Health And Retirement Study Data Research Articles

DIVERSE ENVIRONMENTAL CONTEXTS AND LINKS TO LATER-LIFE WELL-BEING

Abstract The places older adults spend time have profound impacts on their well-being. As new sources of data and methods become available it is imperative to refine work in this area to guide development of interventions with the precision necessary to address persistent disparities in well-being. This symposium brings together four complementary papers that focus on diverse environmental contexts and well-being using multiple methodological approaches. Perzynski, Berg, and Dalton discuss the potential of Digital Twin Neighborhoods, i.e., digital replicas of real communities, to address health inequalities. They present findings from engagement sessions with community members about priorities, preferences, and concerns with regard to use of digital twins. Sol, Clarke and Zahodne link data from the Detroit-area Michigan Cognitive Aging Project with data from the National Neighborhood Data Archive to examine the association between neighborhood disadvantage and cognitive reserve. They show how this association differs between Black and White participants. Cho, Dunkle, and Smith use data from the Health and Retirement Study (HRS) to examine the association between later life relocation and contact frequency with adult children. They illustrate the importance of proximity and mode of contact in this association. Morris and colleagues also use HRS data to examine the extent to which stressors operating at multiple levels (interpersonal, community, and society) explain racial disparities in memory. They show these contexts together explain 11% of the racial disparity in baseline memory. These papers will be discussed by Markus Schafer who will provide an outlook for future research in this area.

GENES FROM PEROXISOME AND INTEGRATED STRESS RESPONSE IN REGULATION OF HUMAN LONGEVITY

Abstract Dietary/caloric restriction shares pathways with peroxisome, a cellular organelle that regulates the synthesis and turnover of complex lipids. PEX3, a peroxisomal biogenesis factor involved in selective type of autophagy. EIF2AK4, a sensor of amino acid deprivation in the integrated stress response (ISR) involved in lifespan regulation. We hypothesize that interplay between EIF2AK4 and PEX3 is associated with human survival. We applied logistic regression model with the interaction term to the analysis of Health and Retirement Study (HRS) and UK Biobank (UKB) parents data to test whether interactions of SNPs from the EIF2AK4 and SNPs from the PEX3 are associated with the dichotomous survival trait (ST) with case group defined as: lifespan ≥ 85; and control as: 75 ≤ (lifespan or age at last follow-up) < 85. Analyses of HRS data showed that interactions between SNPs rs4924405, rs2250402, rs8031319, rs16970049, rs3816899, rs16970112, rs2412459 from the EIF2AK4 gene and SNP rs161070 from the PEX3 gene are significantly associated with ST. In the UKB data, interactions of rs161070 from the PEX3 gene and rs7176881 and rs117778159 SNPs from the EIF2AK4 gene showed nominally significant associations with ST. Analyses of linkage disequilibrium (LD) between the EIF2AK4 SNPs detected in HRS and UKB data showed that their LD are close to the highest values for SNPs with allele frequencies specific for detected variants. Our results confirm association of interplay between EIF2AK4 and PEX3 with human longevity. They also show how information from experimental studies could be transformed to the form useful for human applications.

ASSOCIATION OF EVERYDAY LIFE INFORMATION ACQUISITION METHODS AND DEPRESSION AMONG US OLDER ADULTS

Abstract Everyday life information acquisition refers to the incidental ways people obtain information through daily activities such as reading, watching television, or communicating with friends rather than through intentional or purposive information seeking behaviors. It has been identified as a fundamental way to obtain informational resources and is a potential means to promote social connections and engagement among older adults. Research has shown that social engagement and connections are protective factors of depression for older adults. However, little research has explored how information seeking behaviors effect risks of depression among older adults. To fill this gap in the literature and provide insights into the mechanism of information access impacts the mental health of older adults, this study examines the relationships between different ways for obtaining daily information (reading, using computer, and watching television) and depression among older adults in United States. The data retrieved from 2018’s Health and Retirement Study (HRS). We multiple linear regression analyses on HRS data including adults age 50 years and older (n=3,179). After adjusting for age, gender, race, marital status, education, and chronic diseases, all three different ways to obtaining daily information (reading, using computer, and watching television) were negatively associated with depression (β=-0.19, p< 0.001; β=-0.15, p=0.002; β=-0.15, p< 0.001, respectively). Findings suggest that the association between everyday life information acquisition and depression among older adults varies depending on the specific information acquisition methods used. These findings highlight the importance of implementing targeted interventions to promote effective information access and enhance social inclusion among older adults.

Prevalence and Severity Distribution of Alzheimer’s Disease in the United States from the Health and Retirement Study

AbstractBackgroundThere is limited data regarding Alzheimer’s disease (AD) prevalence by severity across the AD continuum. This study aimed to estimate prevalence of mild cognitive impairment (MCI) and mild, moderate, and severe AD in a representative sample of the United States (US) population.MethodData from the Health and Retirement Study (HRS), a bi‐annual national survey of older adults in the US, were analyzed (2014‐2018), applying cluster sampling weights. AD ascertainment from HRS was based on patient report of a clinical diagnosis (physician diagnosis‐based), or the test score of Modified Telephone Interview of Cognitive Status (TICS‐m) (cognitive performance‐based). Patients with MCI were identified with the cognitive performance‐based method only. MCI and AD severity stages were ascertained using a crosswalk between the scores of TICS‐m and the mini‐mental state examination (MMSE).ResultThe diagnosis‐based method identified 1.2%, 1.2%, 1.3% patients with AD in 2014, 2016 and 2018, respectively (2.2% pooled). The mean (SD) age was 79.4 (10.3) years, with 61% women and 22% had passed college education (combined across surveys). The cognitive performance‐based method identified 23%, 23%, 21% patients with MCI in 2014, 2016 and 2018, respectively (24% pooled), and 27%, 22%, 17% patients with AD in 2014, 2016 and 2018, respectively (30% pooled). For patients with MCI, the mean age was 67.3 (10.1) years, with 51% women and 36% passed college education; for patients with AD, the mean age was 72.8 (11.1) years, with 53% women and 28% had passed college education. Severity distributions are shown in Figure 1. Unweighted results were similar.ConclusionThe cognitive performance‐based method identified a substantially higher prevalence of AD than the diagnosis‐based method and enabled identification of MCI prevalence using HRS data. The prevalence of MCI was consistently higher than AD dementia severity categories over the years (2014, 2016, and 2018). The discordance in AD prevalence estimation between cognitive performance‐based and diagnosis‐based methods underscores a need for better understanding of clinical practice patterns in AD diagnosis, use of clinical assessment tools, and severity classification in the US.

Mendelian randomization study using paired SNPs as instruments supports casual protective effect of overweight on Alzheimer’s disease

AbstractBackgroundMendelian Randomization (MR) is a powerful method of detecting causal effects in observational studies. However, choosing an appropriate SNP as an instrumental variable (IVar) is challenging because of difficulties in satisfying all assumption criteria. We proposed a novel strategy of creating IVar using paired SNPs between genes of interest, and tested this strategy in MR study examining causal relationships between being overweight and onset of Alzheimer’s disease (AD).MethodWe performed MR in the Health and Retirement Study (HRS) subsample of participants aged 75+, using the following binary variables: (1) ‘overweight’ BMI 25‐30 vs. (0) ‘normal’ BMI 18.5‐25, at ages 65‐75, as exposure; and (1) no AD vs. (0) AD onset at ages 75+, as outcome. We found significant association between ‘overweight’ BMI and lower odds of AD in logistic regression, and used MR to test causality. For IVar, we selected eight obesity‐related genes from the literature (ADIPOQ, FTO, LEP, LEPR, INSIG2, MC4R, PCSK1, PPARG), cross‐paired their SNPs, and used counts of minor alleles of the paired SNPs to create new ‘composite SNP’s. Analyses were stratified by sex and race, and non‐stratified. Initial models were estimated in SAS. Resulting summary statistics were used in R‐package MendelianRandomization. F‐value>10 criteria was applied to ensure the IVar strength. The causal effect was evaluated using weighted median, Inverse‐Variance Weighted (IVW), and maximum likelihood (ML).ResultBeing overweight at ages 65‐75 had significant causal protective effect on AD onset after age 75 in White females (logistic regression: OR = 1.99 [higher odds for ‘overweight’ individuals to be in ‘no AD’ group], p‐value = 8.26E‐5; MR: number of instrument ‘SNP’s = 13, weighted median causal estimate = 1.27, p‐value = 1.91E‐5; IVW causal estimate = 1.14, p‐value = 2.30E‐08; ML causal estimate = 1.20, p‐value = 7.67E‐7). No causal effects were detected using single SNPs in the same genes.ConclusionThe proposed novel strategy of creating IVar for MR analysis, based on pairing SNPs between candidate genes, substantially improves the critical step of MR analysis: choosing efficient instrument SNPs. We successfully tested this new strategy in HRS data, and confirmed causal relationship between BMI and AD, specifically, significant causal protective effect of being overweight on AD onset later in life in white females.

Subjective Memory Complaints Predict Decline in Memory, Instrumental Activities of Daily Living, and Social Participation in Older Adults: A Fixed-Effects Model.

Although subjective memory complaints (SMCs) have been suggested to be associated with future memory impairment, limitations in instrumental activities of daily living (IADLs), and social participation restriction, these associations are still inconclusive. To determine whether changes in SMCs over time predict decline in memory, IADLs, and social participation in older adults. Longitudinal study. Community. Sample 1 included 2,493 community-dwelling older adults drawn from the Health and Retirement Study (HRS) data collected between 2004 and 2018. Sample 2 included 1,644 community-dwelling older adults drawn from the HRS data collected between 2008 and 2018. Self-reported SMCs, memory function, self-reported IADL performance, and self-reported social participation. The mean age of Sample 1 at baseline was 70.16 yr; 1,468 (58.88%) were female. In Sample 1, immediate and delayed memory (all ps < .001) and IADL performance (p < .01) declined over time. Increases in SMCs over time significantly predicted future immediate and delayed memory declines (p < .01 and p < .001, respectively) and future IADL performance decline (p < .001), after controlling for depressive symptoms. The mean age of Sample 2 at baseline was 71.52 yr; 928 (56.45%) were female. In Sample 2, social participation declined over time (all ps < .001). Increases in SMCs over time significantly predicted future social participation decline (p < .05), after controlling for depressive symptoms. Increases in SMCs predict future decline in memory, IADL performance, and social participation after accounting for depressive symptoms. What This Article Adds: SMCs can be used as an early indicator of future memory impairment, IADL limitations, and social participation restrictions in older adults. Furthermore, interventions that minimize SMCs may help older adults achieve successful aging.

Association of Functional Status, Cognition, Social Support, and Geriatric Syndrome With Admission From the Emergency Department

The role of patient-level factors that are unrelated to the specific clinical condition leading to an emergency department (ED) visit, such as functional status, cognitive status, social supports, and geriatric syndromes, in admission decisions is not well understood, partly because these data are not available in administrative databases. To determine the extent to which patient-level factors are associated with rates of hospital admission from the ED. This cohort study analyzed survey data collected from participants (or their proxies, such as family members) enrolled in the Health and Retirement Study (HRS) from January 1, 2000, to December 31, 2018. These HRS data were linked to Medicare fee-for-service claims data from January 1, 1999, to December 31, 2018. Information on functional status, cognitive status, social supports, and geriatric syndromes was obtained from the HRS data, whereas ED visits, subsequent hospital admission or ED discharge, and other claims-derived comorbidities and sociodemographic characteristics were obtained from Medicare data. Data were analyzed from September 2021 to April 2023. The primary outcome measure was hospital admission after an ED visit. A baseline logistic regression model was estimated, with a binary indicator of admission as the dependent variable of interest. For each primary variable of interest derived from the HRS data, the model was reestimated, including the HRS variable of interest as an independent variable. For each of these models, the odds ratio (OR) and average marginal effect (AME) of changing the value of the variable of interest were calculated. A total of 42 392 ED visits by 11 783 unique patients were included. At the time of the ED visit, patients had a mean (SD) age of 77.4 (9.6) years, and visits were predominantly for female (25 719 visits [60.7%]) and White (32 148 visits [75.8%]) individuals. The overall percentage of patients admitted was 42.5%. After controlling for ED diagnosis and demographic characteristics, functional status, cognition status, and social supports all were associated with the likelihood of admission. For instance, difficulty performing 5 activities of daily living was associated with an 8.5-percentage point (OR, 1.47; 95% CI, 1.29-1.66) AME increase in the likelihood of admission. Having dementia was associated with an AME increase in the likelihood of admission of 4.6 percentage points (OR, 1.23; 95% CI, 1.14-1.33). Living with a spouse was associated with an AME decrease in the likelihood of admission of 3.9 percentage points (OR, 0.84; 95% CI, 0.79-0.89), and having children living within 10 miles was associated with an AME decrease in the likelihood of admission of 5.0 percentage points (OR, 0.80; 95% CI, 0.71-0.89). Other common geriatric syndromes, including trouble falling asleep, waking early, trouble with vision, glaucoma or cataract, use of hearing aids or trouble with hearing, falls in past 2 years, incontinence, depression, and polypharmacy, were not meaningfully associated with the likelihood of admission. Results of this cohort study suggest that the key patient-level characteristics, including social supports, cognitive status, and functional status, were associated with the decision to admit older patients to the hospital from the ED. These factors are critical to consider when devising strategies to reduce low-value admissions among older adult patients from the ED.

Trajectories of self-reported hearing and their associations with cognition: evidence from the United Kingdom and United States of America.

This study aimed to investigate the relationships between trajectories of change in self-reported hearing over eight years with subsequent effects on cognition, measured using episodic memory. Data were drawn from 5 waves (2008-2016) of the English Longitudinal Study of England (ELSA) and the Health and Retirement Study (HRS), involving 4,875 individuals aged 50+ at the baseline in ELSA and 6,365 in HRS. The latent growth curve modelling was used to identify trajectories of hearing over eight years, and linear regression models were performed to investigate the relationship between hearing trajectory memberships and episodic memory scores, controlling for confounding factors. Five trajectories of hearing (stable very good, stable fair, poor to fair/good, good to fair, and very good to good) were retained in each study. Individuals whose hearing remains suboptimal and those whose hearing deteriorates within suboptimal levels throughout eight years have significantly poorer episodic memory scores at follow-up than those with stable very good hearing. Conversely, individuals whose hearing declines but is within an optimal category at baseline do not see significantly poorer episodic memory scores than those with consistently optimal hearing. There was no significant relationship between individuals whose hearing improved from suboptimal baseline levels to optimal by follow-up and memory in ELSA. However, analysis using HRS data shows a significant improvement for this trajectory group (-1.260, P< 0.001). Either stable fair or deterioration in hearing is associated with worse cognitive function, both stable good or improving hearing is associated with better cognitive function specifically episodic memory.

Readmissions and postdischarge mortality by race and ethnicity among Medicare beneficiaries with multimorbidity.

Disparities in readmission risk and reasons they might exist among diverse complex patients with multimorbidity, disability, and unmet social needs have not been clearly established. These characteristics may be underestimated in claims-based studies where individual-level data are limited. We sought to examine the risk of readmissions and postdischarge mortality by race and ethnicity after rigorous adjustment for multimorbidity, physical functioning, and sociodemographic and lifestyle characteristics. We used Health and Retirement Study (HRS) data linked to Medicare claims. To obtain ICD-9-CM diagnostic codes to compute the ICD-coded multimorbidity-weighted index (MWI-ICD) we used Medicare Parts A and B (inpatient, outpatient, carrier) files between 1991-2015. Participants must have had at least one hospitalization between January 1, 2000 and September 30, 2015 and continuous enrollment in fee-for-service Medicare Part A 1-year prior to hospitalization. We used multivariable logistic regression to assess the association of MWI-ICD with 30-day readmissions and mortality 1-year postdischarge. Using HRS data, we adjusted for age, sex, BMI, smoking, physical activity, education, household net worth, and living arrangement/marital status, and examined for effect modification by race and ethnicity. The final sample of 10,737 participants had mean ± SD age 75.9 ± 8.7 years. Hispanic adults had the highest mean MWI-ICD (16.4 ± 10.1), followed by similar values for White (mean 14.8 ± 8.9) and Black (14.7 ± 8.9) adults. MWI-ICD was associated with a higher odds of readmission, and there was no significant effect modification by race and ethnicity. For postdischarge mortality, a 1-point increase MWI-ICD was associated with a 3% higher odds of mortality (OR=1.03, 95% CI: 1.03-1.04), which did not significantly differ by race and ethnicity. Multimorbidity was associated with a monotonic increased odds of 30-day readmission and 1-year postdischarge mortality across all race and ethnicity groups. There was no significant difference in readmission or mortality risk by race and ethnicity after robust adjustment.

Breast and prostate cancer screening rates by cognitive status in US older adults.

For most older adults with dementia, the short-term harms and burdens of routine cancer screening likely outweigh the delayed benefits. We aimed to provide a more updated assessment of the extent that US older adults with dementia receive breast and prostate cancer screenings. Using the Health and Retirement Study (HRS) Wave 12 (2014-2015) linked to Medicare, we examine rates of breast and prostate cancer screenings in adults 65+ years by cognitive status. We used claims data to identify eligibility for screening and receipt of screening. We used a validated method using HRS data to define cognitive status. The analytic sample included 2439 women in the breast cancer screening cohort and 1846 men in the prostate cancer screening cohort. Average ages were 76.8 years for women and 75.6 years for men, with 9.0% and 7.6% with dementia in each cohort, respectively. Among women with dementia, 12.3% were screened for breast cancer. When stratified by age, 10.6% of those 75+ and have dementia were screened for breast cancer. When stratified by predicted life expectancy, 10.4% of those with predicted life expectancy of <10 years and have dementia were screened for breast cancer. Among men with dementia, 33.9% were screened for prostate cancer. When stratified by age, 30.9% of those 75+ and have dementia were screened for prostate cancer. When stratified by predicted life expectancy, 34.4% of those with predicted life expectancy of <10 years and have dementia were screened for prostate cancer. Using multivariable logistic regression, dementia was associated with lower odds of receiving breast cancer screening (OR 0.36, 95% CI 0.23-0.57) and prostate cancer screening (OR 0.58, 95% CI 0.36-0.96). Our results suggest potential over-screening in older adults with dementia. Better supporting dementia patients and caregivers to make informed cancer screening decisions is critical.

A Large-Scale Prospective Study on Functional Ability and Mortality Following Stroke Based on HRS Data Set: Implications for Rehabilitation.

According to the WHO, 15 million people worldwide suffer a stroke annually. Of these, 5 million die and another 5 million are left permanently disabled. Patients endure the impacts of strokes during their rehabilitation and afterward, placing economical and emotional burdens on family and community. Using data from the Health and Retirement Study (HRS) of the USA, the research performed a large-scale prospective analysis to examine how demographics, socioeconomic factors, cognition, emotion, and physical activity predict functional impairment and mortality. Multiple regression was employed to identify statistically significant variables that predict longitudinal Activities of Daily Life (ADLs). The least absolute shrinkage and selection operator (LASSO) logistic regression, a supervised machine learning approach, was deliberately chosen to obtain the subset of predictors that provide the best possible classification for the dependent variable. The LASSO regression produced a model with a fair mean Area Under the Curve (AUC) of 0.75 in predicting the risk of the patient's mortality. My findings also uncovered the important roles of BMI, mobility, muscle strength, memory, mental status, and socioeconomic status play in the long-term ADLs and survival of patients with stroke.

Enhancing the Utility of the Health and Retirement Study (HRS) to Identify Drivers of Rising Mortality Rates in the United States.

A recent report from the National Academies of Sciences, Engineering, and Medicine (NASEM) highlights rising rates of working-age mortality in the United States, portending troubling population health trends for this group as they age. The Health and Retirement Study (HRS) is an invaluable resource for researchers studying health and aging dynamics among Americans ages 50 and above and has strong potential to be used by researchers to provide insights about the drivers of rising U.S. mortality rates. This paper assesses the strengths and limitations of HRS data for identifying drivers of rising mortality rates in theU.S. and provides recommendations to enhance the utility of the HRS in thisregard. Among our many recommendations, we encourage the HRS to prioritize the following: link cause of death information to respondents; reduce the age of eligibility for inclusion in the sample; increase the rural sample size;enhance the existing HRS Contextual Data Resource by incorporating longitudinal measures of structural determinants of health; develop additional data linkages to capture residential settings and characteristics across the life course; and add measures that capture drug use, gun ownership, and social media use.

Building and Using Data Resources for Research on Job Characteristics, Health, and Retirement

A growing literature seeks to understand the relationship between the experience of work and important later-life outcomes. Rich longitudinal measurement of both sides of this equation in datasets such as the Health and Retirement Study (HRS) have made this research possible. These data take the form of self-reported experiences of work (such as physical demands, job flexibility, job satisfaction etc.). Increasingly, researchers are looking to add potentially complementary information on the work environment available in the Occupational Information Network (O*NET) database through a linkage using occupation and industry codes in the survey data. The session talks will describe research conducted using O*NET linked with HRS data as well as ongoing work to create a new data resource that will allow other researchers to undertake research with O*NET-HRS linked data. Each presentation will include some discussion of both the value and limits of using the linkage to O*NET. Carpenter will provide a detailed description a new project linking the 2019 O*NET data to the HRS for public use.This presentation explains the types of variables that will be made available in the O*NET-HRS occupation project and will provide examples for how the measures can be used in longitudinal HRS studies. Using O*NET-HRS linked data, Carr will present on work examining the role of preretirement job complexity in alternative retirement paths and cognitive performance. Helppie-McFall will used the linked data to discuss the role of mismatch between demands of work and workers’ ability to meet those demands in retirement decisions.

The Effect of Traumatic Brain Injury on Alzheimer’s Disease and Cognitive Decline in Veterans and Non-Veterans

We assess the differences in the effect of traumatic brain injury (TBI) on the decline in cognitive status and the risk of Alzheimer’s disease and related dementia (AD/ADRD) between veteran and non-veteran respondents of the Health and Retirement Study (HRS) and measure the sensitivity of these differences to the incremental introduction of controls for associated risk factors. Three groups of AD/ADRD risk-related variables were used: i) demographic/socioeconomic factors, including gender, race, marital status, education, income, and the number of limitations in activities of daily living; ii) comorbidities, including co-existing depression/post-traumatic stress syndrome (PTSD), substance (alcohol, tobacco and/or prescription drug) abuse, diabetes mellitus, stroke, and heart failure; and iii) genetic factors, including the presence of at least one pair of the APOE4 allele and a series of polygenic risk scores associated with AD hallmarks. The dynamics of changes in cognitive impairment in response to TBI, PTSD, and mild cognitive impairment were validated against respective measures estimated using the Department of Defense Alzheimer’s Disease Neuroimaging Initiative (DoD-ADNI) data. The results of the analyses showed that TBI and PTSD were strongly associated with cognitive decline and the risks of AD/ADRD in both veteran and non-veteran subpopulations in HRS data and the difference between them was not statistically significant. Effect magnitude decreased with the addition of risk-related control variables but remained associated with the increased risks. Prevalence of mild cognitive impairment was associated with TBI at baseline in DoD-ADNI data, but no cognitive decline was observed during one year of follow-up.

Oral Health, Cognitive Function, and Mortality: Findings From National Surveys

Poor oral health, diabetes mellitus (DM), and cognitive impairment are common problems in older adults. Using national surveys, this symposium aims to present new findings regarding the impact of the co-occurrence of DM and poor oral health on cognitive function, cognitive decline, and mortality. This symposium will also cover the topic of dental care use among adult populations in the U.S. Using data from the Health and Retirement Study (HRS) (2006- 2018), the first study shows that adults with both DM and edentulism had the worst cognitive function, followed by those with edentulism alone, and those with DM alone. Using the same HRS data, the second study found that co-occurrence of DM and edentulism had a higher risk of more rapid cognitive decline with advancing age than the presence of each condition alone. The third study used data from the 2006-2016 HRS linked with mortality files, and revealed that the risk of diabetes and edentulism on mortality may vary across racial/ethnic groups. Using the Behavioral Risk Factor Surveillance System survey (2002-2018), the fourth study examined disparities of dental service utilization among racial/ethnic groups (Whites, Hispanics, Blacks, Asians, American Indians or Alaska Natives, and Native Hawaiian or other Pacific Islanders). Age differences in dental services were also compared between older adults and other younger and middle-aged populations. This symposium highlights the role of oral health in improving cognitive health. Policies and programs are needed to increase dental care access, a critical way to help maintain good oral health.

160 Underlying Factors Contributing to Sleep Health Among Middle-aged and Older Adults

Abstract Introduction Although poor sleep is not inherent with aging, an estimated 50-70 million adults in the US have insufficient sleep. Sleep duration is increasingly recognized as incomplete and insufficient. Instead, sleep health (SH), a multidimensional concept describing sleep/wake patterns that promote well-being has been shown to better reflect how sleep impacts the individual. Therefore, focusing on the underlying factors contributing to sleep health may provide the opportunity to develop interventions to improve sleep health in middle-age and older adults. Methods Data from the 2014 wave of the Health and Retirement Study (HRS) were used. Sample size was restricted to those who completed an additional questionnaire containing sleep variables. A derivation of the SH composite was constructed using eight selected sleep variables from the HRS data based on the five dimensions of sleep: Satisfaction, Alertness, Timing, Efficiency, and Duration. Total score ranged from 0-100, with higher scores indicating better SH. Weighting variables were based on complex sampling procedures and provided by HRS. Machine learning-based framework was used to identify determinants for predicting SH using twenty-six variables representing individual health and socio-demographics. Penalized linear regression with elastic net penalty was used to study the impact of individual predictors on SH. Results Our sample included 5,163 adults with a mean age of 67.8 years (SD=9.9; range 50-98 years). The majority were female (59%), white (78%), and married (61%). SH score ranged from 27-61 (mean=50; SD=6.7). Loneliness (coefficient=-1.92), depressive symptoms (coefficient=-1.28), and physical activity (coefficient=1.31) were identified as the strongest predictors of SH. Self-reported health status (coefficient=-1.11), daily pain (coefficient=-0.65), being middle-aged (coefficient=-0.26), and discrimination (coefficient=-0.23) were also significant predictors in this model. Conclusion Our study identified key predictors of SH among middle-aged and older adults using a novel approach of Machine Learning. Improving SH is a concrete target for health promotion through clinical interventions tailored towards increasing physical activity and reducing loneliness and depressive symptoms among middle-aged adults. Support (if any) This study was supported by National Heart, Lung, and Blood Institute (NHLBI) UB Clinical Scholar Program in Implementation Science to Achieve Triple Aims-NIH K12 Faculty Scholar Program in Implementation Science

Financial Decision Making and Survey Responses of Older Couples

The Health and Retirement Study (HRS) asks mid-life and older couples to identify the person most knowledgeable about household finances to provide responses to financial questions (“financial respondent” [FR]). The purpose of this study is to analyze predictors of a female FR, with a focus on gendered financial decision making. This study analyzed HRS data from 2012 to 2018. Our sample consisted of 5,038 married/partnered couples over the age of 50 (at baseline), after eliminating cases with missing data on study variables. Using HLM 6.08 software, we computed the odds of couple households having a female FR. Our main predictor was the amount of female financial decision making power, as measured using eight Likert scale items (0=male always makes this decision, 4=female always makes this decision), for purchases (e.g., cars, major appliances) and financial activities (e.g., paying bills, buying groceries); summed into a scale (range: 0-32). Our models controlled for race, ethnicity, education, income, age, health, cognitive ability, and depressive symptoms. Results indicate that men were more likely to be the FR, but when the female makes more household financial decisions then she is more likely to be the FR (OR = 1.16, p<.01). Black race, lower income, and time also increased the odds of a female FR. This study highlights women’s historical and ongoing limited power in financial matters of married couples. Financial knowledge and decision making skills should be encouraged for both partners in couple households.

PHASED RETIREMENT IN OLDER AMERICAN WORKERS

Older workers are engaging in “phased retirement” in which they transition from full-time working status to part-time working and eventually retire at a later age. This study investigated whether phased retirement was financially and psychologically beneficial for middle-aged and older adults in the United States. The current study examined data on financial and psychological well-being and retirement transitions (i.e., immediate retirement and phased retirement) from the Health and Retirement Study (HRS) at two time points, four years apart. We analyzed 5,106 middle-aged and older adults from the 2010 and 2014 waves of the HRS data set using chi-square and one-way analysis of variance tests. Our results showed that 66.8% of respondents remained full-time working, while 12.7% of respondents chose phased retirement and 15.2% of respondents jumped straight into retirement. Our findings suggest that phased retirement is beneficial for older Americans financially and psychologically. Specifically, the level of total household financial wealth was significantly higher for those who chose phased retirement than immediate retirees from workforce. In terms of psychological perspectives, immediate retirees experienced more depressive symptoms than those who chose phased retirement. Our findings could help individuals and households to be better equipped when preparing for retirement. Our findings could also provide a basis for further research into phased retirement and its impact on well-being in middle-aged and older Americans. Furthermore, policymakers could be better informed about retirement trends and create policies based on our findings to better help older individuals and households be financially and psychologically prepared for retirement.

EXPLORING THE MULTIDIMENSIONAL BIOPSYCHOSOCIAL PATHWAYS LINKING SOCIAL STATUS AND HEALTH AMONG OLDER ADULTS

An abundant literature has documented the social patterning of health, where those with lower social status experience poorer outcomes relative to those with higher status. This symposium examines how social status (e.g., age, race/ethnicity, gender, and SES) impacts various aspects of midlife and older adults’ lives and their psychological and physical health. The research presented in this symposium lend support to utilizing a biopsychosocial framework for understanding mechanisms of health and aging. First, Heather Farmer et al. will explore race and gender differences in elevated C-reactive protein (CRP), a marker of inflammation linked to poor acute and chronic outcomes, using data from the Health and Retirement Study (HRS). Linda Wray and Amy Thierry will use HRS data to test whether race/ethnicity and sex interact to produce unequal outcomes in functional status. Jen Wong et al. will utilize data from the Midlife in the United States (MIDUS) survey to investigate the moderating influences of age, gender, marital status, and social support on caregiving and psychological well-being. Collin Mueller and Heather Farmer will use HRS data to examine how perceptions of unfair treatment are associated with healthcare satisfaction and self-rated health across Black, Latinx, and White subpopulations. Taken together, this work highlights the need for a comprehensive approach to better address physical and mental health disparities over the life course. After attending this session, participants will have a stronger understanding of how social status shapes important outcomes in older adults’ lives and some of the mechanisms responsible for these variations.

VALIDITY OF ALGORITHMIC DIAGNOSES OF DEMENTIA: A CASE STUDY OF APPLICATION TO RESEARCH ON RACE/ETHNIC DISPARITIES IN DEMENTIA

Understanding of race/ethnic disparities in Alzheimer's disease and related dementia (ADRD) is limited by the paucity of diverse cohorts with formal dementia ascertainment. Algorithmic diagnosis of ADRD provides an attractive alternative. Here we evaluate the performance of algorithmic versus gold-standard ADRD classification for the purpose of quantifying and understanding race-ethnic disparities. We used data from the nationally representative Health and Retirement Study (HRS) and its sub-study, the Aging, Demographics, and Memory Study (ADAMS) study, which conducted gold-standard, study-based dementia ascertainment. We identified five existing algorithms for algorithmic classification of dementia using HRS data. We also developed a new algorithm for dementia classification that minimized differences in sensitivity and specificity by race/ethnicity. We classified ADAMS Wave A participants as normal or demented using all six algorithms. In a sample restricted to participants with sufficient data to apply all algorithms, we compared findings using gold-standard and algorithmic diagnoses when estimating (a) the overall prevalence of dementia, (b) the odds ratio (OR) comparing dementia prevalence in non-Hispanic blacks to whites, and (c) the OR quantifying the association between prevalent diabetes and prevalent dementia overall and within race-groups. Models were adjusted for age, age2, race, sex, and HRS proxy status, and analyses were weighted to recover nationally representative estimates. While overall prevalence of dementia using gold-standard diagnoses was 12% (95%CI: 10%, 15%), estimates obtained using algorithmic diagnoses ranged from 6% to 17%. With the exception of our new algorithm designed to minimize sensitivity and specificity differences by race-ethnicity, all of the existing algorithms substantially overestimated the OR comparing blacks to whites (OR: 1.9 (95%CI: 0.8, 4.1) using the gold-standard diagnoses, OR: 1.8 (95%CI: 0.8, 4.3) using our new algorithm, ORs ranging from 2.8 to 5.8 using the other 5 algorithms). In overall or race-stratified adjusted models, use of some, but not all algorithms produced associations between diabetes and dementia that were similar to those obtained using gold-standard diagnoses.