Cognition and Wealth Changes in Mid-to-later Life: A Latent Class Trajectories Approach Using the Health and Retirement Study

Abstract

Objectives

To assess how cognitive trajectories from mid-to-later life relate to wealth change, overall and by mid-life income. Methods: Data were from participants (51–64 years) in the 2000–2018 U.S. Health and Retirement Study who were cognitively healthy at baseline (year 2000; unweighted n = 3821). Longitudinal latent class analyses generated cognitive and wealth trajectories, independently, and multinomial logistic regressions estimated the association between cognitive trajectories and wealth profiles, overall and by median income. Results: We identified three cognitive: cognitively healthy (CH), increasing cognitive impairment (ICI), and increasing dementia (ID) and four wealth profiles: stable wealth loss (SWL), delayed gradual wealth loss (DGWL), stable wealth gain (SWG), and gradual wealth gain (GWG). The ID group had higher probability of being in the SWL group and lower probability of SWG, which was more pronounced in respondents with greater median income. Discussion: Individuals with ID may be vulnerable to wealth loss, particularly for middle-class households.

Keywords

cognition wealth socioeconomic status income aging

Introduction

In the United States (US), over 46 million Americans are 65 years or older, a number that is increasing rapidly and projected to nearly double by 2050 (Jacobsen, 2011). Economic insecurity, or when an individual lacks sufficient financial resources to cover necessary expenses (Mutchler & Li, 2021), affects over 15 million Americans over the age of 65 (Cubanski et al., 2018). As older adults age and retire, they face increased responsibility for managing both financial and health needs. These intersecting needs can make older adults more vulnerable for adverse financial and health outcomes. Socioeconomic indicators, such as wealth, are important for healthy aging (Smith & Kington, 1997; Steptoe & Zaninotto, 2020), both by directly providing financial resources but also indirectly through access to education, social status, and healthy behaviors (Braveman & Gottlieb, 2014). Higher socioeconomic status including wealth has been associated with lower risk of physical conditions including heart disease, diabetes, cancer, stroke, and cognitive impairment (Adler & Stewart, 2010; Fernández-Blázquez et al., 2021), psychiatric conditions including depression and anxiety (Fiske et al., 2009), and increased access to healthcare (McMaughan et al., 2020). Thus, maintaining financial resources into older age is an important factor to promote healthy aging. However, there is a growing body of evidence that suggests that an individuals’ cognitive status can have a significant influence on the financial outcomes of older adults (Angrisani & Lee, 2019; Pachana et al., 2014).

Unhealthy aging can negatively influence the productive capacity of individuals (e.g., income and employment), reduces quality of life, increases the likelihood of incurring medical expenses, and accelerates wealth depletion starting in midlife and into older age (Mira et al., 2022). Older adults may be more vulnerable to adverse financial outcomes due to cognitive impairment and decline (Pachana et al., 2014). Previous research has found that having better cognitive skills is associated with better lifetime savings, stock market participation, and portfolio diversification (Christelis et al., 2009; Korniotis & Kumar, 2011; Smith et al., 2010). Recently, Angrisani and Lee (2019) found wealth reductions among households whose financial decision maker was experiencing cognitive decline. Indeed, financial instability, such as inconsistent bill payments and risky financial decisions, have been recognized as potential early signs of dementia (Boyle et al., 2012; Okonkwo et al., 2008; Spreng et al., 2016; Sudo & Laks, 2017; Widera et al., 2011). Nicholas et al. (2021) found that a diagnosis of dementia was associated with higher rates of missed payments and subprime credit scores years prior to a diagnosis. These studies suggest that cognitive impairment and decline, even prior to a dementia diagnosis, can affect older adults’ financial well-being. However, to our knowledge, few studies have examined the phenotypical trajectories of cognition and wealth starting in mid-life and extending into later-life using a person-centered approach which allows researchers to identify higher-risk individuals by characterizing prototypical trajectories within a given population.

Middle-age income may also modify the associations between cognitive aging and wealth depletion. It is possible that individuals with lower income in middle-age are more vulnerable to adverse health events which can increase the likelihood of wealth depletion. In contrast, higher income in middle-age may expose individuals to accelerated wealth depletion during aging until social safety net protections are activated. Ultimately, it is not fully understood how middle-age income could alter wealth trajectories (wealth accumulation or loss) as individuals’ transition to worse cognitive status. Furthermore, racism has been associated with both socioeconomic status as well as later-life cognition, and as such it is plausible that the mechanisms that link cognitive aging and economic and financial well-being differ between racial and ethnic groups among whom experiences of racism may be more prevalent (Beech et al., 2021; Zahodne et al., 2019, 2021). This is particularly true with wealth as recent estimates show the median net wealth of white families in the US to be much higher than that of Black families (Bhutta et al., 2020). This emphasizes the need to examine influences of race/ethnicity when considering wealth and cognitive trajectories.

Conceptual Framework

The current study draws on concepts from established theoretical frameworks including vulnerable populations, life course theory, and borrows from Andersen’s behavioral model as well as more modern analyses positing how wealth is an important social determinant of health (Woolf et al., 2015). The vulnerable population theory proposes that while everyone is potentially vulnerable to adverse outcomes, this risk is greater for those from populations with fewer social and economic resources (Aday, 1994). A life course approach to understanding later life health requires consideration of the social, physical, and environmental factors at different life stages as well as exposure accumulation over time. This underscores our study’s focus on middle-aged adults transitioning into older adulthood over time. Life course socioeconomic status has been independently linked to both higher risk of cognitive impairment and dementias (Elbejjani et al., 2017; Marden et al., 2017; Staff et al., 2018) as well as more pronounced economic burdens in later-life (Cagney & Lauderdale, 2002; Kuh et al., 2004; Marden et al., 2017). Finally, Andersen’s behavioral model proposes a typology of factors including predisposing (e.g., racism, sexism), enabling (e.g., insurance), and health need factors (e.g., number of comorbid conditions) which can help to explicate the complex pathways of health and illness, including cognitive health. Differences in predisposing, enabling, and health need factors can confer disparities in cognitive health as well as socioeconomic status. In addition to these theories, substantial scientific evidence supports a socioeconomic (SES) health gradient, and several studies have explicitly shown that income and wealth are important social determinants of better health, lower disability, and decreased risk for premature mortality and lifespan longevity (Jackson & Engelman, 2022; Makaroun et al., 2017; Pool et al., 2018; Weida et al., 2020; Woolf et al., 2015). Some of this existing work has proposed life course policy recommendations at the individual, family, and geographic levels that could alleviate the established SES-Health gradient (Woolf et al., 2015). Critically, however, the ascending prevalence of Alzheimer’s Disease and Related Dementias (ADRD) over the past few decades and the expected increase in number of ADRD cases as the population continues to age through mid-century, has pushed gero-researchers to consider the role of cognitive health in influencing financial capacity, well-being, and vulnerability (Widera et al., 2011). As such, while our analyses are rooted in the theoretical frameworks detailed above, we argue in line with these emerging voices, that it is both plausible and necessary to examine how loss of wealth can also accompany and even be accelerated in groups that have higher risk for cognitive impairment. Within this vein of reasoning, we aim to examine how cognitive vulnerability (measured using longitudinal impairment classifications) is linked to wealth gain and loss trajectories and provide empirical evidence that the SES-Health (in our case, wealth-cognitive impairment) relationship is directionally and temporally complex. This contextualization guided the formation of our research questions and our hypotheses, allowing for a more comprehensive consideration of cognitive and wealth phenotypes of older adults while accounting for differential effects by mid-life income.

Current Study

Despite the higher prevalence of cognitive impairment and decline in later-life and the importance of wealth for healthy aging, the links between cognitive health and wealth from mid-to-late life are not well characterized. Furthermore, little empirical research has determined if midlife income modifies wealth phenotypes as individuals’ transition to more severe cognitive impairment. In the present study, we aimed to identify cognitive and wealth phenotypes from mid-to-late life using latent class analysis in a population-representative sample of the Health and Retirement Study from 2000 to 2018 to better understand the relationship between cognitive status and wealth outcomes in older adults. The purpose of this study was to 1) identify latent phenotypes of cognition and wealth from mid-to-later life, 2) determine if latent cognitive phenotypes are associated with latent wealth phenotypes, and 3) assess whether this relationship differs by mid-life median household income. We hypothesize that the transition to higher severity of cognitive impairment (e.g., dementia) will be associated with more pronounced wealth depletion and that wealth depletion will be more prevalent in individuals with higher midlife median income.

Methods

Data Source

We used biennial longitudinal data from 10 waves of the U.S. Health and Retirement Study (HRS) covering the years 2000 (baseline) to 2018 (end wave) for a total of ten waves of data (Sonnega et al., 2014). The HRS is a nationally representative longitudinal study of middle-aged and older adults (51 years and older) in the United States. We focused on middle-aged individuals ages 51–64 in the year 2000 (study baseline), who do not meet research criteria for “dementia” based on the Langa–Weir classification (see cognitive status section below), and have survived over the observation period. Our unweighted analytic sample size was n = 3821 respondents, the weighted equivalent of 15,909,142 middle-aged U.S. adults.

Measures

Cognitive Status

Cognitive status at each biennial study wave (2000–2018) was assessed using the Langa–Weir (LW) algorithm and classified as: “cognitively normal,” “cognitively impaired not dementia,” and “dementia” (Crimmins et al., 2011; Langa et al., 2008; Plassman et al., 2007); assumed to be ordered categorically in the L-LCA models. Briefly, for self-reporting participants, LW classifies respondents relative to thresholds of a modified version of the Telephone Interview for Cognitive Status (TICS-M; range = 0–27)—which consisted of immediate and delayed 10-word free recall and serial 7’s subtraction tests to assess memory and backward count to evaluate attention and processing speed. For proxy respondents, where no direct assessment is available, LW used proxy reports on respondents’ instrumental activities of daily living (range 0–5), memory function (0 = excellent – 4 = poor), and an interviewer report of perceived respondents’ cognitive status (0 = no cognitive impairment, 1 = possible impairment, and 2 = impaired) to generate a proxy-based index (range 0–11) where higher values indicated higher cognitive impairment. For self-reports, TICS-M were used to determine the three cognitive status categories as follows: normal cognition (12–27), cognitive impairment non-dementia (CIND) (7–11), or dementia (0–6). Similarly, proxied participants were classified as: normal cognition (0–2), CIND (3–5), or dementia (6–11). Cut points for both self-reports and proxy responses were set and validated relative to detailed diagnostic assessment derived from the ADAMS, a clinical (including neurological and neuropsychological) study of a sub-sample of HRS participants and have been used in the literature (Crimmins et al., 2011; Langa et al., 2008; Plassman et al., 2007; Rocca et al., 2011).

Wealth

We used wave-specific measures of the net value of wealth (including second home) calculated as the sum of all wealth components (e.g., value of real estate, vehicles, businesses, IRA and Keogh accounts, stocks, mutual funds, and investment trusts, value of checking, savings, or money market accounts, net value of bonds and bond funds, and net value of all other savings) less all debt [for example, value of all mortgages (primary and secondary residences), value of other home loans, and any other debt]. Wealth was reported in nominal dollars and adjusted for inflation using a GDP price index as determined by the Bureau of Economic Analysis, National Income Product Accounts Tables (Dunn et al., 2018). For each biennial wave following the baseline, we created a dichotomous measure to indicate whether the respondent lost or gained (or stayed at the same level) wealth relative to their baseline floor (wealth value in year 2000). Comparing wealth relative to the baseline median allowed us to focus on variations relative to a time-fixed population-specific estimate—in line with the approach that non-institutional, individual, investors use when establishing long-term goals regarding personal wealth building strategies; in our case, lasting for the length of the observation period of the study.

Midlife Income

Midlife income was measured as total household income reported in nominal baseline dollars. This was further dichotomized as either having income below or above the median, for testing income stratified models (see below). Given the focus of this study, the median value was set relative to the target population of middle-aged adults (51–64 years) not satisfying Langa–Weir dementia criteria at baseline.

Covariates

We accounted for several potential confounding variables, all measured at the study baseline, including: age (in years), sex/gender (male; female), educational attainment (in years), race/ethnicity (White, Black, Hispanic, and Other), insurance status (government only, government and private, private only, and no insurance), number of comorbid conditions (high blood pressure, diabetes, cancer, lung disease, heart disease, stroke, psychiatric problems, and arthritis), and United States region (Northeast, Midwest, South, West).

Analytic Approach

First, we used longitudinal latent class analyses (L-LCAs) to extract data-driven phenotypes of cognitive and wealth using the cognitive status classification and binary wealth variables described above, independently. Latent class modeling was done using standard fitting techniques (Collins & Lanza, 2009; Finch & Bronk, 2011; Nylund et al., 2007). For each outcome, we sequentially fit a series of models iterating over potential 2–6 class solutions. For each model, we estimated both absolute and relative fit indices, as recommended in the literature (Masyn, 2013; Nylund et al., 2007). Briefly, L-LCAs assume the presence of a mixture of distributions reflecting underlying heterogeneities in a target population and produce probabilistic person-specific classifications that clusters individuals into more homogenous groups (phenotypes) given similarities in their input data patterns (cognitive status and wealth in our case). All models were estimated using Maximum Likelihood with Robust Standard Errors. We used both statistical fit assessment (Akaike Information Criterion (AIC), Bayesian Information Criterion (BIC), and Entropy as well as the number (%) of participants meeting criteria for each derived phenotype) and substantive interpretation to determine the acceptable number of phenotypes for each considered indicator, independently. Fit statistics derived from the L-LCAs are presented in Supplemental Tables 1 and plotted in Supplemental Figure 1. Following selection of the optimal class solution, labels for the underlying grouping were assigned based on co-authors consensus as supported by expectations derived from existing literature. To characterize and enhance interpretability of the derived phenotypes, we used the wave-specific manifest (measured indicators) variables for cognitive status and wealth to estimate and plot trajectories depicting change in LW-dementia and wealth loss/gain rates across the phenotypes over the study time-period in Figure 1. Estimates of mean and median changes in wealth over the observation period across the derived wealth classifications are provided in Supplemental Figure 2. Second, we used descriptive statistics to characterize the target population across cognitive phenotypes relative to the covariates of interest. We used chi-squared tests to assess statistical significance for differences in categorical indicators and F-tests for continuous measures; both adjusted for the HRS sampling design. Results are presented in Table 1.

Figure 1.

Fit statistics and (a) cognitive and (b) wealth latent classifications.

Table 1.

Baseline Characteristics, Overall and by Cognitive Latent Classes, the Health and Retirement Study, 2000, Unweighted n = 3821 (Weighted n = 15,909,142).

Unweighted N	Cognitive status Classification			Total	p-value
	Increasing Dementia	Increasing Cognitive Impairment	Cognitively Healthy
	N = 219	N = 865	N = 2880
Age, in years (mean, SD)	58.4 (3.9)	58.1 (3.3)	57.0 (3.0)	57.6 (3.3)	<0.001
Female, % (SE)	56.4 (3.5)	54.7 (2.1)	55.1 (1.0)	55.1 (0.8)	p = .91
Race/Ethnicity, % (SE)
White	42.6 (5.2)	61.6 (3.2)	87.1 (1.1)	80.4 (1.5)	<0.001
Black	35.6 (4.2)	17.2 (1.8)	5.9 (0.7)	9.3 (0.7)
Hispanic	18.6 (4.2)	17.1 (3.7)	5.0 (0.8)	7.9 (1.3)
Other	3.2 (1.9)	4.1 (1.1)	2.0 (0.3)	2.4 (0.4)
Education, in years (mean, SD)	9.1 (3.8)	11.5 (3.0)	13.9 (2.1)	13.2 (2.6)	<0.001
United States region, % (SE)
Northeast	13.0 (3.1)	14.9 (2.1)	17.5 (1.7)	16.8 (1.6)	p = .04
Midwest	18.8 (4.5)	22.1 (2.3)	25.8 (2.0)	24.8 (1.8)
South	53.0 (4.8)	42.4 (3.2)	36.3 (1.9)	38.2 (1.7)
West	15.1 (3.6)	20.6 (2.8)	20.3 (2.2)	20.2 (2.1)
Insurance, % (SE)
Government only	28.8 (4.1)	10.5 (1.4)	3.5 (0.4)	5.9 (0.5)	<0.001
Government and private	4.9 (2.2)	4.9 (1.0)	3.5 (0.7)	3.9 (0.6)
Private only	49.2 (4.6)	70.7 (2.0)	87.4 (1.0)	82.6 (0.9)
No insurance	17.1 (2.9)	13.8 (1.2)	5.6 (0.6)	7.7 (0.6)
Baseline income, in 2000$US (mean, SD)	29229 (54824)	53973 (72397)	95344 (105184)	84676 (101371)	<0.001
Baseline wealth (includes 2nd home) in 2000$US (mean, SD)	261387 (1997461)	269189.5 (960164.5)	503504 (1175052)	448456 (1184752)	<0.001
Baseline income, in 2018$US (mean, SD)	41334 (77529)	763 25 (102379)	134831 (148746)	119744 (143354)	<0.001
Baseline wealth (includes 2nd home) in 2018$US (mean, SD)	369640 (2824701)	380673 (1357813)	712028 (1661694)	634182 (1675412)	<0.001
Number of health conditions, (mean, SD)	1.5 (1.3)	1.3 (1.1)	1.0 (0.9)	1.1 (0.9)	<0.001

Third, we used multinomial logistic regression to test the association between our latent cognitive and the latent wealth phenotypes. We sequentially fit five models and incrementally adjusted for covariates. Model one was unadjusted to provide a crude estimate of the association between the cognitive and wealth phenotypes. Model two adjusted for age and gender. Model 3 further adjusted for education. Model 4 additionally included race/ethnicity. Finally, model 5 adjusted for insurance status, US region, and number of comorbid health conditions in addition to all previous covariates. We subsequently plotted the average marginal probabilities of wealth phenotype classifications and their 95% confidence interval, for each of the cognitive phenotypes as derived from the sequentially adjusted models above (Figure 2). We present the relative risk ratios (RRRs) and their 95% confidence intervals in Table 2. Since we treat income as a potential modifier and conduct income stratified analyses (see below), we did not adjust for it in the primary models. However, secondary models that adjusted for baseline income quartiles are included in Supplemental Table 2, and plots of marginal estimates are included in Supplemental Figure 3.

Figure 2.

Probability of being in a wealth latent classification given cognitive latent classification.

Table 2.

Multinomial Logistic Regression Models for the Association Between Latent Cognitive Classifications and Latent Wealth Classifications.

	Model 1	Model 2	Model 3	Model 4	Model 5
	RRR [95% CI]	RRR [95% CI]	RRR [95% CI]	RRR [95% CI]	RRR [95% CI]
Stable wealth loss versus Stable wealth gain (reference)
Cognitively healthy (reference)	ref	ref	ref	ref	ref
Increasing cognitive impairment	1.39 [1.07; 1.81]	1.31 [1.00; 1.71]	1.12 [0.85; 1.47]	1.13 [0.84; 1.52]	1.09 [0.82; 1.46]
Increasing dementia	2.41 [1.51; 3.84]	2.22 [1.37; 3.58]	1.63 [0.96; 2.76]	1.59 [0.91; 2.80]	1.57 [0.86; 2.87]
Delayed gradual wealth loss versus Stable wealth gain (reference)
Cognitively healthy (reference)	ref	ref	ref	ref	ref
Increasing cognitive impairment	1.71 [1.29; 2.27]	1.61 [1.21; 2.15]	1.40 [1.03; 1.90]	1.31 [0.93; 1.83]	1.30 [0.92; 1.82]
Increasing dementia	1.51 [0.88; 2.57]	1.38 [0.80; 2.41]	1.05 [0.61; 1.82]	0.91 [0.53; 1.57]	0.90 [0.50; 1.62]
Gradual wealth gain versus Stable wealth gain (reference)
Cognitively healthy (reference)	ref	ref	ref	ref	ref
Increasing cognitive impairment	1.36 [1.04; 1.77]	1.35 [1.04; 1.76]	1.16 [0.86; 1.57]	1.11 [0.83; 1.49]	1.07 [0.80; 1.44]
Increasing dementia	1.33 [0.72; 2.46]	1.32 [0.72; 2.43]	0.99 [0.55; 1.76]	0.88 [0.50; 1.58]	0.83 [0.47; 1.44]

Model 1 is the unadjusted model; model 2 adjusts for gender and age; model 3 adjusts further for education; model 4 adjusts for race/ethnicity; model 5 adjusts further for insurance status, number of health conditions, and United States region.

Fourth, to determine whether the association between latent cognition and wealth phenotypes differs by mid-life median income, we refit the above-described multinomial logistic regression models in two sub-samples based on stratification by baseline median income (either having an income below or at or above the median income of the target population of middle-aged adults (51–64 years) not meeting Langa–Weir dementia criteria). As with step 3 above, we estimated and plotted the derived average marginal probabilities for each median income strata in Figure 3 and present the fully adjusted RRR and their 95% CIs in Table 3.

Figure 3.

Probability of being in a wealth latent classification given cognitive latent classification according to median income at baseline.

Table 3.

Median Income Stratified Multinomial Logistic Regression Models.

	Less than baseline median income	Greater than or equal to baseline median income
	RRR [95% CI]	RRR [95% CI]
Stable wealth loss versus Stable wealth gain (reference)
Cognitively healthy (reference)	ref	ref
Increasing cognitive impairment	0.97 [0.66; 1.40]	1.26 [0.87; 1.82]
Increasing dementia	1.47 [0.76; 2.86]	3.32 [0.88; 12.56]
Delayed gradual wealth loss versus Stable wealth gain (reference)
Cognitively healthy (reference)	ref	ref
Increasing cognitive impairment	1.35 [0.88; 2.05]	1.29 [0.80; 2.08]
Increasing dementia	0.94 [0.47; 1.89]	1.58 [0.35; 7.07]
Gradual wealth gain versus Stable wealth gain (reference)
Cognitively healthy (reference)	ref	ref
Increasing cognitive impairment	1.25 [0.88; 1.78]	0.77 [0.44; 1.37]
Increasing dementia	0.83 [0.44; 1.58]	1.81 [0.43; 7.69]

Both models are fully adjusted models and include gender, age, education, race/ethnicity, insurance status, number of health conditions, and United States region.

Finally, we conducted two sets of sensitivity analyses to ensure the robustness of our findings. In the first sensitivity analysis, we adjusted for baseline wealth and cognitive function. Baseline wealth was measured in quartiles to reduce the potential effects of skewness and imprecision in self-reporting. Baseline cognitive function was assessed using the TICS-M and treated as a continuous measure (range 0–27). In our second set of sensitivity analyses, we considered an alternative operationalization of our wealth outcome; namely, change in reported wealth over the observation period measured as reported Wealth at the 2018 wave—Wealth at the 2000 wave (both measured in $2018 dollars). As with our first sensitivity analysis, we group the change in wealth into four quartiles to reduce skewness and imprecision in reporting. We use the same set of covariates as with the first sensitivity and model the wealth change quartiles using survey design adjusted ordinal logistic regression. Results from these sensitivity analyses are presented in Supplemental Tables 3(a, b) and 4 (a, b).

Results

Latent Phenotypes of Cognitive Status and Wealth Trajectories

We identified three distinct cognitive classifications, and four distinct wealth classes based on fit statistics and interpretability (Figure 1(a) and (b), respectively). The three cognitive classes were labeled: cognitively healthy (CH; unweighted n (un) = 2,880, 76.6%), increasing cognitive impairment (ICI; un = 865, 19.3%), and increasing dementia (ID; un = 219, 4.1%) (Figure 1(a)). The four wealth classes were: stable wealth loss (SWL; un = 1,043, 24.7%), delayed gradual wealth loss (DGWL; un = 735, 17.3%), stable wealth gain (SWG; un = 1,635, 43.9%), and gradual wealth gain (GWG; un = 551, 14.2%) (Figure 1(b)). For the cognitive classes, the cognitively healthy class was characterized by a high probability of being in the cognitive normal group throughout the time period (Figure 1(a)). The increasing cognitive impairment class was characterized by a stable increase in cognitive impairment with a decrease in cognitive normal status and very little increase in dementia (Figure 1(a)). The increasing dementia class was characterized with a higher probability of being cognitively impaired, a steady increasing probability of dementia, and a steady decrease in being cognitively healthy (Figure 1(a)).

For the wealth latent classes, the stable wealth gain class was characterized by increasing wealth over the time period compared to the baseline wealth (Figure 1(b); also Supplemental Figure 2). The stable wealth loss class was characterized by having lower wealth compared to wealth at baseline during the whole time period (Figure 1(b); also Supplemental Figure 2). The gradual wealth loss class was characterized by lower wealth compared to baseline in the first ten years with wealth increasing after ten years (Figure 1(b); also Supplemental Figure 2). Finally, the delayed gradual wealth loss class was characterized by higher wealth at the beginning of the time period that gradually declined over the time period compared to the baseline wealth (Figure 1(b); also Supplemental Figure 2).

Distribution of the Latent Phenotypes of Trajectories of Cognitive Status

In our target population (unweighted n = 3821; weighted equivalent of 15,909,142 middle-aged adults), the majority were female (55.1%) and non-Hispanic White (80.4%) (Table 1). The mean age at baseline was 57.6 years (SD: 3.3) and the mean years of education was 12.9 (SD: 2.7). The mean, non-inflation adjusted, income and wealth at baseline were $84,676 (SD:$101,371) and $448,456 (SD:$1,184,752), respectively (the median estimates for income and wealth were $57,476 and $186,500), and the majority had private insurance at baseline (82.6%). Of the participants in the increasing dementia class, 35.6% were Black and 18.6% were Hispanic compared to 5.9% and 5.0% of the cognitively healthy class participants being Black and Hispanic, respectively (Table 1). The increasing dementia class had less years of education (9.1 years vs. 13.9 years), higher government only or uninsurance rates, lower average income (mean $29,229 vs. $95,344) and wealth at the baseline, and higher count of comorbid conditions compared to the cognitively healthy class (Table 1). The increasing cognitive impairment class also had higher prevalence of Black (17.2%) and Hispanic (17.1%) participants, lower educational attainment (11.5 years), lower income and wealth accumulation, and higher average number of comorbid conditions (Table 1).

Associations Between Cognitive Status Phenotypes and Latent Trajectories of Wealth

In multinomial regression analyses, we found that, adjusting for age and gender, those in the increasing dementia class (vs. cognitively healthy) were more likely to satisfy criteria for classification in the stable wealth loss class (RRR: 2.2 [95% CI: 1.4, 3.6]; Table 2; Model 2) versus stable wealth gain. The increased relative risk ratios were attenuated and statistically explained by adjustment for education (RRR: 1.6 [95% CI: 0.9, 2.8]; Table 2; Models 3). Additional adjustment for race/ethnicity (Model 4), insurance, region, and comorbid conditions (Model 5) did not have any additional effects on the estimated relative risk ratios. The increasing dementia classification (vs. cognitively healthy) did not relay any additional risk for classification in the delayed gradual wealth loss or the gradual wealth gain classes relative to stable wealth gain.

In age and gender adjusted models, individuals in the increasing cognitive impairment group had higher risk for delayed gradual wealth loss and higher likelihood of gradual wealth gain (vs. stable wealth gain) (RRR: 1.6 [95% CI: 1.2, 2.1] and RRR: 1.1 [95% CI: 0.8, 1.4], respectively; Table 2; Model 2). Adjustment for education explained the difference in RRR for gradual wealth gain but not gradual wealth loss in this group (Table 2; Model 3). Additional adjustment for race/ethnicity explained the difference in RRR for classification in gradual wealth loss (Table 2; Model 4). Further adjustments to insurance and comorbidities did not influence the estimated RRR (Table 2; Model 5). Figure 2 visualizes the probability of being in each wealth classification relative to the cognitive classification exposure. As with above, Figure 2 showcases the slightly higher probability of being in the stable wealth loss and delayed gradual wealth loss class for individuals satisfying criteria for increasing dementia and increasing cognitive impairment, respectively, and the higher probability of being in the stable wealth gain for those classified as cognitively healthy (Figure 2).

Stratification by Baseline Median Income

We found largely similar results when stratifying by median baseline income, with the RRRs being more pronounced in the higher income group (Table 3). Though not significant, individuals in the increasing dementia group had higher risk of stable wealth loss in fully adjusted models (Table 3). This association was stronger in participants who had greater than or equal to baseline median income (RRR: 3.3 [95% CI: 0.9, 12.6]) than for participants with less than the median income at baseline (RRR: 1.5 [95% CI: 0.8, 2.9]) (Table 3; Figure 3).

Sensitivity Analyses

We conducted two sets of sensitivity analyses. First, we adjusted for two additional covariates to account for baseline wealth and cognitive function. The findings from these models show that our reported primary results are qualitatively and quantitatively unchanged after these adjustments (Supplemental Table 3(a, b)). In our second set of sensitivity, we examined the change in reported wealth over the observation period (2018 wave–2000 wave). The findings from these models show that our reported primary results are qualitatively unchanged; individuals meeting criteria for increasing cognitive impairment (OR = 0.78 [0.67; 0.92]) and increasing dementia (OR = 0.61 [0.47; 0.80]) were less likely to be in higher quartiles of wealth gain over the observation period and that group differences are explained by educational attainment (Supplemental Table 4(a, b)). As such, even under this different operationalization of wealth change, that assumes a unique distribution of wealth change, the associations between the cognitive impairment trajectories and wealth gain/loss are consistent.

Discussion

In this nationally representative longitudinal cohort of middle-aged adults in the United States, we examined the association between cognitive impairment and wealth phenotypes. Our study is among the first to estimate the prevalence of heterogenous phenotypical wealth trajectories in aging adults starting in midlife and our results point to several findings. First, we found three distinct patterns of cognitive aging phenotypes among middle-aged individuals and four patterns of wealth phenotype trajectories. Both the increasing dementia and increasing cognitive impairment phenotypes had a more pronounced risk for stable wealth loss while the increasing cognitive impairment phenotype also had an increased likelihood of having delayed gradual wealth loss. However, these increased likelihoods were largely but differentially explained by further adjustment of predisposing, enabling, and need factors including education, and race/ethnicity. Finally, we found that the associations between cognitive impairment and stable wealth loss were more evident in individuals nested within households with median income at and above the baseline income.

Cognitive Aging and Wealth

Wealth depletion during aging is expected as individuals retire and spend down accumulated resources. However, our results suggest that stable wealth loss is more likely to occur among individuals who have similar age profiles at baseline but higher cognitive health needs. This is similar to previous research that has found that unhealthy cognitive aging increases the likelihood of financial vulnerability and insecurity (Angrisani & Lee, 2019; Widera et al., 2011). While our data and models prohibit inferring direct causes for wealth depletion due to the risk and severity of cognitive impairment, there are several proposed mechanisms by which cognitive impairment could influence trajectories of wealth. A primary pathway to lower accumulated wealth in cognitive aging is concomitant physical impairment and disabilities which increases the risk of job loss, increases the need for long-term care and costly medical treatment, and can lead to poor financial decision-making and overspending thus impairing an individual’s ability to manage and maintain wealth (Schimmel & Stapleton, 2012). Furthermore, cognitive aging is linked to decreased affective skills (e.g., emotional dysregulations) (Steptoe & Wardle, 2017), a decline in social integration (Jung et al., 2022), and increased loneliness and social isolation (Clarke & Latham, 2014; Shankar et al., 2011). This lack of social and community support can lead to increased dependence on formal support services, such as long-term care (Gentili et al., 2022; Toot et al., 2017), and other health expenditures associated with “gradual onset health conditions” such as dementia which can induce spend down in assets and depletion in household accumulated wealth (Adams et al., 2003; Banks et al., 2019; Berkowitz & Qiu, 2006).

Another pathway for wealth depletion in older adults is financial exploitation, particularly among those with cognitive impairment. Although still nascent in measurement sophistication, several studies indicate that individuals at risk for dementia or with dementia are more vulnerable to financial exploitation (Fenton et al., 2022). Evidence suggests that 3.5–20% of adults over the age of 65 have been subject to financial fraud or exploitation with cognitive impairment and decline being significant risk factors for financial exploitation (Wood & Lichtenberg, 2017). Finally, cognitively impaired individuals being more prone to financial abuse by family members or caregivers is another mechanism for wealth depletion. Recent research indicates that financial abuse by family members, although varied in severity, is a highly prevalent and common form of elder abuse, and a significant contributor to wealth depletion in older adults (Weissberger et al., 2020).

The Role of Predisposing, Health Enabling, and Health Need Factors

Despite finding an increased likelihood of wealth depletion among those with increasing cognitive impairment and dementia, we also found that these relationships were largely explained by education and race/ethnicity. Predisposing factors (PF) including racism and sexism can differentially impact the risk for physical, mental, and cognitive health disadvantages throughout the life course and increases formal and informal health costs which can affect wealth accumulation and depletion. Our results show that controlling for race/ethnicity, a proxy for racism, attenuated the relative risk for gradual wealth loss among the increasingly cognitively impaired but had minimal explanatory effects among the severely impaired. These mixed results suggest that PFs’ influence on wealth loss and accumulation among individuals at risk for cognitive impairment and dementia is complex. Addressing structural factors, such as inadequate retirement savings and lack of financial literacy, may reduce the potential for unhealthy aging and buffer against wealth depletion in old age.

Health enabling (HE) factors, such as education, are conditions that facilitate or support the maintenance or improvement of an individual’s health (Aday & Andersen, 1974; Andersen, 1995). Our findings suggested that higher educational attainment explained differences in wealth depletion classification by cognitive status for those in the high severity group (i.e., increasing dementia). Education is important for both cognitive aging and wealth management. It is possible that individuals with higher education establish and maintain better financial situations prior to cognitive declines which could buffer the effect of cognitive impairment in later life. However, the influence of education on wealth accumulation and loss as individuals transition to worse cognitive impairment warrants more research.

Health need (HN) factors are demand conditions that influence the type and amount of use of healthcare services (treatment requirements including costly medication regiments, use of expensive medical procedures, and higher likelihood of placement in long-term care) (Andersen, 1995). Our results showed that HN, as measured by number of chronic conditions at baseline did not contribute to the associations between cognitive health and wealth accumulation or loss. This finding is inconsistent with previous research indicating that older adults with more comorbid conditions often experience more healthcare costs which could negatively impact wealth in later life. First, this inconsistency could be due to the younger age of the considered population (the baseline differences in mean count of chronic conditions between the cognitive phenotypes was relatively small). Second, we only accounted for baseline chronic conditions and as such incident conditions and increasing comorbidities could not be assessed in the context of our study. Third, this inconsistency could also be related to the type of chronic conditions included in this calculation as some chronic conditions are more severe and require higher drug costs and more medical interventions. Regardless, the identification of high-risk HN patient clusters can help devise strategies for better managed (through both formal and informal venues) and cost-effective care, insuring (either individually or through public resources) that households are protected against the potential fallout from costly care, and allocating resources to assist in funding healthcare costs that protects individuals and their families accelerated wealth depletion.

The Role of Middle-Age Income

Finally, we found that individuals with income above the baseline median were more prone to wealth loss due to cognitive impairment and dementia. There are several possibilities for this finding. First, individuals with lower income in midlife may not lose as much because there is not enough to lose. Additionally, lower income in midlife can increase the risk for onset of complex chronic health conditions, increasing the likelihood of incurring costly medical expenses and further inhibiting accumulation of assets and wealth. Indeed, previous work suggests that socioeconomic status and financial burdens increase the likelihood of incident dementia (Samuel et al., 2020), which could in turn lead to adverse financial outcomes in later-life. Further, midlife socioeconomic status is influenced by employment grade, impacting earning potentials which can blunt the potential for wealth accumulation. However, higher socioeconomic status in midlife exposes individuals to accelerated wealth depletion as individuals meeting these criteria are less likely to qualify for access or receipt of publicly funded resources of healthcare. Individuals with low socioeconomic status are eligible for social security and safety net programs (e.g., supplemental security income insurance as well as Medicaid for individuals meeting financial criteria), which provides support for older adults with reduced ability to work and can enable access to needed health and healthcare resources starting in late middle age and into older age following onset of cognitive dysfunction. Therefore, individuals with higher socioeconomic status in midlife, particularly those in what’s been traditionally known as the middle class, could be more susceptible to wealth loss as a result of spend down requirements before qualifying for any safety net programs that fund care utilization required for dementia care.

Limitations

Our study has limitations that are worth noting. First, since the goal of this analysis was to determine associations between cognitive and wealth trajectories, we did not aim to establish causality nor temporality. In our conceptual framework, we placed our arguments within the SES-health gradient and referenced the substantial literature documenting how wealth (a social determinant of health and a proxy for multiple health enabling factors) influences cognition and cognitive impairment. Yet, we also argued, in line with recent calls by gero-oriented research, that it is critical to examine how loss of wealth can also accompany and even be accelerated in groups that have higher cognitive impairment risk. We noted that the wealth-health (in our case, wealth-cognitive impairment) relationship might prove cross-causal and bidirectional. While testing such models are beyond the scope of this narrowly focused paper and might require an alternative analytic design, we hope that our findings provide incremental evidence that contribute to the scientific discussion on the complex links between SES factors and cognitive health. Second, the use of a binary threshold for wealth depletion trajectories limits our ability to understand smaller nuances in wealth changes across time. Further, we only used one measure of wealth and did not investigate other measures of socioeconomic status. An important future direction is to explore the influence of other measures of socioeconomic status including educational attainment. Second, we used one algorithm for classifying cognitive status. Future research should consider the robustness of these findings under alternative algorithms, and the use of clinical and administrative data which could provide better identification of cognitive status. Furthermore, the Langa–Weir algorithm to classify cognitive status has been shown to have differential psychometric properties across race/ethnic groups which limit our understanding of how these cognitive statuses could differ by race/ethnicity (Gianattasio et al., 2019). It was beyond the scope of this study to explore racial and ethnic disparities in cognitive and wealth phenotypes. This is a current focus of future research. There is also a need to better understand how intersecting disadvantages could influence cognition and wealth trajectories in later-life. Though we adjusted for various predisposing, health enabling, and health needs factors, future research should further explore how these factors can influence wealth accumulation and loss during aging, particularly among those with cognitive impairment. Fourth, we only accounted for time invariant factors at baseline and did not account for potential time-varying factors including change in health status and disability measures. These are important factors which should be examined in subsequent research. Finally, future research should consider potential mechanisms, including physical disability, financial fraud exploitation, and lack of social support, affect wealth accumulation and loss during cognitive aging.

Conclusion

Dementia and cognitive impairment increased vulnerability to sustained and delayed wealth loss, particularly for middle-class households. Our findings help identify risk groups for targetable interventions, and can potentially inform public health policies (e.g., long-term insurance), programming and infrastructure building for informal and formal care support, and development of resources for financial planning starting in middle-age.

Supplemental Material

Supplemental Material - Cognition and Wealth Changes in Mid-to-later Life: A Latent Class Trajectories Approach Using the Health and Retirement Study

Supplemental Material for Cognition and Wealth Changes in Mid-to-later Life: A Latent Class Trajectories Approach Using the Health and Retirement Study by Ashly C. Westrick, Darlingtina K. Esiaka, Helen C.S. Meier, Ronica N. Rooks, Mark Manning, and Wassim Tarraf in Journal of Aging and Health

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

This work was supported by the Michigan Center for Contextual Factors in Alzheimer’s Disease (MCCFAD) [P30 AG059300], funded by the National Institute on Aging of the National Institutes of Health. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

ORCID iDs

Ashly C. Westrick

Wassim Tarraf

Supplemental Material

Supplemental material for this article is available online

References

Adams

Hurd

M. D.

McFadden

Merrill

Ribeiro

(2003). Healthy, wealthy, and wise? Tests for direct causal paths between health and socioeconomic status. Journal of Econometrics, 112(1), 3–56. https://doi.org/10.1016/S0304-4076(02)00145-8

Aday

L. A.

(1994). Health status of vulnerable populations. Annual Review of Public Health, 15(1), 487–509. https://doi.org/10.1146/annurev.pu.15.050194.002415.

Aday

L. A.

Andersen

(1974). A framework for the study of access to medical care. Health Services Research, 9(3), 208–220.

Adler

N. E.

Stewart

(2010). Health disparities across the lifespan: Meaning, methods, and mechanisms. Annals of the New York Academy of Sciences, 1186, 5–23. https://doi.org/10.1111/j.1749-6632.2009.05337.x.

Andersen

R. M.

(1995). Revisiting the behavioral model and access to medical care: Does it matter? Journal of Health and Social Behavior, 36(1), 1–10. https://doi.org/10.2307/2137284

Angrisani

Lee

(2019). Cognitive decline and household financial decisions at older ages. Journal of the Economics of Ageing, 13, 86–101. https://doi.org/10.1016/j.jeoa.2018.03.003.

Banks

Blundell

Levell

Smith

J. P.

(2019). Life-cycle consumption patterns at older ages in the United States and the United Kingdom: Can medical expenditures explain the difference? American Economic Journal: Economic Policy, 11(3), 27–54. https://doi.org/10.1257/pol.20170182

Beech

B. M.

Ford

Thorpe

R. J.

Bruce

M. A.

Norris

K. C.

(2021). Poverty, racism, and the public health crisis in America. Frontiers in Public Health, 9, 699049. https://doi.org/10.3389/fpubh.2021.699049.

Berkowitz

M. K.

Qiu

(2006). A further look at household portfolio choice and health status. Journal of Banking and Finance, 30(4), 1201–1217. https://doi.org/10.1016/j.jbankfin.2005.05.006

10.

Bhutta

Chang

A. C.

Dettling

L. J.

Hewitt

J. W. H.

, With assistance from J. (2020). Disparities in Wealth by Race and Ethnicity in the 2019 Survey of Consumer Finances. Economic Research. https://www.federalreserve.gov/econres/notes/feds-notes/disparities-in-wealth-by-race-and-ethnicity-in-the-2019-survey-of-consumer-finances-20200928.html

11.

Boyle

P. A.

Wilson

R. S.

Gamble

Buchman

A. S.

Bennett

D. A.

(2012). Poor decision making is a consequence of cognitive decline among older persons without Alzheimer’s disease or mild cognitive impairment. PLoS One, 7(8), e43647. https://doi.org/10.1371/journal.pone.0043647

12.

Braveman

Gottlieb

(2014). The social determinants of health: It’s time to consider the causes of the causes. Public Health Reports, 129 Suppl 2, 19–31. https://doi.org/10.1177/00333549141291S206

13.

Cagney

K. A.

Lauderdale

D. S.

(2002). Education, wealth, and cognitive function in later life. Journals of Gerontology Series B: Psychological Sciences and Social Sciences, 57(2), P163–P172. https://doi.org/10.1093/geronb/57.2.p163

14.

Christelis

Jappelli

Padula

(2009). Cognitive abilities and portfolio choice. European Economic Review, 54(1), 18–38. U-M Articles Search. https://doi.org/10.1016/j.euroecorev.2009.04.001

15.

Clarke

Latham

(2014). Life course health and socioeconomic profiles of Americans aging with disability. Disability and Health Journal, 7(1 Suppl), S15–S23. https://doi.org/10.1016/j.dhjo.2013.08.008

16.

Collins

Lanza

(2009). Latent class and latent transition analysis: With applications in the social, behavioral, and health sciences (718). John Wiley and Sons.

17.

Crimmins

E. M.

Kim

J. K.

Langa

K. M.

Weir

D. R.

(2011). Assessment of cognition using surveys and neuropsychological assessment: The health and retirement study and the aging, demographics, and memory study. Journals of Gerontology Series B: Psychological Sciences and Social Sciences, 66(Suppl 1), i162–i171. https://doi.org/10.1093/geronb/gbr048

18.

Cubanski

Koma

Damico

Neuman

(2018). How many seniors live in poverty? Kaiser Family Foundation. https://files.kff.org/attachment/Issue-Brief-How-Many-Seniors-Live-in-Poverty

19.

Dunn

Grosse

Zuvekas

(2018). Adjusting health expenditures for inflation: A review of measures for health services research in the United States. Health Services Research, 53(1), 175–196. https://doi.org/10.1111/1475-6773.12612

20.

Elbejjani

Fuhrer

Abrahamowicz

Mazoyer

Crivello

Tzourio

Dufouil

(2017). Life-course socioeconomic position and hippocampal atrophy in a prospective cohort of older adults. Psychosomatic Medicine, 79(1), 14–23. https://doi.org/10.1097/PSY.0000000000000365

21.

Fenton

Weissberger

G. H.

Boyle

P. A.

Mosqueda

Yassine

H. N.

Nguyen

A. L.

Lim

A. C.

Han

S. D.

(2022). Cognitive and neuroimaging correlates of financial exploitation vulnerability in older adults without dementia: Implications for early detection of Alzheimer’s disease. Neuroscience and Biobehavioral Reviews, 140, 104773. https://doi.org/10.1016/j.neubiorev.2022.104773.

22.

Fernández-Blázquez

M. A.

Noriega-Ruiz

Ávila-Villanueva

Valentí-Soler

Frades-Payo

Del Ser

Gómez-Ramírez

(2021). Impact of individual and neighborhood dimensions of socioeconomic status on the prevalence of mild cognitive impairment over seven-year follow-up. Aging & Mental Health, 25(5), 814–823. https://doi.org/10.1080/13607863.2020.1725803

23.

Finch

W. H.

Bronk

K. C.

(2011). Conducting confirmatory latent class analysis using mplus. Structural Equation Modeling: A Multidisciplinary Journal, 18(1), 132–151. https://doi.org/10.1080/10705511.2011.532732

24.

Fiske

Wetherell

J. L.

Gatz

(2009). Depression in older adults. Annual Review of Clinical Psychology, 5(1), 363–389. https://doi.org/10.1146/annurev.clinpsy.032408.153621.

25.

Gentili

Riccardi

Gialloreti

L. E.

Scarcella

Stievano

Proietti

M. G.

Rocco

Liotta

(2022). Admission to the long-term care facilities and institutionalization rate in community-dwelling frail adults: An observational longitudinal cohort study. Healthcare, 10(2), 317. https://doi.org/10.3390/healthcare10020317

26.

Gianattasio

K. Z.

Glymour

M. M.

Power

M. C.

(2019). Comparison of methods for algorithmic classification of dementia status in the health and retirement study. Epidemiology, 30(2), 291–302. https://doi.org/10.1097/EDE.0000000000000945

27.

Jackson

Engelman

(2022). Deaths, disparities, and cumulative (Dis)Advantage: How social inequities produce an impairment paradox in later life. The Journals of Gerontology. Series A, Biological Sciences and Medical Sciences, 77(2), 392–401. https://doi.org/10.1093/gerona/glab181

28.

Jacobsen

(2011). America’s aging population. Population Bulletin, 66(1), 1–16. https://assets.prb.org/pdf11/aging-in-america.pdf.

29.

Jung

Thompson

H. J.

Byun

(2022). Social integration: A concept analysis. Nursing Forum, 57(6), 1551–1558. https://doi.org/10.1111/nuf.12843

30.

Korniotis

G. M.

Kumar

(2011). DO older investors make better investment decisions? The Review of Economics and Statistics, 93(1), 244–265. U-M Articles Search. https://doi.org/10.1162/REST_a_00053

31.

Kuh

Ben Shlomo

Ezra

(Eds.), (2004). A life course approach to chronic disease epidemiology. Oxford University Press. https://doi.org/10.1093/acprof:oso/9780198578154.001.0001

32.

Langa

K. M.

Larson

E. B.

Karlawish

J. H.

Cutler

D. M.

Kabeto

M. U.

Kim

S. Y.

Rosen

A. B.

(2008). Trends in the prevalence and mortality of cognitive impairment in the United States: Is there evidence of a compression of cognitive morbidity? Alzheimer’s and Dementia: The Journal of the Alzheimer’s Association, 4(2), 134–144. https://doi.org/10.1016/j.jalz.2008.01.001

33.

Makaroun

L. K.

Brown

R. T.

Diaz-Ramirez

L. G.

Ahalt

Boscardin

W. J.

Lang-Brown

Lee

(2017). Wealth-associated disparities in death and disability in the United States and england. JAMA Internal Medicine, 177(12), 1745–1753. https://doi.org/10.1001/jamainternmed.2017.3903

34.

Marden

J. R.

Tchetgen Tchetgen

E. J.

Kawachi

Glymour

M. M.

(2017). Contribution of socioeconomic status at 3 life-course periods to late-life memory function and decline: Early and late predictors of dementia risk. American Journal of Epidemiology, 186(7), 805–814. https://doi.org/10.1093/aje/kwx155

35.

Masyn

K. E.

(2013). 551Latent class analysis and finite mixture modeling. In Little

T. D.

(Ed.), The oxford handbook of quantitative methods in psychology: Vol. 2: Statistical analysis. Oxford University Press. https://doi.org/10.1093/oxfordhb/9780199934898.013.0025

36.

McMaughan

D. J.

Oloruntoba

Smith

M. L.

(2020). Socioeconomic status and access to healthcare: Interrelated drivers for healthy aging. Frontiers in Public Health, 8, 231. https://doi.org/10.3389/fpubh.2020.00231.

37.

Mira

Newton

Sabbah

(2022). Inequalities in the progress of multiple chronic conditions: A systematic review of longitudinal studies. PLoS One, 17(2), e0263357. https://doi.org/10.1371/journal.pone.0263357

38.

Mutchler

(2021). Aging in the 100 Largest metropolitan areas: How do older adults fare? Center for Social and Demographic Research on Aging Publications, 50.

39.

Nicholas

L. H.

Langa

K. M.

Bynum

J. P. W.

Hsu

J. W.

(2021). Financial presentation of alzheimer disease and related dementias. JAMA Internal Medicine, 181(2), 220–227. https://doi.org/10.1001/jamainternmed.2020.6432

40.

Nylund

K. L.

Asparouhov

Muthén

B. O.

(2007). Deciding on the number of classes in latent class analysis and growth mixture modeling: A Monte Carlo simulation study. Structural Equation Modeling: A Multidisciplinary Journal, 14(4), 535–569. https://doi.org/10.1080/10705510701575396

41.

Okonkwo

O. C.

Wadley

V. G.

Griffith

H. R.

Belue

Lanza

Zamrini

E. Y.

Harrell

L. E.

Brockington

J. C.

Clark

Raman

Marson

D. C.

(2008). Awareness of deficits in financial abilities in patients with mild cognitive impairment: Going beyond self-informant discrepancy. American Journal of Geriatric Psychiatry: Official Journal of the American Association for Geriatric Psychiatry, 16(8), 650–659. https://doi.org/10.1097/JGP.0b013e31817e8a9d

42.

Pachana

N. A.

Byrne

G. J.

Wilson

Tilse

Pinsker

D. M.

Massavelli

Vearncombe

K. J.

Mitchell

L. K.

(2014). Predictors of financial capacity performance in older adults using the Financial Competence Assessment Inventory. International Psychogeriatrics, 26(6), 921–927. https://doi.org/10.1017/S1041610214000209

43.

Plassman

B. L.

Langa

K. M.

Fisher

G. G.

Heeringa

S. G.

Weir

D. R.

Ofstedal

M. B.

Burke

J. R.

Hurd

M. D.

Potter

G. G.

Rodgers

W. L.

Steffens

D. C.

Willis

R. J.

Wallace

R. B.

(2007). Prevalence of dementia in the United States: The aging, demographics, and memory study. Neuroepidemiology, 29(1–2), 125–132. https://doi.org/10.1159/000109998

44.

Pool

L. R.

Burgard

S. A.

Needham

B. L.

Elliott

M. R.

Langa

K. M.

Mendes de Leon

C. F.

(2018). Association of a negative wealth shock with all-cause mortality in middle-aged and older adults in the United States. JAMA, 319(13), 1341–1350. https://doi.org/10.1001/jama.2018.2055

45.

Rocca

W. A.

Petersen

R. C.

Knopman

D. S.

Hebert

L. E.

Evans

D. A.

Hall

K. S.

Gao

Unverzagt

F. W.

Langa

K. M.

Larson

E. B.

White

L. R.

(2011). Trends in the incidence and prevalence of Alzheimer’s disease, dementia, and cognitive impairment in the United States. Alzheimer's and Dementia: The Journal of the Alzheimer's Association, 7(1), 80–93. https://doi.org/10.1016/j.jalz.2010.11.002

46.

Samuel

L. J.

Szanton

S. L.

Wolff

J. L.

Ornstein

K. A.

Parker

L. J.

Gitlin

L. N.

(2020). Socioeconomic disparities in six-year incident dementia in a nationally representative cohort of U.S. older adults: An examination of financial resources. BMC Geriatrics, 20(1), 156. https://doi.org/10.1186/s12877-020-01553-4

47.

Schimmel

Stapleton

D. C.

(2012). The financial repercussions of new work-limiting health conditions for older workers. Inquiry: A Journal of Medical Care Organization, Provision and Financing, 49(2), 141–163. https://doi.org/10.5034/inquiryjrnl_49.02.04

48.

Shankar

McMunn

Banks

Steptoe

(2011). Loneliness, social isolation, and behavioral and biological health indicators in older adults. Health Psychology: Official Journal of the Division of Health Psychology, American Psychological Association, 30(4), 377–385. https://doi.org/10.1037/a0022826

49.

Smith

J. P.

Kington

(1997). Demographic and economic correlates of health in old age. Demography, 34(1), 159–170. https://doi.org/10.2307/2061665

50.

Smith

M. A.

Willis (2010). Financial decision making and cognition in a family context. The Economic Journal, 120(548), F363–F380. U-M Articles Search. https://doi.org/10.1111/j.1468-0297.2010.02394.x

51.

Sonnega

Faul

J. D.

Ofstedal

M. B.

Langa

K. M.

Phillips

J. W. R.

Weir

D. R.

(2014). Cohort profile: The health and retirement study (HRS). International Journal of Epidemiology, 43(2), 576–585. https://doi.org/10.1093/ije/dyu067

52.

Spreng

R. N.

Karlawish

Marson

D. C.

(2016). Cognitive, social, and neural determinants of diminished decision-making and financial exploitation risk in aging and dementia: A review and new model. Journal of Elder Abuse and Neglect, 28(4–5), 320–344. https://doi.org/10.1080/08946566.2016.1237918

53.

Staff

R. T.

Hogan

M. J.

Whalley

L. J.

(2018). The influence of childhood intelligence, social class, education and social mobility on memory and memory decline in late life. Age and Ageing, 47(6), 847–852. https://doi.org/10.1093/ageing/afy111

54.

Steptoe

Wardle

(2017). Life skills, wealth, health, and wellbeing in later life. Proceedings of the National Academy of Sciences of the United States of America, 114(17), 4354–4359. https://doi.org/10.1073/pnas.1616011114

55.

Steptoe

Zaninotto

(2020). Lower socioeconomic status and the acceleration of aging: An outcome-wide analysis. Proceedings of the National Academy of Sciences of the United States of America, 117(26), 14911–14917. https://doi.org/10.1073/pnas.1915741117

56.

Sudo

F. K.

Laks

(2017). Financial capacity in dementia: A systematic review. Aging and Mental Health, 21(7), 677–683. https://doi.org/10.1080/13607863.2016.1226761

57.

Toot

Swinson

Devine

Challis

Orrell

(2017). Causes of nursing home placement for older people with dementia: A systematic review and meta-analysis. International Psychogeriatrics, 29(2), 195–208. https://doi.org/10.1017/S1041610216001654

58.

Weida

E. B.

Phojanakong

Patel

Chilton

(2020). Financial health as a measurable social determinant of health. PLoS One, 15(5), e0233359. https://doi.org/10.1371/journal.pone.0233359

59.

Weissberger

G. H.

Goodman

M. C.

Mosqueda

Schoen

Nguyen

A. L.

Wilber

K. H.

Gassoumis

Z. D.

Nguyen

C. P.

Han

S. D.

(2020). Elder abuse characteristics based on calls to the national center on elder abuse resource line. Journal of Applied Gerontology: The Official Journal of the Southern Gerontological Society, 39(10), 1078–1087. https://doi.org/10.1177/0733464819865685

60.

Widera

Steenpass

Marson

Sudore

(2011). Finances in the older patient with cognitive impairment: “He didn’t want me to take over.”. JAMA, 305(7), 698–706. https://doi.org/10.1001/jama.2011.164

61.

Wood

Lichtenberg

P. A.

(2017). Financial capacity and financial exploitation of older adults: Research findings, policy recommendations and clinical implications. Clinical Gerontologist, 40(1), 3–13. https://doi.org/10.1080/07317115.2016.1203382

62.

Woolf

Aron

Dubay

Simon

Zimmerman

Luk

(2015). How are income and wealth linked to health and longevity? https://www.urban.org/sites/default/files/publication/49116/2000178-How-are-Income-and-Wealth-Linked-to-Health-and-Longevity.pdf

63.

Zahodne

L. B.

Sol

Kraal

(2019). Psychosocial pathways to racial/ethnic inequalities in late-life memory trajectories. Journals of Gerontology Series B: Psychological Sciences and Social Sciences, 74(3), 409–418. https://doi.org/10.1093/geronb/gbx113

64.

Zahodne

L. B.

Sharifian

Kraal

A. Z.

Zaheed

A. B.

Sol

Morris

E. P.

Schupf

Manly

J. J.

Brickman

A. M.

(2021). Socioeconomic and psychosocial mechanisms underlying racial/ethnic disparities in cognition among older adults. Neuropsychology, 35(3), 265–275. https://doi.org/10.1037/neu0000720

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.19 MB