Skip to main content
NCBI home page
Alzheimer's & Dementia logoLink to Alzheimer's & Dementia
. 2024 Mar 13;20(5):3147–3156. doi: 10.1002/alz.13768

Racial and ethnic differences in the association between depressive symptoms and cognitive outcomes in older adults: Findings from KHANDLE and STAR

Marcia P Jimenez 1,✉, Emma L Gause 2, Kendra D Sims 1, Eleanor Hayes‐Larson 3, Emily P Morris 4, Evan Fletcher 5, Jennifer Manly 6, Paola Gilsanz 7, Yenee Soh 7, Maria Corrada 8, Rachel A Whitmer 9, Medellena Maria Glymour 1
PMCID: PMC11095484  PMID: 38477489

Abstract

INTRODUCTION

Depressive symptoms are associated with higher risk of dementia, but how they impact cognition in diverse populations is unclear.

METHODS

Asian, Black, Latino, or White participants (n = 2227) in the Kaiser Healthy Aging and Diverse Life Experiences (age 65+) and the Study of Healthy Aging in African Americans (age 50+) underwent up to three waves of cognitive assessments over 4 years. Multilevel models stratified by race/ethnicity were used to examine whether depressive symptoms were associated with cognition or cognitive decline and whether associations differed by race/ethnicity.

RESULTS

Higher depressive symptoms were associated with lower baseline verbal episodic memory scores (−0.06, 95% CI: −0.12, −0.01; −0.15, 95% CI: −0.25, −0.04), and faster decline annually in semantic memory (−0.04, 95% CI: −0.07, −0.01; −0.10, 95% CI: −0.15, −0.05) for Black and Latino participants. Depressive symptoms were associated with lower baseline but not decline in executive function.

DISCUSSION

Depressive symptoms were associated with worse cognitive outcomes, with some evidence of heterogeneity across racial/ethnic groups.

HIGHLIGHTS

  • We examined whether baseline depressive symptoms were differentially associated with domain‐specific cognition or cognitive decline by race/ethnicity.

  • Depressive symptoms were associated with worse cognitive scores for all racial/ethnic groups across different domains examined.

  • Higher depressive symptoms were associated with faster cognitive decline for semantic memory for Black and Latino participants.

  • The results suggest a particularly harmful association between depressive symptoms and cognition in certain racial/ethnic groups.

Keywords: Alzheimer's disease, cognitive function, depressive symptoms, longitudinal data, mental health, race and ethnicity

1. BACKGROUND

The older adult US population is projected to both grow and become more racially and ethnically diverse. 1 Given the higher burden of dementia among Black and Latino individuals, uncovering modifiable risk factors of Alzheimer's disease (AD) and related dementias (ADRD) is essential. 2 Convincing evidence – derived predominantly in samples comprising non‐Hispanic Black and White‐identified participants – supports the notion that depression increases the risk of cognitive decline. 3 , 4 , 5 , 6 , 7 , 8 , 9 Non‐clinical depressive symptoms associated with disease risk can be effectively screened with commonly used indices in the clinical setting or on a population level. 10 Since depression can be effectively treated with a conjunction of pharmacotherapy, intervention from a mental health professional, and behavior change, it is an identifiable modifiable risk factor for the prevention or attenuation of cognitive decline in older adults. 11 However, the relationships between race/ethnicity and depressive symptoms and between depressive symptoms and cognition are complex. 12 , 13

Minority stress theory (MST) asserts that exposure to racial discrimination and acculturative stress constitutes a hostile psychological environment leading to negative health consequences. In addition, studies have shown that racial differences in socioeconomic status (SES) are an important contributor to racial disparities in health. 14 The reserve capacity model (RCM) posits that certain subgroups of the population, such as individuals of lower SES, are more vulnerable to the negative health effects of depression because increased exposure to chronic stress diminishes reserve capacity. 15 In other words, MST in conjunction with RCM suggests that racially minoritized individuals may or may not experience depression more often than White individuals, but they experience more severe health consequences of depression when it is present. The literature on the prevalence of depressive symptoms across race and ethnicity is mixed; some studies found the prevalence to be lower among White participants compared to Black and Latino participants, 16 whereas other studies found a lower prevalence of depressive symptoms among Black and Latino participants. 17 , 18 , 19 Furthermore, studies show that when depression affects Black individuals, it is usually untreated, more severe, and disabling than in White individuals. This suggests that the burden of depressive disorders may be higher among Black than White participants. 17

Currently, the relation of depressive symptoms with late life cognition across racial/ethnic groups is only partially understood. To our knowledge, this is among the first studies to investigate, in a multi‐ethnic older adult sample, the associations of depressive symptoms with baseline cognition and cognitive decline. We hypothesized that the prevalence of depressive symptoms among racially minoritized versus socially dominant older adults would contribute to both lower baseline cognition and accelerated cognitive decline.

2. METHODS

2.1. Data

This study used data from the Kaiser Healthy Aging and Diverse Life Experiences (KHANDLE) and the Study of Healthy Aging in African Americans (STAR) cohorts. KHANDLE comprises community‐dwelling older adults in the San Francisco Bay and Sacramento regions of California who were aged 65 or older as of January 1, 2017 and were long‐term members of Kaiser Permanente Northern California. KHANDLE participants were eligible for inclusion if they spoke either English or Spanish and participated in at least one voluntary health checkup (Multiphasic Health Checkup [MHC]) with Kaiser Permanente between 1964 and 1985. Exclusion criteria included a previous diagnosis of dementia or other neurodegenerative disease (frontotemporal dementia, Lewy body disease, Pick's disease, Parkinson's disease with dementia, Huntington's disease) or a diagnosis related to poor health that might interfere with study participation or study interviews (defined by hospice activity in the past 12 months, history of severe chronic obstructive pulmonary disease in the past 6 months, congestive heart failure hospitalizations in the past 6 months, and history of end‐stage renal disease or dialysis in the past 12 months). Participants selected for inclusion in KHANDLE were randomly sampled from this eligible participant pool within strata of race and ethnicity and educational attainment to achieve balanced numbers of participants who reported Asian, Black, Latino, and White race and ethnicity with diverse educational backgrounds.

The STAR study is similarly composed of community‐dwelling older adults from the San Francisco Bay area but enrolled only participants who identified as Black or African American that participated in at least one MHC exam and who were aged 50 or older as of January 1, 2018. The STAR study used the same exclusion as KHANDLE, outlined previously. More information about the KHANDLE and STAR cohorts can be found in cited materials. 20 , 21 , 22 , 23 At wave 1, which we will hereafter refer to as baseline, KHANDLE had 1712 participants and STAR had 764 participants (N = 2389). Participants were excluded from the analysis set if they were missing or refused to answer race and ethnicity data or if they were missing information on any of the primary model covariates (including age at baseline, sex, education, income, and marital status). Four participants who self‐reported as Native American were excluded due to concerns about small numbers and identifiability. The final analysis cohort was made up of 2227 participants. (Figure S1).

2.2. Exposure

The primary exposure of interest was depressive symptoms reported at baseline interviews. The KHANDLE cohort used the four‐item Patient‐Reported Outcomes Measurement Information System (PROMIS) computer adaptive testing (CAT) version 1.0 administered through National Institutes of Health (NIH) Toolbox and the STAR cohort used the eight‐item PROMIS‐57 administered through DatStat. We followed the depressive symptoms scoring manual from PROMIS to standardize the PROMIS scores across forms between the two cohorts. These depressive symptom scores were transformed into theta scores, which were standardized to the US adult population so that a difference in one unit represented a standardized difference in depressive symptoms, which were centered around zero for the average US adult. Depressive symptom theta scores were treated as a continuous variable.

RESEARCH IN CONTEXT

  1. Systematic review: The authors reviewed the existing literature on the association between depressive symptoms with cognition and cognitive decline among older adults across racial and ethnic groups.

  2. Interpretation: Our findings suggest a particularly harmful association for Black and Latino participants, between greater depressive symptoms with worse cognitive scores and faster cognitive decline in specific domains examined.

  3. Future directions: Future research should include more representative data of underrepresented racial and ethnic groups and investigate the robustness of these results to better understand how depressive symptoms are associated with late‐life cognitive function longitudinally.

2.3. Outcomes

Cognitive function was assessed across three waves using the Spanish and English Neuropsychological Assessment Scales (SENAS). Both KHANDLE and STAR assessed cognition using SENAS three times between 2017 and 2021. The SENAS includes three cognitive domains (executive function, verbal episodic memory, and semantic memory) and was administered in either English or Spanish, with language of administration determined by an algorithm that considered preferred language and everyday language usage in a variety of settings. The SENAS is a battery of cognitive tests that has undergone extensive development for valid comparisons of cognitive change across racial and ethnic and linguistically diverse groups. 24 Briefly, item response theory and confirmatory factor analysis methods were used to construct measures that are psychometrically matched across domains with respect to level of reliability across the ability continuum. 25 The Episodic Memory Score is derived from a multitrial word‐list‐learning test. The semantic memory measure is a composite of highly correlated verbal (object‐naming) and non‐verbal (picture association) tasks. The executive function composite is constructed from component tasks of category fluency, phonemic (letter) fluency, and working memory (digit‐span backward, list sorting). Administration procedures, measure development, and psychometric characteristics of the SENAS battery are described in detail elsewhere. 24 , 26 We were unable to assess semantic memory, which uses visual stimuli, during the third waves of KHANDLE and STAR because SENAS was administered by phone during the COVID‐19 pandemic. SENAS measures were standardized into z‐scores using baseline values for analysis across the two study cohorts prior to restricting to complete cases. These z‐scores can be interpreted as standard deviations away from the mean of zero for the average individual in the analysis set at baseline.

2.4. Covariates

Race and ethnicity were self‐reported at baseline. Individuals were able to select more than one racial and ethnic category, which were collapsed for these analyses as Latino, Black, Asian, or White using that assignment order. Other baseline sociodemographic information included age (in years at time of assessment), biological sex (female/male), educational attainment (operationalized as college degree or higher vs no college degree), household income (operationalized as greater than 55 k annually vs less than 55 k), and marital status (operationalized as married or living with partner vs not). Sensitivity analyses included additional covariates that could potentially mediate associations between depressive symptoms and cognitive decline: daily socializing, daily physical activity, smoking history, heavy drinking, and self‐reported health. More information on these variables is provided in the supplementary material (Methods S1).

2.5. Statistical analysis

We first descriptively compared the distributions of depressive symptoms across race/ethnicity subgroups (following Ward et al.), 27 as well as the distributions of SENAS cognitive domains across these groups. To assess the association between depressive symptoms and cognition and cognitive decline, we used multilevel linear regression with random intercepts for individuals to separately assess the associations of depressive symptoms with baseline and longitudinal change in each SENAS domain (executive function, verbal episodic memory, and semantic memory). We used a multilevel approach as studies have shown that using the paired difference as an outcome is less efficient and vulnerable to bias. 28 An interaction term between depressive symptoms and years since baseline was included to test for differences in decline in the cognitive domain over time. In an additional sensitivity analysis, we also fit a model with interactions between time, and all included covariates to control for confounding on estimates of cognitive decline. Specifying a random intercept model without random slopes allowed us to test the hypothesis that the change in cognitive decline might differ by race/ethnicity in the stratified analysis. In addition, with only —two or three time points, the random intercept model is more efficient (tighter confident intervals) and less vulnerable to convergence difficulties. To examine race/ethnicity as a potential effect modifier of the association between depressive symptoms and cognitive function and subsequent cognitive decline, we used a three‐step approach. First, we included an interaction term between race/ethnicity and depression for each cognitive outcome and used a joint likelihood ratio test to determine whether the model including the interaction term was a better fit to the data compared to the nested model without interaction terms. Second, we evaluated associations in models stratified by race and ethnicity. Third, we evaluated heterogeneity in the associations between depressive symptoms and cognitive health by computing the differences of the estimate of interest (eg, depressive symptoms) obtained from each race/ethnicity subgroup in stratified models. 29 In this analysis, we selected Black participants as the comparison group for two reasons: (1) it was the subgroup with the biggest sample size, which would allow more statistical power for comparisons; and (2) we are interested in de‐institutionalizing the requirement for White participants as the comparison group to demonstrate health inequity (following Adkins‐Jackson). 30 Confidence intervals for these differences were calculated through bootstrapping with 1000 iterations. We did not adopt format multiple comparisons because reducing the type I error for null associations increases the type II error for those associations that are not null. 31

Our main results are based on complete case analysis (N = 2227). We also performed supplementary analysis with multiple imputation using chained equations (MICE) on the sample with and without missing covariates (N = 2469). 32 We imputed missing values across 10 datasets, conditional on the distribution of complete case data for all main model variables in addition to study cohort, potential mediating variables (daily socializing, daily physical activity, smoking history, heavy drinking, and self‐reported health), and practice setting where the interview took place in each wave. Models were fit to each of the 10 imputed datasets, and then results were pooled, adjusting for variability within and between the standard errors of the datasets using Rubin's Rules. 33 We also performed a sensitivity analysis that additionally adjusted for the lifestyle covariates described above. We further adjusted for a wave indicator and the practice setting to account for practice effects. 34 To assess non‐linear associations of depressive symptoms with cognitive health, we used natural splines with three to four knots based on the Akaike information criterion. Finally, given the age difference between KHANDLE and STAR participants, and given that age is a known risk factor for cognitive function and decline, we further analyzed these associations only among KHANDLE participants. We followed current STROBE reporting guidelines that recommend not to include a column containing inferential statistics (eg, p values). 35 , 36 , 37 Instead, we based meaningful differences on both the magnitude of observed estimates and consistency with causal theory. Analyses were conducted using R software version 4.1.3. 38 All code used for this analysis can be found at https://github.com/SpatialHealth/KHANDLE‐STAR_Depression_Cognition. 39

3. RESULTS

Among the 2227 individuals included in the analysis, 17% (n = 380) were classified as Asian, 48% (n = 1076) as Black, 14% (n = 318) as Latino, and 20% (n = 453) as White (Table 1). The average age of the cohort at baseline was 72.7 years (SD = 8.0), and the majority were female (62%). Education, income, and other sociodemographic factors varied across race and ethnicity (Table 1). Depressive symptom levels were somewhat lower among Asian and Black participants compared to Latino and White participants (Figure 1).

TABLE 1.

Characteristics of complete case participants in KHANDLE and STAR by race and ethnicity at baseline (N = 2227).

Characteristic Overall Cohort N = 2227 Asian American participants N = 380 Black participants N = 1076 Latino participants N = 318 White participants N = 453
Cohort N(%)
KHANDLE 1535 (68.9) 380 (100.0) 390 (36.2) 312 (98.1) 453 (100.0)
STAR 692 (31.1) 0 (0.0) 686 (63.8) 6 (1.9) 0 (0.0)
Depression theta score median [SD] −0.18 [0.75] −0.19 [0.79] −0.25 [0.77] −0.01 [0.77] −0.01 [0.63]
Female gender N (%) 1386 (62.2) 199 (52.4) 739 (68.7) 186 (58.5) 262 (57.8)
Age at baseline median [SD] 72.7 [8.0] 74.3 [6.4] 70.5 [8.5] 74.5 [6.3] 75.5 [6.9]
Education N (%)
Less than college degree 1224 (55.0) 128 (33.7) 690 (64.1) 214 (67.3) 192 (42.4)
College degree or further education 1003 (45.0) 252 (66.3) 386 (35.9) 104 (32.7) 261 (57.6)
Total household income N (%)
Less than or equal to USD55,000 750 (33.7) 83 (21.8) 396 (36.8) 139 (43.7) 132 (29.1)
More than USD55,000 1477 (66.3) 297 (78.2) 680 (63.2) 179 (56.3) 321 (70.9)
Marital status N (%)
Married or living with a partner 1044 (46.9) 118 (31.1) 604 (56.1) 134 (42.1) 188 (41.4)
Not married or partnered 1183 (53.1) 262 (68.9) 472 (43.9) 184 (57.9) 265 (58.5)
Baseline executive function median [SD] 0.04 [0.98] −0.19 [0.9] −0.03 [0.9] −0.20 [0.9] 0.59 [1.1]
Baseline verbal episodic memory median [SD] 0.03 [0.98] 0.10 [1] 0.06 [0.9] −0.23 [1] 0.06 [1.1]
Baseline semantic memory median [SD] 0.05 [0.98] −0.13 [1.1] −0.21 [0.8] 0.31 [0.9] 0.82 [0.8]
Open in a new tab

FIGURE 1.

FIGURE 1

Open in a new tab

Distribution of baseline depressive symptoms by race and ethnicity group in KHANDLE and STAR (N = 2227).

3.1. Baseline cognitive function

Table 2 shows the estimates for the association of depressive symptoms with baseline cognitive function for the full sample and stratified by race and ethnicity. In the full sample, higher depressive symptoms were associated with −0.06 (95% CI: −0.10, −0.02) SD units lower executive function and −0.05 (95% CI: −0.09, −0.007) SD units lower verbal episodic memory. We did not observe an association between depressive symptoms and semantic memory score in the full sample. The likelihood ratio test suggested a better goodness of fit when interaction terms between depressive symptoms and race/ethnicity group were included. Analyses stratified by race and ethnicity suggested that higher depressive symptoms at baseline were associated with lower executive function and verbal episodic memory scores among Black and Latino participants (eg, −0.06 SD, 95% CI: −0.12, −0.01; −0.17 SD, 95% CI: −0.28, −0.06 for executive function for Black and Latino participants, respectively). Higher depressive symptoms at baseline were associated with lower semantic memory scores among Latino participants (−0.15, 95% CI:−0.27, −0.04), but not among Asian, Black, or White participants. This finding aligns with our analysis of the difference in stratified effect estimates for depressive symptoms and baseline cognition, which showed that the estimate for Latino participants was significantly different compared to the estimate from Black participants (eg, difference in point estimate for executive function −0.11, 95% CI: −0.20, −0.01; Table S1). This suggested a particularly harmful association between depressive symptoms and cognition for Latino participants. We did not observe a difference in the point estimates for Asian or White participants (Table S1). The results of the sensitivity analysis using the multiply imputed data showed similar associations between baseline depressive symptoms and lower scores on specific cognitive domains. For example, in the full sample, higher depressive symptoms were associated with −0.07 (95% CI: −0.11, −0.03) SD units lower executive function and −0.06 (95% CI: −0.10, −0.02) SD units lower verbal episodic memory (Table S2). The spline analysis did not show evidence of non‐linear relationships between depressive symptoms and baseline cognitive function, though results of the executive memory models might suggest the existence of a threshold for worse cognitive health among Asian participants with greater than average depressive symptoms (Figure S2). Further research would be needed to investigate this.

TABLE 2.

Estimated associations between depressive symptoms and baseline cognitive function among participants in KHANDLE and STAR (N = 2227).

Group N Executive function score Verbal episodic memory score Semantic memory score a
Overall 2227 −0.06 (−0.10, −0.02) −0.05 (−0.09, −0.007) −0.03 (−0.07, 0.01)
Asian 380 −0.03 (−0.13, 0.06) −0.01 (−0.11, 0.10) −0.01 (−0.13, 0.12)
Black 1076 −0.06 (−0.12, −0.01) −0.06 (−0.12, −0.01) −0.01 (−0.07, 0.04)
Latino 318 −0.17 (−0.28, −0.06) −0.15 (−0.26, −0.04) −0.15 (−0.27, −0.04)
White 453 0.04 (−0.09, 0.17) 0.02 (−0.09, 0.14) −0.00 (−0.10, 0.10)
Open in a new tab

Notes: Models adjusted for age at baseline, year of follow‐up, sex, college education, income greater than USD55,000, and marital status. Overall model also adjusted for race/ethnicity.

a

Only two waves of data available for analysis.

3.2. Cognitive decline

In the full sample, we observed that higher depressive symptoms were associated with accelerated decline in semantic memory score (difference in annual rate of change = −0.04 SD, 95% CI: −0.06, −0.02; Table 3). In the stratified models, higher depressive symptoms among Black participants were associated with accelerated decline in verbal episodic memory (difference in annual rate of change = −0.03 SD, 95% CI: −0.05, −0.001) and semantic memory score (difference in annual rate of change = −0.04 SD, 95% CI: −0.07, −0.01; Table 3). Higher depressive symptoms were associated with a rate of decline in semantic memory score that was faster by −0.10 SD (95% CI: −0.15, −0.05) among Latino participants. Depressive symptoms were not associated with rate of decline in semantic memory among Asian or White participants (Table 3). We did not observe associations between depressive symptoms and rate of decline in executive function score in the full sample or in stratified results. In the analysis assessing the difference in stratified effect estimates for depressive symptoms and cognitive decline, we observed that more depressive symptoms were associated with a slower decline in verbal episodic memory for White participants compared to Black participants (eg, difference in point estimate for decline in verbal episodic memory 0.07, 95% CI: 0.02, 0.12; Table S3). This suggested a particularly harmful association between depressive symptoms and cognitive decline for Black participants. The results of the sensitivity analysis using the multiply imputed data showed similar associations between baseline depressive symptoms and accelerated decline (eg, difference in annual rate of change in semantic memory for the full sample = −0.04 SD, 95% CI: −0.06, −0.02; Table S4). Further, in models including interaction terms of time for all covariates, results were similar (eg, difference in annual rate of change in semantic memory for the full sample = −0.04 SD, 95% CI: −0.06, −0.02; Table S5). Finally, further adjusting for additional lifestyle covariates, wave indicator and practice setting attenuated and reduced precision around the estimates (Tables S6–S7). Results considering only participants in the KHANDLE cohort were similar in magnitude and direction, although the confidence interval for Black participants included the null potentially due to a smaller sample size (Tables S8 and S9).

TABLE 3.

Estimated associations between depressive symptoms and annual rate of cognitive decline among participants in KHANDLE and STAR (N = 2227).

Group N Executive function score Verbal episodic memory score Semantic memory score a
Overall 2227 −0.01 (−0.02, 0.01) 0.00 (−0.02, 0.02) −0.04 (−0.06, −0.02)
Asian 380 0.00 (−0.02, 0.03) 0.00 (−0.04, 0.04) −0.03 (−0.08, 0.01)
Black 1076 −0.02 (−0.03, 0.003) −0.03 (−0.05, −0.001) −0.04 (−0.07, −0.01)
Latino 318 −0.03 (−0.06, 0.01) 0.02 (−0.03, 0.07) −0.10 (−0.15, −0.05)
White 453 0.01 (−0.02, 0.04) 0.04 (−0.002, 0.09) 0.00 (−0.04, 0.04)
Open in a new tab

Notes: Models adjusted for: age at baseline, sex, college education, income greater than USD55,000, and marital status. Overall model also adjusted for race/ethnicity.

a

Only two waves of data available for analysis.

4. DISCUSSION

In a diverse sample of adults aged at least 50, more depressive symptoms were associated with worse baseline cognitive function as well as faster cognitive decline over 4 years for Black and Latino participants in specific cognitive domains over time. Associations differed across cognitive domains and differed by race and ethnicity. Associations were especially large for Latino individuals, with effect sizes several times the magnitude as in White respondents.

Our findings on lower baseline depressive symptoms among Asian and Black participants compared to Latino and White participants are consistent with previously documented racial and ethnic differences in depression. For example, Asian American adults, particularly if foreign‐born, have been reported to have the lowest lifetime prevalence of mental health disorders, such as depression, compared to Black, Latino, and White participants. 40 , 41 In addition, according to the National Center for Health Statistics, non‐Hispanic Asian adults in 2019 were least likely to experience depressive symptoms compared with Hispanic, non‐Hispanic White, and non‐Hispanic Black adults. 42 Black individuals have been reported to have lower lifetime and 12‐month rates of major depressive disorder compared with White individuals. 43 The “Black‐White depression paradox” describes the lower prevalence of depression among non‐Hispanic Black (relative to non‐Hispanic White) individuals despite their greater exposure to major life stressors, a phenomenon that remains unexplained. 44 The combination of higher average levels of depressive symptoms and greater association with cognitive outcomes among Latino individuals elevates the likely population impact of depressive symptoms on cognitive outcomes among Latino older adults.

In models including the full sample, depressive symptoms were associated with worse scores in both executive function and episodic memory. This is consistent with previous studies reporting that depression may negatively impact risk of mild cognitive impairment 45 and neurocognitive functions. We found that associations were particularly marked in a multi‐ethnic older sample for cognitive domains that are governed by fronto‐subcortical networks, such as executive functions. 46 Further, our results suggested effect modification by race and ethnicity in associations of depressive symptoms and baseline cognitive function for some cognitive domains. We observed inverse associations between depressive symptoms and cognition (executive function and verbal episodic memory scores) among Black and Latino participants, but not among Asian or White participants. Our results on higher depressive symptoms associated with lower executive function and verbal episodic memory scores among Black participants are consistent with an analysis that showed that depressive symptoms were associated with worse task‐switching, inhibition, and episodic memory among Black participants; taken together, this suggests that Black older adults may be more vulnerable than their White counterparts to negative effects of depressive symptoms on cognition. 47 Our results on higher depressive symptoms associated with lower executive function and verbal episodic memory scores among Latino participants aligns with previous work among Caribbean Hispanic older adults showing that greater depressive symptoms were associated with worse episodic memory. 48 Among Asian participants, prior findings are mixed regarding associations between depressive symptoms and cognition, with a cross‐sectional study showing increased prevalence of depressive symptoms in cognitively impaired Chinese American participants, 49 while a longitudinal study found that more severe depressive symptoms were associated with better cognitive function among patients in China. 50 When taken together, the results suggest a particularly harmful association between depressive symptoms and baseline executive function scores for Black and Latino participants, but not for Asian or White participants. These results are in agreement with previous studies that showed that depressive symptoms were more strongly related to executive functioning among Black older adults. 47 , 51 , 52

With regard to cognitive decline, in models including the full sample, we observed that more depressive symptoms were associated with a faster rate of cognitive decline in semantic memory scores but not in other cognitive domains. Null associations between depressive symptoms and change in cognitive performance across time were also observed in a randomized trial of 445 middle‐aged to older adults designed to compare the effectiveness of Internet cognitive behavioral therapy relative to an online attention control. 53 In our sample, the association between depressive symptoms and cognitive decline was modified by race and ethnicity. More depressive symptoms were associated with a faster rate of cognitive decline in verbal episodic memory among Black participants. This is consistent with a recent finding from the Washington/Hamilton Heights Inwood Columbia Aging Project (WHICAP), a multi‐ethnic cohort including White, Black, and Hispanic participants, that reported that a higher baseline level of depressive symptoms was associated with a faster decline in episodic memory. 5 The results on rate of cognitive decline for Black participants despite the lower scores for depression symptoms (compared to Latino and White participants) could be due to the fact that when depression affects Black participants, it is often untreated and more severe compared to White participants, and therefore, when Black participants report heightened depressive symptoms, they may more significantly influence cognitive functioning than in other groups. 17 Simultaneously, researchers have observed that Black participants have reduced access to mental health services and often receive poorer quality care than White participants. 54 Along these lines, disparities in primary care (eg, counselling/referrals for counseling, antidepressant medications) persist, with Black and Latino participants having less access to care than White participants. 12 Thus, higher depression symptoms in Black participants might be indicative of untreated depression and seemed associated with a steeper decline in cognition. More depressive symptoms were associated with a faster rate of cognitive decline in semantic memory scores among Black and Latino participants, although these results should be interpreted with caution due to limited time point availability. Finally, we observed no association between depressive symptoms and rate of cognitive decline in analyses restricted to Asian or White participants. The reasons for this may include the low prevalence of depressive symptoms among Asian and White participants, resulting in imprecise estimates with wide confidence intervals for these groups. 55 When taken together, the results suggest a particularly harmful association between depressive symptoms and a faster rate of cognitive decline in verbal episodic memory for Black participants, but not for Latino, Asian, or White participants.

Several limitations in this study should be noted. First, disease mechanisms common to depression and cognitive health may explain the associations observed. It remains controversial whether depressive symptoms increase risk of dementia, are an early symptom of neurodegeneration, or are a reaction to early cognitive deficit. To account for this, we used longitudinal data including three repeated measurements of cognitive performance to evaluate trajectories of cognitive decline. We note that prior studies with much longer follow‐up helped establish temporal order; for example, a study from Denmark demonstrated that depression diagnoses predicted incident dementia over 20 years later. 5 The major contribution in the current study is to demonstrate the relevance of depressive symptoms for Black and Latino adults, who were often not represented in prior work. However, we only had two assessments of semantic memory outcome due to the shift to phone administration of KHANDLE visits during the COVID‐19 pandemic; our estimates of semantic memory decline over time have a shorter follow‐up and are less precise. Second, Latino participants include ethnic/regional subgroups (eg, Mexicans, Dominicans, Puerto Ricans) with important intergroup differences; Asian Americans are similarly heterogeneous. However, we were underpowered to assess Latino or Asian subgroup differences in the association between depression symptoms and cognitive health. Larger‐scale studies that are racially and ethnically diverse are needed.

Despite these limitations, this study has notable strengths. The use of longitudinal data in the KHANDLE and STAR cohorts allowed us to examine prospective associations between depression and cognitive decline. The study was conducted in an exceptionally diverse cohort, with rich data on covariates that allowed us to adjust for important confounders. Using a well‐validated neuropsychological assessment scale enabled us to examine domain‐specific associations.

Among participants in KHANDLE and STAR, greater depressive symptoms were associated with worse cognitive health and potentially faster cognitive decline among certain cognitive domains and race and ethnicity groups. Our findings provide support for the hypothesis that Black and Latino participants are more susceptible to the negative effects of depressive symptoms on cognitive health.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflicts of interest. Author disclosures are available in the supporting information.

CONSENT STATEMENT

All human subjects provided informed consent.

Supporting information

Supporting Information

ALZ-20-3147-s002.docx (10MB, docx)

Supporting Information

ALZ-20-3147-s001.pdf (2.5MB, pdf)

ACKNOWLEDGMENTS

The authors thank the staff and participants of the KHANDLE and STAR studies for their important contributions. The authors would also like to thank Dr. Ruijia Chen for her assistance with our bootstrapped difference estimates. This study is supported by the National Institute on Aging (NIA) R01AG052132, RF1AG050782. M.P.J. is supported by 4R00AG066949‐03 and the Alzheimer's Association 23AARGD‐1030259.

Jimenez MP, Gause EL, Sims KD, et al. Racial and ethnic differences in the association between depressive symptoms and cognitive outcomes in older adults: Findings from KHANDLE and STAR. Alzheimer's Dement. 2024;20:3147–3156. 10.1002/alz.13768

REFERENCES

  • 1. Colby SL, Ortman JM. Projections of the Size and Composition of the U.S. Population: 2014 to 2060. U.S. Department of Commerce Economics and Statistics Administration. U.S. Census Bureau; 2014:p25‐1143. Accessed October 2, 2023. https://www.census.gov/content/dam/Census/library/publications/2015/demo/p25‐1143.pdf [Google Scholar]
  • 2. Kawas CH, Corrada MM, Whitmer RA. Diversity and disparities in dementia diagnosis and care: a challenge for all of us. JAMA Neurol. 2021;78(6):650‐652. doi: 10.1001/jamaneurol.2021.0285 [DOI] [PubMed] [Google Scholar]
  • 3. Bennett S, Thomas AJ. Depression and dementia: cause, consequence or coincidence? Maturitas. 2014;79(2):184‐190. doi: 10.1016/j.maturitas.2014.05.009 [DOI] [PubMed] [Google Scholar]
  • 4. Whitmer RA, Gilsanz P, Mungas DM, et al. Is depression associated with cognitive decline in a diverse cohort of the oldest‐old? Findings from the LifeAfter90 Study. Alzheimers Dementia. 2021;17(S6). https://alz‐journals.onlinelibrary.wiley.com/doi/10.1002/alz.056104 [Google Scholar]
  • 5. Zahodne LB, Stern Y, Manly JJ. Depressive symptoms precede memory decline, but not vice versa, in non‐demented older adults. J Am Geriatr Soc. 2014;62(1):130‐134. doi: 10.1111/jgs.12600 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6. Plassman BL. Systematic review: factors associated with risk for and possible prevention of cognitive decline in later life. Ann Intern Med. 2010;153(3):182. doi: 10.7326/0003-4819-153-3-201008030-00258 [DOI] [PubMed] [Google Scholar]
  • 7. Semkovska M, Quinlivan L, O'Grady T, et al. Cognitive function following a major depressive episode: a systematic review and meta‐analysis. Lancet Psychiatry. 2019;6(10):851‐861. doi: 10.1016/S2215-0366(19)30291-3 [DOI] [PubMed] [Google Scholar]
  • 8. Zaninotto P, Batty GD, Allerhand M, Deary IJ. Cognitive function trajectories and their determinants in older people: 8 years of follow‐up in the English Longitudinal Study of Ageing. J Epidemiol Community Health. 2018;72(8):685‐694. doi: 10.1136/jech-2017-210116 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9. Köhler S, van Boxtel MPJ, van Os J, et al. Depressive symptoms and cognitive decline in community‐dwelling older adults. J Am Geriatr Soc. 2010;58(5):873‐879. doi: 10.1111/j.1532-5415.2010.02807.x [DOI] [PubMed] [Google Scholar]
  • 10. Vilagut G, Forero CG, Barbaglia G, Alonso J. Screening for depression in the general population with the center for epidemiologic studies depression (CES‐D): a systematic review with meta‐analysis. PLoS One. 2016;11(5):e0155431. doi: 10.1371/journal.pone.0155431 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11. Zenebe Y, Akele B, W/Selassie M, Necho M. Prevalence and determinants of depression among old age: a systematic review and meta‐analysis. Ann Gen Psychiatry. 2021;20(1):55. doi: 10.1186/s12991-021-00375-x [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12. Stockdale SE, Lagomasino IT, Siddique J, McGuire T, Miranda J. Racial and ethnic disparities in detection and treatment of depression and anxiety among psychiatric and primary health care visits, 1995‐2005. Med Care. 2008;46(7):668‐677. doi: 10.1097/MLR.0b013e3181789496 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13. Jorm AF. Is depression a risk factor for dementia or cognitive decline? A review. Gerontology. 2000;46(4):219‐227. doi: 10.1159/000022163 [DOI] [PubMed] [Google Scholar]
  • 14. Williams DR, Priest N, Anderson NB. Understanding associations among race, socioeconomic status, and health: patterns and prospects. Health Psychol. 2016;35(4):407‐411. doi: 10.1037/hea0000242 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15. Gallo LC, Matthews KA. Understanding the association between socioeconomic status and physical health: do negative emotions play a role? Psychol Bull. 2003;129(1):10‐51. doi: 10.1037/0033-2909.129.1.10 [DOI] [PubMed] [Google Scholar]
  • 16. Peplinski B, McClelland R, Szklo M. Associations between socioeconomic status markers and depressive symptoms by race and gender: results from the Multi‐Ethnic Study of Atherosclerosis (MESA). Ann Epidemiol. 2018;28(8):535‐542. doi: 10.1016/j.annepidem.2018.05.005 [DOI] [PubMed] [Google Scholar]
  • 17. Williams DR, González HM, Neighbors H, et al. Prevalence and distribution of major depressive disorder in African Americans, Caribbean Blacks, and Non‐Hispanic Whites: results from the national survey of American Life. Arch Gen Psychiatry. 2007;64(3):305. doi: 10.1001/archpsyc.64.3.305 [DOI] [PubMed] [Google Scholar]
  • 18. Gavin AR, Walton E, Chae DH, Alegria M, Jackson JS, Takeuchi D. The associations between socio‐economic status and major depressive disorder among Blacks, Latinos, Asians and non‐Hispanic Whites: findings from the Collaborative Psychiatric Epidemiology Studies. Psychol Med. 2010;40(1):51‐61. doi: 10.1017/S0033291709006023 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19. Barnes DM, Keyes KM, Bates LM. Racial differences in depression in the United States: how do subgroup analyses inform a paradox? Soc Psychiatry Psychiatr Epidemiol. 2013;48(12):1941‐1949. doi: 10.1007/s00127-013-0718-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20. Hayes‐Larson E, Mobley TM, Mungas D, et al. Accounting for lack of representation in dementia research: generalizing KHANDLE study findings on the prevalence of cognitive impairment to the California older population. Alzheimers Dement. 2022;18(11):2209‐2217. doi: 10.1002/alz.12522 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21. George KM, Gilsanz P, Peterson RL, et al. Impact of cardiovascular risk factors in adolescence, young adulthood, and midlife on late‐life cognition: study of healthy aging in African Americans. J Gerontol A Biol Sci Med Sci. 2021;76(9):1692‐1698. doi: 10.1093/gerona/glab143 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22. Hernandez Saucedo H, Whitmer RA, Glymour M, et al. Measuring cognitive health in ethnically diverse older adults. J Gerontol B Psychol Sci Soc Sci. 2022;77(2):261‐271. doi: 10.1093/geronb/gbab062 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23. Mungas D, Shaw C, Hayes‐Larson E, et al. Cognitive impairment in racially/ethnically diverse older adults: accounting for sources of diagnostic bias. Alzheimers Dement (Amst). 2021;13(1):e12265. doi: 10.1002/dad2.12265 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24. Mungas D, Reed BR, Haan MN, González H. Spanish and English neuropsychological assessment scales: relationship to demographics, language, cognition, and independent function. Neuropsychology. 2005;19(4):466‐475. doi: 10.1037/0894-4105.19.4.466 [DOI] [PubMed] [Google Scholar]
  • 25. Crane PK, Narasimhalu K, Gibbons LE, et al. Item response theory facilitated cocalibrating cognitive tests and reduced bias in estimated rates of decline. Journal of Clinical Epidemiology. 2008;61(10):1018‐1027.e9. doi: 10.1016/j.jclinepi.2007.11.011 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26. Crane PK, Narasimhalu K, Gibbons LE, et al. Composite scores for executive function items: demographic heterogeneity and relationships with quantitative magnetic resonance imaging. J Int Neuropsychol Soc. 2008;14(5):746‐759. doi: 10.1017/S1355617708081162 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27. Ward JB, Gartner DR, Keyes KM, Fliss MD, McClure ES, Robinson WR. How do we assess a racial disparity in health? Distribution, interaction, and interpretation in epidemiological studies. Ann Epidemiol. 2019;29:1‐7. doi: 10.1016/j.annepidem.2018.09.007 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28. Tennant PWG, Arnold KF, Ellison GTH, Gilthorpe MS. Analyses of “change scores” do not estimate causal effects in observational data. Int J Epidemiol. 2022;51(5):1604‐1615. doi: 10.1093/ije/dyab050 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29. Altman DG, Bland JM. Interaction revisited: the difference between two estimates. BMJ. 2003;326(7382):219. doi: 10.1136/bmj.326.7382.219 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30. Adkins‐Jackson PB, Kraal AZ, Hill‐Jarrett TG, et al. Riding the merry‐go‐round of racial disparities in ADRD research. Alzheimer's & Dementia. 2023;19(10):4735‐4742. doi: 10.1002/alz.13359 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31. Rothman KJ. No adjustments are needed for multiple comparisons. Epidemiology. 1990;1(1):43‐46. [PubMed] [Google Scholar]
  • 32. Harel O, Mitchell EM, Perkins NJ, et al. Multiple imputation for incomplete data in epidemiologic studies. Am J Epidemiol. 2018;187(3):576‐584. doi: 10.1093/aje/kwx349 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33. Rubin DB. Multiple Imputation for Nonresponse in Surveys. 1st ed. Wiley; 1987. doi: 10.1002/9780470316696 [DOI] [Google Scholar]
  • 34. Chen R, Calmasini C, Swinnerton K, et al. Pragmatic approaches to handling practice effects in longitudinal cognitive aging research. Alzheimers Dement. 2023;19(9):4028‐4036. doi: 10.1002/alz.13067 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35. Vandenbroucke JP, Von Elm E, Altman DG, et al. Strengthening the Reporting of Observational Studies in Epidemiology (STROBE): explanation and Elaboration. Epidemiology. 2007;18(6):805‐835. doi: 10.1097/EDE.0b013e3181577511 [DOI] [PubMed] [Google Scholar]
  • 36. Hayes‐Larson E, Kezios KL, Mooney SJ, Lovasi G. Who is in this study, anyway? Guidelines for a useful Table 1. Journal of Clinical Epidemiology. 2019;114:125‐132. doi: 10.1016/j.jclinepi.2019.06.011 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37. Palesch YY. Some common misperceptions about P values. Stroke. 2014;45(12):e244‐246. doi: 10.1161/STROKEAHA.114.006138 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38. R Core Team . R: A language and environment for statistical computing. Published online 2019. https://www.r‐project.org
  • 39. Gause E. R code used for the analysis of depression and cognition within the KHANDLE and STAR cohorts. Published online 2023. Accessed October 2, 2023. https://github.com/SpatialHealth/KHANDLE‐STAR_Depression_Cognition
  • 40. Budhwani H, Hearld KR, Chavez‐Yenter D. Depression in racial and ethnic minorities: the impact of nativity and discrimination. J Racial Ethn Health Disparities. 2015;2(1):34‐42. doi: 10.1007/s40615-014-0045-z [DOI] [PubMed] [Google Scholar]
  • 41. Alvarez K, Fillbrunn M, Green JG, et al. Race/ethnicity, nativity, and lifetime risk of mental disorders in US adults. Soc Psychiatry Psychiatr Epidemiol. 2019;54(5):553‐565. doi: 10.1007/s00127-018-1644-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42. Villarroeal MA, Terlizzi EP. Symptoms of Depression Among Adults: United States, 2019. U.S. DEPARTMENT OF HEALTH AND HUMAN SERVICES Centers for Disease Control and Prevention; 2020. Accessed June 7, 2023. https://www.cdc.gov/nchs/data/databriefs/db379‐H.pdf [Google Scholar]
  • 43. Kessler RC, Berglund P, Demler O, et al. The epidemiology of major depressive disorder: results from the National Comorbidity Survey Replication (NCS‐R). JAMA. 2003;289(23):3095. doi: 10.1001/jama.289.23.3095 [DOI] [PubMed] [Google Scholar]
  • 44. Pamplin JR, Bates LM. Evaluating hypothesized explanations for the Black‐white depression paradox: a critical review of the extant evidence. Soc Sci Med. 2021;281:114085. doi: 10.1016/j.socscimed.2021.114085 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 45. Barnes DE, Alexopoulos GS, Lopez OL, Williamson JD, Yaffe K. Depressive symptoms, vascular disease, and mild cognitive impairment: findings from the Cardiovascular Health Study. Arch Gen Psychiatry. 2006;63(3):273‐279. doi: 10.1001/archpsyc.63.3.273 [DOI] [PubMed] [Google Scholar]
  • 46. Gillis JC, Chang SC, Devore EE, Rosner BA, Grodstein F, Okereke OI. Patterns of late‐life depressive symptoms and subsequent declines in cognitive domains. Int J Geriatr Psychiatry. 2017;32(12):1330‐1341. doi: 10.1002/gps.4618 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 47. Zahodne LB, Nowinski CJ, Gershon RC, Manly JJ. Depressive symptoms are more strongly related to executive functioning and episodic memory among African American compared with non‐Hispanic White older adults. Arch Clin Neuropsychol. 2014;29(7):663‐669. doi: 10.1093/arclin/acu045 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 48. Zeki Al Hazzouri A, Caunca MR, Nobrega JC, et al. Greater depressive symptoms, cognition, and markers of brain aging: northern Manhattan Study. Neurology. 2018;90(23):e2077‐e2085. doi: 10.1212/WNL.0000000000005639 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 49. Chao SZ, Matthews BR, Yokoyama JS, et al. Depressive symptoms in Chinese‐American subjects with cognitive impairment. Am J Geriatr Psychiatry. 2014;22(7):642‐652. doi: 10.1016/j.jagp.2012.10.029 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 50. Duan Y, Wei J, Geng W, et al. Research on cognitive function in anxious depression patients in China. J Affect Disord. 2021;280(Pt A):121‐126. doi: 10.1016/j.jad.2020.11.078 [DOI] [PubMed] [Google Scholar]
  • 51. Sharifian N, Sol K, Zaheed AB, et al. Depressive symptoms, leisure activity engagement, and global cognition in non‐hispanic black and white older adults. J Gerontol B Psychol Sci Soc Sci. 2022;77(11):2137‐2147. doi: 10.1093/geronb/gbab153 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 52. Sol K, Zaheed AB, Kraal AZ, Sharifian N, Arce Rentería M, Zahodne LB. Psychological predictors of memory decline in a racially and ethnically diverse longitudinal sample of older adults in the United States. Int J Geriatr Psychiatry. 2020;35(2):204‐212. doi: 10.1002/gps.5236 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 53. LaMonica HM, Biddle DJ, Naismith SL, Hickie IB, Maruff P, Glozier N. The relationship between depression and cognitive function in adults with cardiovascular risk: evidence from a randomised attention‐controlled trial. PLoS One. 2018;13(9):e0203343. doi: 10.1371/journal.pone.0203343 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 54. Office of the Surgeon General (US), Center for Mental Health Services (US), National Institute of Mental Health (US) . Mental Health: Culture, Race, and Ethnicity: A Supplement to Mental Health: A Report of the Surgeon General. Substance Abuse and Mental Health Services Administration (US); 2001. Accessed January 26, 2023. http://www.ncbi.nlm.nih.gov/books/NBK44243/ [PubMed] [Google Scholar]
  • 55. Panza F, Frisardi V, Capurso C, et al. Late‐life depression, mild cognitive impairment, and dementia: possible continuum? Am J Geriatr Psychiatry. 2010;18(2):98‐116. doi: 10.1097/JGP.0b013e3181b0fa13 [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supporting Information

ALZ-20-3147-s002.docx (10MB, docx)

Supporting Information

ALZ-20-3147-s001.pdf (2.5MB, pdf)

Articles from Alzheimer's & Dementia are provided here courtesy of Wiley

RESOURCES