former’, and ‘never’ smokers may miss crucial DNAm patterns associated with other smoking behaviors such as duration, intensity, and frequency of cigarette smoking, resulting in an underestimation of the contribution of smoking behaviors to DNAm and potentially biasing EWAS results. We investigated associations between multiple smoking behavioral phenotypes (smoking pack years, smoking duration, smoking start age, and smoking end age) and single site DNAm using linear regressions adjusting for age, sex, race/ethnicity, education, and cell-type proportions in a subsample of individuals who participated in the HRS 2016 Venous Blood Study (N=1,775). DNAm was measured using the Infinium Methylation EPIC BeadChip. All 4 phenotypes had significant associations (FDR < 0.05) with many methylation sites (packyears=6859, smoking duration=6572, start age=11374, quit age=773). There was not much overlap in DNAm sites between the full set of models with only 6 overlapping between all 4. However, the phenotypes packyears and smoking duration showed large overlap (N=3782). Results suggest additional smoking phenotypes beyond current/former/never smoker classification should be included in EWAS analyses to appropriately account for the influence of smoking behaviors on DNAm.

Introduction: Obesity/body mass index (BMI) is one of the leading risk factors of cardiovascular disease (CVD) in both African American and non-Hispanic white populations. Identifying the range of factors that lead to the persistent and pronounced differences in the BMI of populations is critical for developing effective strategies for reducing these inequalities and ultimately reducing the incidence of CVD. It is posited that these differences arise from the interplay between both biological (genes) and social factors (environments). Standard approaches to this complex problem of incorporating “bio-social” data remain fractured; examining only biological, or only social determinants of the diseases/risk factors in question. This paradigm is not effective in uncovering the origins of diseases that arise from the relationships between multiple systems. By incorporating both biological and sociological data, we seek to better understand the health disparities underlying obesity. Hypotheses: We hypothesize that there will be gene regions associated with BMI in both African and non-Hispanic whites, but not necessarily the same regions across ethnicity. We expect stronger associations in each ethnicity with gene regions defined from meta-analysis of that ethnicity. We hypothesize that low socioeconomic position, lower levels of social support and higher levels of anger, ongoing stressful life events, and depressive symptoms will modify the effects of genetic predictors on BMI. Methods: We used novel, gene region-based analysis methods (Sequence Kernel Association Testing and Gene-Environment Set Association Test) with gene regions identified from 103 known obesity loci (97 from published GWAS meta-analysis in European ancestry and 6 from African ancestry), to examine the effect of gene regions on BMI in up to 3,000 African Americans and up to 12,000 non-Hispanic whites from the Health and Retirement Study. We also investigated the modification of these gene regions’ effects on BMI by social and psychosocial environment and compared across ethnicity. Results and Conclusions: Preliminary results indicate strong gene-region associations in African Americans with several genes including GPRC5B (p=0.024), SH2B1 (p=0.016) and TNNI3K (p=0.004) as well as statistically significant associations between BMI and social/psychosocial variables (bivariate p

This project evaluated if by focusing on process changes and tool use rather than key indicator rates, the use of evidence-based therapies in patients with acute myocardial infarction (AMI) would increase.The use of tools designed to improve quality of care in the American College of Cardiology AMI Guidelines Applied in Practice Pilot Project resulted in improved adherence to evidence-based therapies for patients, but overall, tool use was modest.The current project, implemented in five hospitals, was modeled after the previous project, but with greater emphasis on tool use. This allowed early identification of barriers to tool use and strategies to overcome barriers. Main outcome measures were AMI quality indicators in pre-measurement (January 1, 2001 to June 30, 2001) and post-measurement (December 15, 2001 to March 31, 2002) samples.One or more tools were used in 93% of patients (standard orders = 82%, and discharge document = 47%). Tool use was associated with significantly higher adherence to most discharge quality indicator rates with increases in aspirin, angiotensin-converting enzyme inhibitors, and smoking cessation and dietary counseling. Patients undergoing coronary artery bypass grafting (CABG) had low rates of discharge indicators. Patients undergoing percutaneous coronary revascularization were more likely to receive evidence-based therapies.These data validate the results of the pilot project that quality of AMI care can be improved through the use of guideline-based tools. Identifying and overcoming barriers to tool use led to substantially higher rates of tool use. The low rates of adherence to quality indicators in patients undergoing CABG suggest that these patients should be particularly targeted for quality improvement efforts.

Quality of care of patients with acute myocardial infarction (AMI) has received intense attention. However, it is unknown if a structured initiative for improving care of patients with AMI can be effectively implemented at a wide variety of hospitals.To measure the effects of a quality improvement project on adherence to evidence-based therapies for patients with AMI.The Guidelines Applied in Practice (GAP) quality improvement project, which consisted of baseline measurement, implementation of improvement strategies, and remeasurement, in 10 acute-care hospitals in southeast Michigan.A random sample of Medicare and non-Medicare patients at baseline (July 1998--June 1999; n = 735) and following intervention (September 1--December 15, 2000; n = 914) admitted at the 10 study centers for treatment of confirmed AMI. A random sample of Medicare patients at baseline (January--December 1998; n = 513) and at remeasurement (March--August 2001; n = 388) admitted to 11 hospitals that volunteered, but were not selected, served as a control group.The GAP project consisted of a kickoff presentation; creation of customized, guideline-oriented tools designed to facilitate adherence to key quality indicators; identification and assignment of local physician and nurse opinion leaders; grand rounds site visits; and premeasurement and postmeasurement of quality indicators.Differences in adherence to quality indicators (use of aspirin, beta-blockers, and angiotensin-converting enzyme [ACE] inhibitors at discharge; time to reperfusion; smoking cessation and diet counseling; and cholesterol assessment and treatment) in ideal patients, compared between baseline and postintervention samples and among Medicare patients in GAP hospitals and the control group.Increases in adherence to key treatments were seen in the administration of aspirin (81% vs 87%; P =.02) and beta-blockers (65% vs 74%; P =.04) on admission and use of aspirin (84% vs 92%; P =.002) and smoking cessation counseling (53% vs 65%; P =.02) at discharge. For most of the other indicators, nonsignificant but favorable trends toward improvement in adherence to treatment goals were observed. Compared with the control group, Medicare patients in GAP hospitals showed a significant increase in the use of aspirin at discharge (5% vs 10%; P.001). Use of aspirin on admission, ACE inhibitors at discharge, and documentation of smoking cessation also showed a trend for greater improvement among GAP hospitals compared with control hospitals, although none of these were statistically significant. Evidence of tool use noted during chart review was associated with a very high level of adherence to most quality indicators.Implementation of guideline-based tools for AMI may facilitate quality improvement among a variety of institutions, patients, and caregivers. This initial project provides a foundation for future initiatives aimed at quality improvement.