Alzheimer’s disease (AD) is a complex neurodegenerative disorder influenced by various factors, including genetic and exposure-related. Certain combinations of these factors may promote AD more substantially than others. APOE4 is the strongest genetic risk factor for AD. Traffic-related air pollution (TRAP) and infections are important exposure-related AD risk factors. Here we investigated how the interplay between a history of infections and chronically high exposure to TRAP (highTRAP) impacts the subsequent risk of AD and other dementias (AD+) in carriers and non-carriers of APOE4 in UK Biobank (UKB) participants aged 60–75 years. HighTRAP was approximated by the proximity (50 meters or less) of a participant’s primary residence to a major road. Chi-square, Wilson score interval, Wald interval, Wald risk ratio, Welch tests, and regression were used to examine statistical significance. We found that UKB participants with a history of various infections (by ICD-10 codes), but without highTRAP, had a 54% increase in AD+ risk. HighTRAP alone did not significantly influence AD+ risk. Individuals with both a history of infections and highTRAP had a 164% higher risk of AD+ compared to those without either factor. That risk was much higher (349%) in non-carriers of APOE4 but became non-significant in APOE4 carriers. We conclude that avoiding high exposure to TRAP may significantly reduce the risk of AD in non-carriers of APOE4 with a history of infections but not in carriers. One potential explanation could be that APOE4 is a stronger AD risk factor, whose AD-promoting effects may outweigh those of other risk factors.

Background and Objectives:The distribution of life expectancy at age 65 (LE65 ) in the United States is characterized by profound sex- and locality-related disparities. Quantifying the disease-specific drivers of these disparities prior to the coronavirus disease 2019 (COVID-19) pandemic provides a critical baseline for understanding subsequent health shocks. Methods:Using CDC WONDER data (1999-2018) and Pollard's decomposition method, we analyzed cause-specific contributions to disparities in LE65, establishing a pre-pandemic baseline for trends in years of life lost (YLL) at age 65. Results:Sex-related disparities (YLLsex) narrowed, particularly in disadvantaged localities, driven by reductions in YLL from macrovascular diseases and lung cancer. Conversely, locality-related disparities (YLLloc) widened, especially for females, due to increasing YLL from Alzheimer's disease and the persistent impacts of diabetes, chronic lower respiratory diseases, and heart failure. This created a pre-existing landscape of vulnerability concentrated in low-LE states. Discussion and Implications:The 2 decades before COVID-19 saw a critical divergence: progress in reducing sex-based gaps was offset by rapidly widening geographic disparities. The systemic failures that drove the increasing burden of chronic conditions in disadvantaged regions likely predetermined the populations most vulnerable to the pandemic's shock. Our findings highlight that building future resilience requires targeted investments to address these specific, pre-existing health challenges.

A new mathematically exact method of population attributable fraction (PAF) decomposition free of limitations inherent in existing approaches was developed and compared to existing methods based on the Miettinen, Norton, and Niedhammer-Chastang formulae. The developed approach is applicable for two broadly used study designs involving either a single or multiple data sources to measure the predictors and outcomes of interest. The approach was applied to Medicare data to estimate six disease-specific contributions to the overall PAF of Alzheimer's disease risk: stroke (7.2%), hypertension (6.5%), diabetes (4.4%), renal disease (0.9%), traumatic brain injury (0.9%), and depression (9.6%). We found that the approximation based on the Norton formula was the best approach among the methods utilized prior to the development of our approach. However, the quality of such approximations and the respective biases should be re-estimated on a case-by-case basis. An extension of the approach to health disparities was proposed and discussed.

Although numerous risk factors for Alzheimer’s disease (AD) and related dementias (ADRD) have been identified, their combined influence on risk remains uncertain. In this study, we leverage a high-power 5%-sample of the Medicare population to assess the joint impact of AD/ADRD risk-related diseases and low income—as indicated by dual Medicare eligibility—on the risk of clinical diagnosis of AD/ADRD. We leveraged the univariable and multivariable Cox models for estimating AD/ADRD risks, identified most powerful predictors, and constructed predictive multivariable models for AD/ADRD risks, and evaluated their population attributable fractions (PAFs) for the general populations of older adults and race- and sex-specific subpopulations. The identified model included nine diseases—heart failure, hypertension, arrhythmia, stroke, hypotension, renal disease, depression, traumatic brain injury, and diabetes mellitus—as primary determinants of variation in AD/ADRD risk. The fraction of the total PAF explained by the predictors increased with age. The decomposition of the total PAF in terms of disease-specific PAFs showed that hypertension, stroke, and depression provided the strongest contribution for all subpopulations: Hypertension reached 45-50% of total PAF for Black, Asian, and Hispanic subpopulations. Stroke was the primary contributor for males (>30%); depression was highest among females and White individuals (approximately 30%). Hypertension and depression were the leading contributors for Native Americans (both approximately 30%). Heart Failure was the fourth strongest contributor for all subpopulations, with the contribution exceeding 5% for Native American individuals. Additional noteworthy contributors with PAFs exceeding 3% included hypotension (males), diabetes (Asian, Hispanic), arrhythmia (males), and TBI (Black, Asian).

Existing literature lacks information on the relationship between exposure to ambient air contaminants and Alzheimer’s disease (AD) risk represented by simultaneous assessment of absolute and relative concentration-response patterns. This study focuses on quantifying PM2.5-exposure and AD-risk relationships with detailed analyses of the role of methodologic limitations and sources of bias (e.g., population heterogeneity) and identifying notable subpopulation differences and populations most vulnerable to PM2.5 exposure. Using Medicare and SEDAC data, we evaluated the effects of PM2.5 exposure on AD risk stratified by sex, race/ethnicity, and exposure level using absolute concentration-response functions based on age-adjusted rates and relative functions estimated via Cox models. The results show that chronic low-level exposures to PM2.5 were associated with higher AD risk which varied substantially across sex- and race-specific population groups. Females showed higher absolute incidence, while males exhibited steeper relative risk increases with rising PM2.5 concentrations. Native Americans, residing primarily in low-exposure areas, demonstrated the steepest rise in hazard with increasing PM2.5 concentrations; Black and Asian subpopulations exhibited the lowest relative ratios. Nonlinearities were detected in exposure-response relationships. AD risk in Hispanics was subject to notable geographic heterogeneity. This study provides a foundation for assessing the health impacts of recent U.S. air quality standards and their effects on AD risk in vulnerable populations. Future research should incorporate flexible modeling techniques (e.g., spline-based or piecewise methods) and integrate mechanistic studies to clarify the biological and social pathways linking PM2.5 to AD, explain observed disparities, and identify underlying causal factors.

Alzheimer’s disease and related dementias (AD/ADRD) pose a critical public health challenge, further complicated by the high prevalence of neuropsychiatric symptoms (NPS). Using a 5% Medicare sample, we quantified the epidemiology of NPS in relation to AD/ADRD onset and survival. We identified the onset dates for nine NPS—aggression, delirium, wandering, restlessness/agitation, hallucinations, anxiety, demoralization/apathy, sleep disturbances, and psychosis—and estimated their prevalence, the hazard ratios for developing AD and non-AD ADRD, and the hazard ratios for mortality following AD/ADRD diagnoses. We found that prevalence increases with age; maximum values at AD diagnosis were detected for anxiety (38.3%), psychosis (33.1%), and delirium (17.6%). Hazard ratios for AD/ADRD risk were evaluated using the Cox model with predictors measured with a one-year lag to avoid accounting ties between NPS and AD/ADRD onsets. The highest effects were detected for restlessness and agitation (HR = 4.30, CL = 4.21-4.39), psychosis (4.16; 4.12-4.20), and demoralization and apathy (4.15; 3.28-5.26). The effect of any NPS was 2.47 (2.45-2.48). Finally, we calculated the death hazard ratios for cohorts of individuals with diagnoses of AD and ADRD using a left-truncation design and age as the time-scale variable. The highest effect was detected for restlessness and agitation (1.61; 1.59-1.62) and delirium (1.52; 1.51-1.53). The death hazard ratio for any NPS was (1.46; 1.45-1.47). Our robust and biologically interpretable estimates provide new knowledge on underlying pathological processes—including an improved understanding of NPS and their treatment, and additional clues to potential underlying neuropathology—and to facilitate applications at individual-patient and population levels.

INTRODUCTION:Disparities in Alzheimer's disease (AD) and related dementias (ADRD) persist across race/ethnicity, sex, and US geographic regions, but limited quantitative information exists to explain how specific predictors contribute to these disparities. Many traditional methods lack precision in addressing both exposure (higher prevalence of a predictor) and vulnerability (higher risk associated with a predictor) effects. This study introduces an approach that leverages population attributable fraction (PAF) to analyze and explain AD/ADRD disparities using Medicare data. METHODS:We applied our method to Medicare claims data from a nationally representative sample of the US adults aged 70, 75, 80, and 85. The analysis focused on six types of disparities: Black-White, Hispanic-White, Native American-White, Asian-White, female-male, and stroke-belt versus non-stroke-belt states. Predictors included Medicare/Medicaid dual eligibility as an indicator of low income and 10 AD/ADRD-related diseases. The method quantified the exposure and vulnerability effects of each predictor on the observed disparities. RESULTS:Low income and vulnerability to arterial hypertension were the primary contributors to AD/ADRD disparities, with cerebrovascular diseases and depression as notable secondary predictors. The exposure effect dominated for income-related disparities, while hypertension's effect was largely driven by increased vulnerability. Racial disparities (Black-White, Hispanic-White) were most affected by income and hypertension, while female-male and stroke-belt disparities were less influenced by the examined predictors. DISCUSSION:Our findings indicate that different intervention strategies are needed to address AD/ADRD disparities. Income-related disparities require targeting exposure (e.g., socioeconomic improvements), while hypertension-related disparities suggest a focus on managing vulnerability (e.g., better control of hypertension). The developed approach offers a robust framework for explaining disparities and designing targeted interventions. Further application to other datasets and exploration of additional predictors could enhance understanding and lead to more effective prevention strategies for AD/ADRD disparities. Highlights:Our new approach addresses disparities leveraging the concept of population attributable fraction for Cox models.Exposure and vulnerability mechanisms of health disparity generation are evaluated.Vulnerability to hypertension is a consistent dominant factor in Alzheimer's disease (AD) risk disparities.Predictors explain AD disparities better in Black and Hispanic populations.Disparities in AD are driven by exposure to socioeconomic status suggesting targeted interventions.

Disparities in Alzheimer’s disease (AD) and related dementias (ADRD) persist across race/ethnicity, sex, and U.S. regions, yet few quantitative studies clarify how specific predictors drive these differences. Traditional methods often fall short by not addressing both the higher prevalence (exposure) and the increased risk (vulnerability) associated with a predictor. We applied two advanced approaches—the Powers-Yun decomposition technique and our recent PAF decomposition method—to quantify the exposure and vulnerability effects of each predictor using Medicare claims data from a nationally representative sample of U.S. adults aged 70, 75, 80, and 85. The analysis focused on six types of disparities: Black-White, Hispanic-White, Native American-White, Asian-White, Female-Male, and Stroke-Belt vs. non-Stroke-Belt states. Predictors included low-income status (ascertained through Medicare/Medicaid dual eligibility) and ten AD/ADRD-related diseases. We found that low income and vulnerability to arterial hypertension were the primary contributors to AD/ADRD disparities, with cerebrovascular diseases and depression as notable secondary predictors. The exposure effect dominated for income-related disparities, while hypertension’s effect was largely driven by increased vulnerability. Racial disparities (Black-White, Hispanic-White) were most affected by low income and hypertension, while Female-Male and Stroke-Belt disparities were less influenced by these predictors. Our findings indicate that different intervention strategies are needed to address AD/ADRD disparities. Low income-related disparities require targeting exposure (e.g., socioeconomic improvements), while hypertension-related disparities suggest a focus on managing vulnerability (e.g., better control of hypertension). The developed methodology offers a robust framework for explaining disparities and designing targeted interventions.

Background Existing literature lacks information on the relationship between exposure to ambient air contaminants and Alzheimer's disease (AD) risk represented by simultaneous assessment of absolute and relative concentration-response patterns. Objective This study quantifies fine particulate matter (PM 2.5 )-exposure and AD-risk relationships with detailed analyses of the role of methodologic limitations and sources of bias (e.g., population heterogeneity) and identifies notable subpopulation differences and subpopulations most vulnerable to PM 2.5 exposure. Methods Using Medicare and SEDAC data, we evaluated effects of PM 2.5 exposure on AD risk using absolute and relative concentration-response functions through, respectively, age-adjusted rates and Cox-model estimates by sex/race/ethnicity/exposure groups. Results Analyses demonstrated that chronic PM 2.5 exposure was significantly associated with increased AD risk, with substantial variation observed across sex and race/ethnicity-specific subgroups. Females showed higher absolute incidence, while males exhibited steeper relative risk increases with rising PM 2.5 concentrations. Native Americans, residing primarily in low-exposure areas, demonstrated the steepest rise in hazard with increasing PM 2.5 concentrations; Black and Asian subpopulations exhibited the lowest relative ratios. Nonlinearities were detected in exposure-response relationships. AD risk in Hispanics displayed notable geographic heterogeneity. Conclusions This study provides a background for evaluating health impacts of recent air-quality standards across the U.S. and their specific effects on AD risk in vulnerable subpopulation groups. Future research should employ more-flexible modeling techniques (e.g., spline-based or piecewise methods), integrate mechanistic studies, and utilize data on social determinants to further elucidate the biological and structural pathways linking PM 2.5 to AD diagnosis, explain observed disparities through associated risk factors, and identify causal factors mediating the relationship.

Abstract Background Heart failure (HF) is a challenging clinical and public health problem characterized by high prevalence and mortality among US older adults, along with a recent decline in HF prevalence and increase in mortality. The changes of prevalence can be decomposed into pre-existing disease prevalence, disease incidence, and respective survival, while the changes of mortality can be decomposed into mortality in the general population independent from HF, pre-existing HF prevalence, incidence, and respective survival. These epidemiological components may contribute differently to the changes in prevalence and mortality. Objective We aimed to investigate and compare the relative contributions of epidemiologic determinants in HF prevalence and mortality trends. Methods This study was a secondary data analysis of 5% of Medicare claims data for 1992‐2017 in the United States. Medicare is a federal health insurance program for older adults aged 65+ years as well as people with specific disabilities and end-stage renal disease. Age-adjusted prevalence and incidence-based mortality (IBM; all-cause mortality that occurred in patients with HF) were partitioned into their respective epidemiologic determinants using the partitioning analysis approach. Results The age-adjusted HF prevalence (1/100 person-years) increased from 11 in 1994 to 14.6 in 2005, followed by a decline to 12.6 in 2017, and the age-adjusted HF IBM (1/100,000) increased from 2220.8 in 1994 to 2563.7 in 2000, then declined to 2075.9 in 2016, followed by an increase to 2094.7 in 2017. The HF incidence (1/1000 person-years) declined from 29.4 in 1992 to 19.9 in 2017. The 1-, 3-, and 5-year survival trend showed declines in recent years. Partitioning of HF prevalence showed three phases: (1) decelerated increasing prevalence (1994‐2006), (2) accelerated declining prevalence (2007‐2014), and (3) decelerated declining prevalence (2015‐2017). During the whole period, the decreasing HF incidence contributed to the declines in prevalence, overpowering prevalence increases contributed from survival. Likewise, partitioning of HF IBM showed three phases: (1) decelerated increasing mortality (1994‐2001), (2) accelerated declining mortality (2002‐2012), and (3) decelerated declining mortality (2013‐2017). The decreasing HF incidence in 1994‐2017 and increasing survival in 2002‐2006 contributed to the declines in mortality, while the decreasing survival in 2007‐2017 contributed to the mortality increase. Conclusions Decade-long declines in HF prevalence and mortality mainly reflected decreasing incidence, while the most recent increase of mortality was predominantly due to the declining survival. If current trends persist, HF prevalence and mortality are forecasted to grow substantially in the next decade. Prevention strategies should continue the prevention of HF risk factors as well as improvement of treatment and management of HF after diagnosis.

Abstract Improving Life expectancy (LE) in the LE-lagging states could significantly improve overall LE and eliminate health inequalities in the United States. We analyzed the disparity in mortality rates between eight states with high and low LE at 65 using individual morbidity trajectories in Medicare administrative claims data from 2000-2020. Disease-specific contributions were identified using the Powers-Yun decomposition technique for hazard functions (the Blinder-Oaxaca algorithm extended for time-to-event data). This approach identifies two effects per disease: exposure (the disparity is generated because of the difference in the prevalence of a disease in two compared populations) and vulnerability (the disparity is generated because of the difference in mortality in the subpopulation with the disease). Two population groups (80-) and (80+) separated by age 80 were analyzed. Disease-specific contributions explained 44% and 63% of the disparities in all-cause mortality in the 80- and 80+ age groups, respectively. Exposure effects dominated in the 80- group (47.8% vs. 3.5% of vulnerability effect), while vulnerability was the main contributor (40.1% vs. 23.3% of exposure effect) in the 80+ group. Heart failure, influenza, and pneumonia explained 21.3% of the disparity in the 80- group, while the other circulatory system diseases, especially the vulnerability of cardiac dysrhythmias (5%), dominated the 80+ group. Additionally, cerebrovascular diseases (10.7%) and blood diseases (14.1%) are the other two major contributors to vulnerability in 80+. Reducing disparities in LE demands multiple fronts, such as tobacco control, heightened preventive measures against influenza and pneumonia, and enhanced access to arrhythmia treatments.

Abstract Ambient air contaminants have been associated with higher risk of Alzheimer’s disease (AD), related dementia, and cognitive impairment; however, mechanisms that underlay these effects remain unclear. Pre-existing arterial hypertension, diabetes mellitus, cerebrovascular disease, and renal disease have a strong impact on AD risk and disparities. These diseases, in turn, are associated with exposure to air contaminants, including particulate matter (PM2.5). We applied causal mediation analysis in the formulation of VanderWeele which provides causal estimates of the direct and indirect (through the effects on pre-existing diseases) effects of PM2.5 exposure on the risk of AD. We used Administrative Health Insurance Claims from 5%-Medicare (N=6,042,239; 1991-2020) to identify the date of onset for AD and risk related diseases such as arterial hypertension, diabetes, cerebrovascular, and kidney diseases that were considered mediators in these analyses. Information on ambient PM2.5 at the zip code level were derived from the Socioeconomic Data and Applications Center (SEDAC) that incorporates air monitoring data, satellite aerosol optical depth, meteorological conditions, chemical transport model simulations, and land-use variables. The estimates supported the hypotheses on the linear dependence of the dose-response function with higher effects for Males vs. Females, White vs. Black subpopulations, and stroke-belt states vs. other U.S. regions. We found that the contributions of indirect effects were stronger for younger older adults and reached up to 13% for cerebrovascular disease, 10% for arterial hypertension, 5% for diabetes, and 4% for chronic kidney disease. Analysis in subpopulations demonstrated health disparities in the effects of PM2.5 on AD risk.

OBJECTIVES:Health forecasting is an important aspect of ensuring that the health system can effectively respond to the changing epidemiological environment. Common models for forecasting Alzheimer's disease and related dementias (AD/ADRD) are based on simplifying methodological assumptions, applied to limited population subgroups, or do not allow analysis of medical interventions. This study uses 5 %-Medicare data (1991-2017) to identify, partition, and forecast age-adjusted prevalence and incidence-based mortality of AD as well as their causal components. METHODS:The core underlying methodology is the partitioning analysis that calculates the relative impact each component has on the overall trend as well as intertemporal changes in the strength and direction of these impacts. B-spline functions estimated for all parameters of partitioning models represent the basis for projections of these parameters in future. RESULTS:Prevalence of AD is predicted to be stable between 2017 and 2028 primarily due to a decline in the prevalence of pre-AD-diagnosis stroke. Mortality, on the other hand, is predicted to increase. In all cases the resulting patterns come from a trade-off of two disadvantageous processes: increased incidence and disimproved survival. Analysis of health interventions demonstrates that the projected burden of AD differs significantly and leads to alternative policy implications. DISCUSSION:We developed a forecasting model of AD/ADRD risks that involves rigorous mathematical models and incorporation of the dynamics of important determinative risk factors for AD/ADRD risk. The applications of such models for analyses of interventions would allow for predicting future burden of AD/ADRD conditional on a specific treatment regime.

Background: Alzheimer’s disease (AD) and related dementia (ADRD) risk is affected by multiple dependent risk factors; however, there is no consensus about their relative impact in the development of these disorders. Objective: To rank the effects of potentially dependent risk factors and identify an optimal parsimonious set of measures for predicting AD/ADRD risk from a larger pool of potentially correlated predictors. Methods: We used diagnosis record, survey, and genetic data from the Health and Retirement Study to assess the relative predictive strength of AD/ADRD risk factors spanning several domains: comorbidities, demographics/socioeconomics, health-related behavior, genetics, and environmental exposure. A modified stepwise-AIC-best-subset blanket algorithm was then used to select an optimal set of predictors. Results: The final predictive model was reduced to 10 features for AD and 19 for ADRD; concordance statistics were about 0.85 for one-year and 0.70 for ten-year follow-up. Depression, arterial hypertension, traumatic brain injury, cerebrovascular diseases, and the APOE4 proxy SNP rs769449 had the strongest individual associations with AD/ADRD risk. AD/ADRD risk-related co-morbidities provide predictive power on par with key genetic vulnerabilities. Conclusion: Results confirm the consensus that circulatory diseases are the main comorbidities associated with AD/ADRD risk and show that clinical diagnosis records outperform comparable self-reported measures in predicting AD/ADRD risk. Model construction algorithms combined with modern data allows researchers to conserve power (especially in the study of disparities where disadvantaged groups are often grossly underrepresented) while accounting for a high proportion of AD/ADRD-risk-related population heterogeneity stemming from multiple domains.

Topics discussed at the "Leveraging Existing Data and Analytic Methods for Health Disparities Research Related to Aging and Alzheimer's Disease and Related Dementias" workshop, held by Duke University and the Alzheimer's Association with support from the National Institute on Aging, are summarized. Ways in which existing data resources paired with innovative applications of both novel and well-known methodologies can be used to identify the effects of multi-level societal, community, and individual determinants of race/ethnicity, sex, and geography-related health disparities in Alzheimer's disease and related dementia are proposed. Current literature on the population analyses of these health disparities is summarized with a focus on identifying existing gaps in knowledge, and ways to mitigate these gaps using data/method combinations are discussed at the workshop. Substantive and methodological directions of future research capable of advancing health disparities research related to aging are formulated.

Abstract Evidence is accumulating that individuals with cancer diagnoses exhibit Alzheimer’s disease (AD) and related dementia (ADRD) risk profiles that differ from the general population of U.S. older adults. In this study we used SEER-Medicare data to compare the relative risk of AD/ADRD between individuals with slow-progressive cancers and the non-cancer general population. The study cohort included individuals age 65+ (N=2,023,054) with a primary diagnosis (1999-2017) of one of nine slow progressive cancers (breast, colorectal, prostate, uterine, kidney, ovarian, and urinary bladder cancers, as well as lymphomas and melanoma) and no clinical record of AD/ADRD prior to cancer diagnosis. This cohort was then matched by age to a comparable non-cancer population (N=1,142,641). The hazard ratios of AD/ADRD for each cancer compared to the non-cancer cohort were evaluated individually in 29 age-specific groups for each cancer type. All cancers had similar patterns of dependence for post-cancer AD/ADRD risks. We found that the presence of cancer was associated with higher risk of AD/ADRD at age at diagnosis 65-75; the relative risks decline with age at diagnoses becoming protective at advanced ages. Furthermore, for any given age at diagnosis the relative risk of AD/ADRD (i.e., cancer vs. non-cancer) also declines with time. Detailed discussion of possible causes of these effects including cancer treatment, genetic variation, possible trade-off effects, common risk and protective factors, possibly lower administration and adherence of AD/ADRD diagnostic procedures for individuals with cancer, and the roles of competing risks (first of all due to death cases) is presented.