Abstract
Background
Little is known about the protective effects of social support and mental well-being for late-life cognition among different immigrant generations.
Methods
Kaiser Healthy Aging and Diverse Life Experiences participants were categorised as 1st-generation arriving age <18 years (n=73), first-generation arriving ≥18 years (n=282), 2nd-generation (n=279) or ≥3rd-generation (n=174). Social support (emotional, instrumental), loneliness and depressive symptoms were assessed at baseline. Verbal episodic memory (VEM) and executive function (EF) were assessed up to four times (max. years=6.6). Linear mixed-effects models examined associations of social support, loneliness and depression with EF and VEM, adjusting for covariates overall and in race/ethnic-stratified models. Interactions by immigrant generation were tested.
Results
First-generation immigrants arriving <18 years old had the lowest instrumental support (mean (SD)=−0.18 (1.0)) and the highest loneliness (mean (SD)=0.25 (0.93)) and depressive symptom (mean (SD)=−0.04 (0.80)) scores. Instrumental (β=0.05 (95% CI 0.003 to 0.10)) support and emotional support (β=0.06 (95% CI 0.01 to 0.11)) were positively associated with baseline EF. Loneliness (β=−0.08 (95% CI −0.13 to –0.03)) and depressive symptoms (β=−0.09 (95% CI −0.15 to –0.04)) were negatively associated with baseline EF. For associations with VEM, instrumental support (β=0.02 (95% CI -0.04 to 0.07)) and emotional support (β=0.03 (95% CI −0.03 to 0.09)) were not significantly associated. Loneliness (β=−0.07 (95% CI −0.13 to –0.02)) was negatively associated and depression (β=−0.04 (95% CI −0.11 to 0.03)) trended negatively with VEM. All associations with cognitive change were null. In race/ethnic stratified models, associations were more pronounced for Latino participants, whereas associations for Asian participants were generally weaker and non-significant.
Conclusions
Social support and mental well-being may be most pertinent for late-life EF among older Latinos, as associations were generally stronger among Latino adults and closer to null among Asian adults. These findings reiterate the importance of considering both immigrant generation and race/ethnicity in evaluations of late-life cognition risk and resilience factors.
Keywords: Age Factors, Sociodemographic Factors, Mental Health, Social Cohesion, statistics and numerical data
WHAT IS ALREADY KNOWN ON THIS TOPIC
Prior research on immigrant populations has found variation in late-life cognitive outcomes, yet little is known about the specific risk and resilience factors that may explain these relationships. Social support is widely recognised as a protective factor for healthy ageing, and this study aimed to build on prior research in the Kaiser Healthy Aging and Diverse Life Experiences Study to assess its associations of social support and mental well-being with cognitive function in late life varied by immigrant generational status.
WHAT THIS STUDY ADDS
This study confirms the link between social support/mental well-being and late-life cognition and adds a more nuanced understanding of heterogeneity by highlighting more pronounced differences among Latinos than Asians and by immigrant generation.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
The study findings underscore the importance of examining distinct life course pathways for understanding risk and resilience for cognitive ageing outcomes. It informs future public health interventions tailored to these specific challenges, such as acculturation stress, social networks and socioeconomic mobility. Our study findings could also support policies that allocate resource types by immigrant generations to provide assistance where people are most vulnerable (ie, instrumental support for first generation and mental health among third generation immigrants) and culturally and generation-sensitive approaches to cognitive health-associated care.
Introduction
With increasing heterogeneity among the population of US older adults, it is important to understand how different life course experiences may contribute to disparities in late-life cognition. Prior research has found variation in cognitive outcomes among immigrant groups,1,3 but little is known about the specific risk and resilience factors that may help to explain these relationships.
Social support is widely recognised for its positive impact on healthy ageing, including cognitive function and brain health, and may buffer against some of the stressful experiences faced by immigrant groups, including processes of acculturation4,6 and discrimination.7 8 Social support encompasses various forms of assistance that may independently serve as a protective factor for late-life cognition.9,12 Specifically, instrumental support—or the availability of material or functional support—and emotional support—receiving empathy and having a confidante within one’s social network—may uniquely contribute to late-life cognition.13 Conversely, loneliness and depression—two risk factors for poor cognitive ageing—may be experienced differently among different immigrant generations and diaspora, potentially diminishing the positive benefits of social support.914,16
We assessed associations of emotional support, instrumental support, loneliness and depression with cognitive function and decline after age 65 among Asian and Latino diaspora in the USA and examined if these associations varied by immigrant generational status (1st, 2nd or ≥3). We employed the US Census definition of immigrant generation, where first-generation immigrants are those born outside of the USA, second-generation are those with at least one parent born outside of the USA and third-generation or higher includes those with two parents who are native to the USA.17 We hypothesised higher levels of social support and lower levels of loneliness and depression would associate with improved cognitive function and slower cognitive decline in all groups, and that these effects would be most pronounced among first-generation immigrants. Prior research has identified links between mental well-being, immigrant generational status and age at immigration, with first-generation immigrants often facing acute acculturative stress, language barriers, discrimination and disruptions to their social networks—due in part to the recency of family presence in the USA. All of these acculturation processes increase risk for loneliness and depression,14 16 18 while social support has been shown to buffer against acculturation processes among immigrant populations.4,6 These factors may, in turn, exacerbate cognitive impacts of these mental health challenges. If first-generation immigrants have smaller social networks, a greater variety of chronic stressors and less cognitive reserve compared with second-generation and third-generation immigrants, the impact of depressive symptoms, loneliness and lower social support for late-life cognitive function could be amplified. On the other hand, later-generation immigrants may have more robust social networks, acquired more language fluency, have more years of educational attainment in the USA and more comfort navigating US social systems, which may lessen any adverse cognitive consequences of depression, loneliness and lower social support.
Methods
Study sample
Kaiser Healthy Aging and Diverse Life Experiences (KHANDLE) is a longitudinal cohort composed of community-dwelling older adults residing in the San Francisco Bay and Sacramento areas of California that was initiated in 2017. Individuals eligible for KHANDLE are long-term members of Kaiser Permanente Northern California, an integrated healthcare delivery system, aged 65 years or older who speak English or Spanish and previously participated in Kaiser Permanente voluntary health check-up exams between 1964 and 1985. Stratified random sampling is used with the goal of recruiting approximately equal proportions of Asian, Black, Latino and White participants and diversity in educational attainment. Kaiser members were excluded from the sampling frame if they had an electronic medical record diagnosis of dementia or other neurodegenerative disease, or presence of other severe health conditions that would impede participation including hospice activity, history of severe chronic obstructive pulmonary disease, congestive heart failure hospitalisations and end stage renal disease. In March 2020, both the survey and cognitive assessment were transitioned from in-person to telephone due to the COVID-19 pandemic and semantic memory was no longer assessed due to its reliance on visual materials. Following the COVID-19 pandemic, participants were provided with the option of in-person versus phone interview.
Measures
Cognition: Verbal episodic memory, executive function and semantic memory were assessed up to four times at approximately 18-month intervals using the Spanish and English Neuropsychological Assessment Scales (SENAS). Administration procedures, development and psychometric characteristics have been described in detail elsewhere.19 SENAS was administered in either English or Spanish, based on the participant’s preference. Each cognitive domain was z-standardised using the wave 1 mean and SD from the combined KHANDLE and STAR cohorts.
Immigrant generational status was categorised into four groups. First-generation immigrants were defined as those participants born outside the USA, which is also known as being ‘foreign born” .17 Following Meyer et al, we differentiated first-generation immigrants who arrived in the USA before age 18 from those who immigrated after to capture the potential influence of acculturation through schooling in the USA. Second-generation immigrants were born in the USA to at least one foreign-born parent .1 Third-generation immigrants and beyond were those for whom both parents were also born in the USA. Participants’ country of birth was derived through self-reporting. When applicable, age at immigration to the USA was also obtained through self-report.
Baseline perceived emotional and instrumental support and loneliness were self-reported using the NIH PROMIS Toolbox Emotion Battery.20,23 The NIH PROMIS instruments were developed using Item Response Theory (IRT) and have been extensively validated in diverse populations.24 25 Instrumental support assessed perceived access to practical help (eg, financial assistance, chores, caregiving). Emotional support assessed perceived availability of comfort, understanding and companionship from family, friends or peers. Loneliness assessed perceived social isolation. Depression symptoms were assessed using the clinically valid PROMIS Depression Instrument for Adults.23 26 27 Scores for each measure were standardised (mean=0, SD=1) to the US adult population (18+).
Covariates: Age and sex were collected at wave 1 or imputed from electronic medical records if missing. Race was categorised as Asian, Black, Latino or White. Years of education were ascertained from two interview questions: highest completed grade level (0–12); highest level of schooling or degree completed. For those with education beyond a high school diploma or General Education Development (GED) credential (12 years), years were assigned as follows: some college or trade school (13), associate’s degree (14), bachelor’s degree (16), master’s degree (18), doctorate (20). Late-life income was self-reported in 13 categories and recoded as ≤US$54 999, US$55 000–US$99 000 and ≥US$100 000. Childhood socioeconomic status was captured via parental educational attainment. Participants reported their mother’s and father’s highest completed year (0–12) or level of education (college but no degree, associate’s degree, bachelor’s degree, master’s degree or doctoral degree). We applied the same coding scheme as participant education. To retain participants with missing data (maternal education: 18%, paternal: 23%), missing values were coded as 0 as a missingness indicator and included in all models.
Analytic sample
Due to low numbers of Black and White first-generation participants, we restricted analyses to Asian and Latino participants. Of 836 participants, we excluded 15 for missing scores for social support, loneliness and/or depression, 10 for missing immigration status, 2 for missing all cognitive assessments and 1 for missing education, yielding a final analytic sample of 808. Within this sample, 73 participants were first-generation immigrants who arrived in the USA before age 18, while 282 immigrated at age 18 or older. Additionally, 280 participants were second-generation and 173 were third-generation.
Statistical analysis
We examined the correlation between measures of social support and mental well-being and summarised participant characteristics overall, stratified by immigrant generation and stratified by race/ethnicity. To account for the repeated measures within our data and accommodate different number of follow-ups across individuals, we fit a series of linear mixed effects models for each cognitive outcome using years between assessments as timescale. Separate models were fit to assess associations of each measure of social support/well-being (instrumental support, emotional support, loneliness and depression) with executive function and verbal episodic memory at baseline (model set 1) and longitudinally by adding a cross-level interaction with time (model set 2). Next, we tested for interactions (p<0.10) between immigrant generation (1st, 2nd and ≥3rd generation) and social support/well-being variables (model set 3). Given the distribution of immigrant generation was not equivalent by race/ethnicity, we refit model sets 1 and 2 stratified by race/ethnicity as a sensitivity analysis. All models adjusted for baseline age, race/ethnicity (except race-stratified models), sex, educational attainment, maternal and paternal education, missing maternal and paternal education, language of SENAS administration and mode of interview (phone vs in-person). Analyses were performed in Stata V.17.0 (StataCorp).
Results
Measures of late-life social support (emotional and instrumental) and mental well-being (depression and loneliness) were moderately correlated (range: −0.41 to 0.64; online supplemental table 1). Within the analytic sample (N=808), participants’ mean age was 75.7 (SD=6.8), 57% were female, and 51% had at least a bachelor’s degree (table 1). Maternal and paternal education was highest among ≥3rd-generation immigrants (maternal mean (SD)=9.8 (4.8) years; paternal mean (SD)=10.5 (5.4) years).
Table 1. Characteristics of the analytic sample of KHANDLE participants.
| Overall (n=808) | 1st generation (age immigrated <18 years) (n=73) | 1st generation (age immigrated >18 years) (n=282) | 2nd generation (n=279) | 3rd generation (n=174) | |
|---|---|---|---|---|---|
| Age (mean, SD) | 75.73±6.76 | 74.22±6.75 | 75.75±6.06 | 76.95±7.72 | 74.36±5.80 |
| Sex (n, %) | |||||
| Male | 351 (44) | 34 (47) | 118 (42) | 135 (48) | 64 (37) |
| Female | 457 (57) | 39 (53) | 164 (58) | 144 (51) | 110 (63) |
| Race/ethnicity (n, %) | |||||
| Asian | 429 (53) | 43 (59) | 161 (57) | 153 (55) | 72 (41) |
| Latino | 379 (47) | 30 (41) | 121 (43) | 126 (45) | 102 (59) |
| Educational attainment (mean, SD) | 14.66±3.22 | 15.01±2.37 | 13.93±4.03 | 15.09±2.61 | 15.02±2.68 |
| <College | 400 (49%) | 33 (45%) | 150 (53%) | 133 (48%) | 84 (48%) |
| ≥College | 408 (51%) | 40 (55%) | 132 (47%) | 146 (52%) | 90 (52%) |
| Maternal education (mean, SD) | 7.40±5.53 | 6.48±5.32 | 6.59±5.58 | 6.95±5.55 | 9.83±4.81 |
| Missing maternal education (n,%) | 142 (18%) | 17 (23%) | 49 (17%) | 60 (22%) | 16 (9%) |
| Paternal education (mean, SD) | 8.11±6.45 | 9.11±6.97 | 8.27±6.75 | 6.20±6.04 | 10.51±5.42 |
| Missing paternal education (n, %) | 190 (23%) | 15 (21%) | 64 (23%) | 92 (33%) | 16 (11%) |
| Instrumental support (mean, SD)* | −0.001 (0.98) | −0.18 (1.00) | −0.01 (0.94) | −0.05 (1.00) | 0.16 (1.01) |
| Emotional support (mean, SD)* | −0.45 (0.99) | −0.56 (0.99) | −0.61 (1.04) | −0.43 (0.93) | −0.19 (0.96) |
| Depression (mean, SD)* | −0.09 (0.78) | −0.04 (0.80) | −0.07 (0.87) | −0.11 (0.71) | −0.12 (0.73) |
| Loneliness (mean, SD)* | 0.05 (0.97) | 0.25 (0.93) | 0.05 (1.06) | 0.05 (0.93) | −0.02 (0.88) |
Calculated as a theta score standardised to the US adult population ages 18 and older.
KHANDLE, Kaiser Healthy Aging and Diverse Life Experiences.
When comparing descriptively across immigrant generations, first-generation immigrants—including those <18 and ≥18 at immigration—reported the lowest levels of late-life emotional support and mental well-being. Those arriving before age 18 scored lowest in instrumental support (mean (SD)=−0.18 (1.0)) and highest in loneliness (mean (SD)=0.25 (0.93)) and depressive symptoms (mean (SD)=−0.04 (0.80)). First-generation immigrants who arrived after age 18 scored lowest in emotional support (mean (SD)=−0.56(0.99)). Among race/ethnic subsamples, we observed that 48% of Asian participants were 1st-generation immigrants and 17% were ≥3rd-generation immigrants. By contrast, 40% of Latinos were 1st-generation and 27% were ≥3rd-generation immigrants.
When examining levels of social support and well-being within each race/ethnic subsample, we observed distinct generation-related patterns. Specifically, among first-generation participants who arrived <18, we observed lower instrumental support among Asian participants than Latino participants (Asian participant mean=−0.23 vs Latino participant mean=−0.11); in contrast, among third-generation immigrants, we observed higher instrumental support scores among Asian versus Latino participants (Asian mean=0.30 vs Latino mean=0.05). Among first-generation participants who arrived <age 18, Latinos scored higher on depressive symptoms (Latino mean=0.16 vs Asian participant mean=−0.18) and loneliness (mean=0.39 vs Asian participant mean=0.15), though we observed limited differences in these scores between Asian participants and Latino participants who were ≥3rd-generation (online supplemental table 2).
Associations of social support and mental well-being with cognition
In adjusted linear mixed effects models, we observed positive associations between instrumental (β=0.05 (95% CI 0.003 to 0.10)) and emotional support (β=0.06 (95% CI 0.01 to 0.11)) with baseline executive function, and negative associations between loneliness (β=−0.08 (95% CI −0.13 to –0.03)) and depressive symptoms (β=−0.09 (95% CI −0.15 to –0.04); table 2). Associations with change in executive function over time were null (instrumental support β=0.01 (95% CI −0.01 to 0.02); emotional support (β=0.004 (95% CI −0.01 to 0.02); loneliness β=−0.003 (95% CI −0.02 to 0.01); depression β=−0.01 (95% CI −0.02 to 0.01)).
Table 2. Associations of social support and mental well-being with late-life cognition among 1st, 2nd and 3rd+ generation immigrants in the Kaiser Healthy Aging and Diverse Life Experiences cohort.
| Executive function | Verbal episodic memory | |||
|---|---|---|---|---|
| Estimate | 95% CI | Estimate | 95% CI | |
| Instrumental support | 0.05 | 0.003 to 0.10 | 0.02 | −0.04 to 0.07 |
| Emotional support | 0.06 | 0.01 to 0.11 | 0.03 | −0.03 to 0.09 |
| Loneliness | −0.08 | −0.13 to –0.03 | −0.07 | −0.13 to –0.02 |
| Depression | −0.09 | −0.15 to –0.04 | −0.04 | −0.11 to 0.03 |
| Years | −0.03 | −0.04 to –0.02 | −0.06 | −0.08 to –0.05 |
| Instrumental support*years | 0.01 | −0.01 to 0.02 | 0.003 | −0.01 to 0.02 |
| Emotional support*years | 0.004 | −0.01 to 0.02 | −0.004 | −0.02 to 0.01 |
| Loneliness*years | −0.003 | −0.02 to 0.01 | 0.002 | −0.01 to 0.02 |
| Depression*years | −0.01 | −0.02 to 0.01 | −0.003 | −0.01 to 0.02 |
Separate mixed effects models were estimated for each social support/mental well-being exposure. All models adjusted for baseline age, sex, race/ethnicity, participant education, parental education and mode of interview (phone vs in-person) and use years between cognitive assessments as timescale.
Baseline effect estimates were from models that did not include a cross-level interaction with years between cognitive assessments.
In models estimating associations with baseline verbal episodic memory, higher levels of instrumental support (β=0.02 (95% CI −0.04 to 0.07)) and higher levels of emotional support were non-significantly associated with baseline verbal episodic memory (β=0.03 (95% CI −0.03 to 0.09)). Higher levels of loneliness were significantly associated with lower baseline verbal episodic memory (β=−0.07 (95% CI −0.13 to –0.02)). Lower depressive symptom scores trended towards better baseline verbal episodic memory (β=−0.04 (95% CI −0.11 to 0.03)). Associations with verbal episodic memory change over time were non-significant and null (instrumental support β=0.003 (95% CI −0.01 to 0.02); emotional support β=−0.004 (95% CI −0.02 to 0.01); loneliness β=0.002 (95% CI −0.01 to 0.02); depressive symptoms β=−0.003 (95% CI −0.01 to 0.02)).
Effect modification by immigrant generational status
In models testing for differences in adjusted associations by immigrant generation, we observed a significant interaction between immigrant generation and instrumental support in association with baseline verbal episodic memory (p=0.06). Specifically, among first-generation immigrants, having higher instrumental support was marginally associated with lower baseline verbal episodic memory (est.=−0.06 (95% CI −0.14 to 0.03)), while among second-generation (est.=0.09 (95% CI −0.003 to 0.18)) and third-generation (est.=0.04 (95% CI −0.07 to 0.16)) immigrants, having higher instrumental support was marginally associated with better baseline verbal episodic memory (figure 1). We also observed a significant interaction (p=0.09) between depressive symptoms and generational status with executive function change. To further explore this three-way interaction (immigrant-generation*depressive symptoms*time), we dichotomised depressive symptoms at the standardised national mean=0 and estimated marginal effects for each immigrant generation. Among third-generation immigrants, we found those with high depressive symptoms had a faster rate of executive function decline (est=−0.07 (95% CI −0.11 to –0.04)) compared with third-generation immigrants who had lower depressive symptoms (est.=−0.01 (95% CI −0.04 to 0.02)). We did not see any differences between those with higher versus lower depressive symptoms among first- or second-generation immigrants (figure 2). All other interactions between immigrant generation and social support or mental well-being did not reach significance at p<0.10 (online supplemental table 3).
Figure 1. Estimated baseline verbal episodic memory z-score and 95% CI across levels of instrumental support (theta score), stratified by first-generation, second-generation and third-generation immigrant status in the KHANDLE cohort. KHANDLE, Kaiser Healthy Aging and Diverse Life Experiences.
Figure 2. Estimated rate of executive function z-score change over time (years) and 95% CI among those with higher versus lower (ref. y=0) depressive symptoms, stratified by first-generation, second-generation and third-generation immigrant status.
Race/ethnicity-stratified associations of social support and mental well-being with cognition
Among Asians, associations of instrumental support (β=0.04 (95% CI −0.03 to 0.11)) and emotional support (β=0.02 (95% CI −0.05 to 0.09)) with baseline executive function were positive (table 3). Associations with loneliness (β=−0.04 (95% CI −0.11 to 0.03)) and depressive symptoms (β=−0.05 (95% CI −0.13 to 0.03) with baseline executive function were negative. All associations with changes in executive function over time were null.
Table 3. Associations of social support and mental well-being with late-life cognition among 1st, 2nd and 3rd+ generation immigrants in the Kaiser Healthy Aging and Diverse Life Experiences cohort, stratified by race/ethnicity.
| Executive function | Verbal episodic memory | |||
|---|---|---|---|---|
| Estimate | 95% CI | Estimate | 95% CI | |
| Asian participants (n=429) | ||||
| Instrumental support | 0.04 | −0.03 to 0.11 | −0.01 | −0.10 to 0.07 |
| Emotional support | 0.02 | −0.05 to 0.09 | 0.03 | −0.05 to 0.12 |
| Loneliness | −0.04 | −0.11 to 0.03 | −0.05 | −0.13 to 0.03 |
| Depression | −0.05 | −0.13 to 0.03 | 0.001 | −0.10 to 0.10 |
| Years | −0.03 | −0.05 to –0.02 | −0.08 | −0.10 to –0.06 |
| Instrumental support*years | −0.01 | −0.02 to 0.01 | −0.003 | −0.02 to 0.02 |
| Emotional support*years | −0.001 | −0.02 to 0.01 | 0.01 | −0.02 to 0.03 |
| Loneliness*years | 0.001 | −0.02 to 0.02 | −0.003 | −0.03 to 0.02 |
| Depression*years | −0.004 | −0.02 to 0.02 | −0.01 | −0.04 to 0.02 |
| Latino participants (n=379) | ||||
| Instrumental support | 0.03 | −0.04 to 0.10 | 0.03 | −0.05 to 0.10 |
| Emotional support | 0.10 | 0.03 to 0.16 | 0.01 | −0.06 to 0.09 |
| Loneliness | −0.09 | −0.16 to –0.03 | −0.09 | −0.16 to –0.01 |
| Depression | −0.14 | −0.23 to –0.05 | −0.09 | −0.19 to 0.003 |
| Years | −0.02 | −0.04 to –0.01 | −0.04 | −0.06 to –0.01 |
| Instrumental support*years | 0.02 | −0.002 to 0.04 | 0.01 | −0.01 to 0.03 |
| Emotional support*years | 0.01 | −0.01 to 0.03 | −0.02 | −0.04 to –0.01 |
| Loneliness*years | −0.003 | −0.02 to 0.02 | 0.01 | −0.02 to 0.03 |
| Depression*years | −0.02 | −0.04 to 0.01 | −0.001 | −0.03 to 0.03 |
Separate mixed effects models were estimated for each social support/mental well-being exposure. All models adjusted for baseline age, sex, race/ethnicity, participant education, parental education and mode of interview (phone vs in-person) and use years between cognitive assessments as timescale.
Baseline effect estimates were from models that did not include a cross-level interaction with years between cognitive assessments.
Also among Asians, the association of instrumental support (β=−0.01 (95% CI −0.10 to 0.07)) with baseline verbal episodic memory was null; the association of emotional support (β=0.03 (95% CI −0.05 to 0.12)) was positive; the association of loneliness (β=−0.05 (95% CI −0.13 to 0.03)) was negative; and the association of depressive symptoms with baseline verbal episod memory was null (β=0.001 (95% CI −0.10 to 0.10)). All associations with verbal episodic memory change over time were null.
Among Latino participants, associations of instrumental support (β=0.03 (95% CI −0.04 to 0.10)) and emotional support (β=0.10 (95% CI 0.03 to 0.16)) with baseline executive function were positive. Associations of loneliness (β=−0.09 (95% CI −0.16 to –0.03)) and depressive symptoms (β=−0.14 (95% CI −0.23 to –0.05)) with baseline executive function were negative. All associations with executive function change were null.
Also among Latinos, associations of instrumental support (β=0.03 (95% CI −0.05 to 0.10)) and emotional support (β=0.01 (95% CI −0.06 to 0.09)) with verbal episodic memory were positive. Associations of loneliness (β=−0.09 (95% CI −0.16 to –0.01)) and depressive symptoms (β=−0.09 (95% CI −0.19 to 0.003)) with baseline verbal episodic memory were negative. All associations with verbal episodic memory change among Latino participants were null.
Discussion
In our sample of Asian and Latino older adults, we found third-generation participants reported higher levels of social support and lower depression and loneliness, while first-generation immigrants reported the highest levels of depression and loneliness. Adjusted models confirmed our hypothesised patterns of higher instrumental and emotional support with higher baseline cognitive function, and higher loneliness and depression with lower baseline cognitive function. In race/ethnicity-stratified models, these patterns of association were more pronounced among Latinos, but moved closer to the null among Asians. We also observed interactions by immigrant generation in the association of instrumental support with baseline verbal episodic memory and of depressive symptoms with executive function decline.
In contrast to our hypothesis, having higher instrumental support was not protective for baseline verbal episodic memory among first-generation immigrants. This may be because the types of instrumental support (eg, financial assistance; caregiving) available within the social networks of first-generation immigrants systematically differ from second-generation immigrants because of differences in socioeconomic status, acculturation and access to broader community resources among more recent immigrants. Alternatively, this finding may be an artefact of language proficiency for the cognitive assessment, which was available in Spanish and English, but none of the first languages of the Asian diaspora. While we are underpowered in this analysis to further examine these differences, future studies in other cohorts should explore potential drivers of these differences among first-generation immigrants.
When investigating the significant difference in effect of depression symptoms on executive function decline, we found that having higher versus lower depressive symptoms was associated with more than twice the rate of executive function decline among third-generation immigrants. Specifically, while the overall rate of executive function decline in this sample was 0.03 SD per year, the estimated rate of executive function decline for third-generation immigrants with higher depressive symptoms was 0.07 SD per year. Importantly, this finding is not likely driven by differences in depression symptom prevalence, which were close to the US population mean and had differences of less than one-tenth of a SD across immigrant generation.
Our findings build on prior work in KHANDLE1 2 by highlighting the potential role of social support and mental well-being in understanding immigrant generational differences in late-life cognition. The observed differences in this study by race/ethnicity and immigrant generation may reflect different lived experiences with regards to stresses of assimilation and discrimination. In our sample, first-generation immigrants were more likely to be Asian, while third-generation immigrants were more likely to be Latino. Prior theoretical work on segmented assimilation has suggested that immigration experiences may vary by diaspora group and individual characteristics.28,30 Segmented assimilation theory highlights the importance of parental human capital (including education and income), modes of incorporation (eg, degree of discrimination), family structure and transnational ties among immigrant groups as contributors to upward or downward assimilation and economic mobility of immigrants,31 which may have long-term implications for late-life cognition.
Our study also supports findings from similar research on cognition among immigrants, showing associations between age at migration,32 33 social support9,1134 and mental health12 35 with cognitive outcomes. Of note, Ge et al10 found social support and social strain have mixed effects on cognition in late life among immigrants. Like our study, which showed lower instrumental support non-significantly associated with higher verbal episodic memory in second- and third-generation immigrants, Ge et al found elevated social strain was associated with higher cognitive function among older US Chinese.10 These results emphasise the need to differentiate sources of social support and social strain in analyses of late-life cognition among immigrants and their varying effects across immigrant categories.
The strengths of this study are its contributions to growing literature seeking to understand health and ageing among an increasingly heterogeneous population. While previous studies have explored associations between social support, loneliness, depression and late-life cognition among immigrants, our investigation provides new contributions to this literature by explicitly examining these associations across both age and generation of immigration. Though most associations were non-significant, diverging effects across first-generation, second-generation and third-generation immigrants may point to differences in how social support is experienced across immigrant generations.
We also note some limitations. The longitudinal data were unable to control for time-varying confounders of social support and mental health throughout the life course. Instead, self-reported social support and mental health measures were captured at the baseline interview, when they may already be changing alongside cognitive function. Additionally, we lack measures of early life and midlife levels of social support, depression and isolation, as well as early experiences that may exacerbate (eg, assimilation, discrimination) well-being, which may influence outcomes in later life. Also, while differences in depression trends were observed across immigrant generations, future studies that compare these generational differences with population-based data or national trends could reveal the meaningfulness of these results, including whether these trends are specific to the study population or more broadly observed on a national level. While SENAS cognitive assessments are equivalent across English and Spanish, the assessments cannot be completed in any Asian languages, which may contribute to linguistic or cultural bias among Asian participants. However, all Asian participants in KHANDLE must be sufficiently fluent in English to participate successfully in all portions of the interview—a trade-off that may limit generalisability but minimises the linguistic bias in cognitive assessment. Finally, this Northern California-based cohort may be limited in generalisability to other regions of the USA. Specifically, recruited participants have had long-term healthcare access, which may positively influence cognitive outcomes. Additionally, individuals with dementia or major chronic conditions were excluded from the baseline sampling frame, which may bias findings toward healthier participants—especially among those who initiated the study at age 80 and older. Replicating this study in a nationally representative sample with longer follow-up would help to confirm our study findings. Additionally, individuals with dementia or major chronic conditions were excluded from the baseline sampling frame, which may bias findings toward healthier participants—especially among those who initiated the study at age 80 and older. Replicating this study in a nationally representative sample with longer follow-up would help to confirm our study findings.
Conclusions
Our study findings support the need for public health practitioners to attune interventions to the differing needs of older adults of different immigrant generations, such as addressing acculturation stress, enhancing social networks and supporting socioeconomic mobility. Public health policies and programmes targeting low instrumental support for first-generation immigrants and high depressive symptoms for third-generation immigrants may be important intervention targets to improve mental well-being, enhance social support networks and protect late-life cognition. Additionally, we suggest future research comparing the KHANDLE cohort with national health statistics on mental health and well-being to contextualise the findings and assess whether depressive symptoms are elevated in this population or reflect broader trends observed across immigrant generations.
Supplementary material
Footnotes
Funding: Support for this study comes from NIH/NIA: R00AG073457 (PI: RLP); R01AG052132 (PIs: RAW, PG, MG, Mayeda); R01AG050782 (PIs: RAW, PG).
Provenance and peer review: Not commissioned; externally peer reviewed.
Patient consent for publication: Not applicable.
Patient and public involvement: Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.
Data availability statement
Data may be obtained from a third party and are not publicly available.
References
- 1.Meyer OL, Eng CW, Ko MJ, et al. Generation and age of immigration on later life cognitive performance in KHANDLE. Int Psychogeriatr. 2023;35:17–28. doi: 10.1017/S1041610220003774. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Peterson RL, George KM, Gilsanz P, et al. Lifecourse socioeconomic changes and late-life cognition in a cohort of U.S.-born and U.S. immigrants: findings from the KHANDLE study. BMC Public Health. 2021;21:920. doi: 10.1186/s12889-021-10976-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Xu H, Zhang Y, Wu B. Association between migration and cognitive status among middle-aged and older adults: a systematic review. BMC Geriatr. 2017;17:184. doi: 10.1186/s12877-017-0585-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Kim BJ, Sangalang CC, Kihl T. Effects of acculturation and social network support on depression among elderly Korean immigrants. Aging Ment Health. 2012;16:787–94. doi: 10.1080/13607863.2012.660622. [DOI] [PubMed] [Google Scholar]
- 5.Li LW, McLaughlin SJ, Zhang J. Healthy Aging in Older Chinese Americans: Associations With Immigrant Experiences. J Aging Health. 2020;32:1098–108. doi: 10.1177/0898264319889122. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Sung JH, Lee JE, Lee J-Y. Effects of Social Support on Reducing Acculturative Stress-Related to Discrimination between Latin and Asian Immigrants: Results from National Latino and Asian American Study (NLAAS) J Adv Med Med Res. 2018;27:1–10. doi: 10.9734/JAMMR/2018/42728. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Cariello AN, Perrin PB, Williams CD, et al. Moderating Influence of Social Support on the Relations between Discrimination and Health via Depression in Latinx Immigrants. J Lat Psychol. 2022;10:98–111. doi: 10.1037/lat0000200. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Szaflarski M, Bauldry S. The Effects of Perceived Discrimination on Immigrant and Refugee Physical and Mental Health. Adv Med Sociol. 2019;19:173–204. doi: 10.1108/S1057-629020190000019009. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Costa-Cordella S, Arevalo-Romero C, Parada FJ, et al. Social Support and Cognition: A Systematic Review. Front Psychol. 2021;12:637060. doi: 10.3389/fpsyg.2021.637060. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Ge S, Wu B, Bailey DE, et al. Social support, social strain, and cognitive function among community-dwelling US Chinese older adults. J Gerontol B Psychol Sci Soc Sci. 2017;72:1030–40. doi: 10.1093/geronb/glw221. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Ohman A, Maxwell CJ, Tyas SL, et al. Subtypes of social support availability are not differentially associated with memory: a cross-sectional analysis of the Comprehensive Cohort of the Canadian Longitudinal Study on Aging. Aging, Neuropsychology, and Cognition. 2023;30:354–69. doi: 10.1080/13825585.2022.2030294. [DOI] [PubMed] [Google Scholar]
- 12.Wang Y, Chen X, Hu Y. Relationship between social support and 7-year trajectories of cognitive decline: results from the China Health and Retirement Longitudinal Study. J Epidemiol Community Health. 2023;77:578–86. doi: 10.1136/jech-2022-219733. [DOI] [PubMed] [Google Scholar]
- 13.Muñoz-Laboy M, Severson N, Perry A, et al. Differential impact of types of social support in the mental health of formerly incarcerated Latino men. Am J Mens Health. 2014;8:226–39. doi: 10.1177/1557988313508303. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Lee J, Hong J, Zhou Y, et al. The Relationships Between Loneliness, Social Support, and Resilience Among Latinx Immigrants in the United States. Clin Soc Work J. 2020;48:99–109. doi: 10.1007/s10615-019-00728-w. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Mogic L, Rutter EC, Tyas SL, et al. Functional social support and cognitive function in middle- and older-aged adults: a systematic review of cross-sectional and cohort studies. Syst Rev. 2023;12:86. doi: 10.1186/s13643-023-02251-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Tang F, Jiang Y, Li K, et al. Residential Segregation and Depressive Symptoms in Older Chinese Immigrants: The Mediating Role of Social Processes. Gerontologist. 2023;63:1376–84. doi: 10.1093/geront/gnad027. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Bureau UC. Washington (DC): US Census Bureau; [20-Jan-2026]. Frequently asked questions about foreign-born.https://www.census.gov/topics/population/foreign-born/about/faq.html Available. Accessed. [Google Scholar]
- 18.Tonui BC, Miller VJ, Adeniji DO. Older immigrant adults experiences with social isolation: a qualitative interpretive meta synthesis. Aging Ment Health. 2023;27:1068–76. doi: 10.1080/13607863.2022.2068131. [DOI] [PubMed] [Google Scholar]
- 19.Mungas D, Reed BR, Crane PK, et al. Spanish and English Neuropsychological Assessment Scales (SENAS): further development and psychometric characteristics. Psychol Assess. 2004;16:347–59. doi: 10.1037/1040-3590.16.4.347. [DOI] [PubMed] [Google Scholar]
- 20.Babakhanyan I, McKenna BS, Casaletto KB, et al. National Institutes of Health Toolbox Emotion Battery for English- and Spanish-speaking adults: normative data and factor-based summary scores. Patient Relat Outcome Meas. 2018;9:115–27. doi: 10.2147/PROM.S151658. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Cyranowski JM, Zill N, Bode R, et al. Assessing social support, companionship, and distress: National Institute of Health (NIH) Toolbox Adult Social Relationship Scales. Health Psychol. 2013;32:293–301. doi: 10.1037/a0028586. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.National Institutes of Health . Bethesda (MD): NIH; 2025. [20-Jan-2026]. Emotion assessments – NIH toolbox.https://nihtoolbox.org Available. Accessed. [Google Scholar]
- 23.Salsman JM, Butt Z, Pilkonis PA, et al. Emotion assessment using the NIH Toolbox. Neurology (ECronicon) 2013;80:S76–86. doi: 10.1212/WNL.0b013e3182872e11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.National Institutes of Health . Bethesda (MD): NIH; [20-Jan-2026]. Validation and norming – NIH toolbox.https://nihtoolbox.org/validation-team/ Available. Accessed. [Google Scholar]
- 25.Mather MA, Ho EH, Bedjeti K, et al. Measuring Multidimensional Aspects of Health in the Oldest Old Using the NIH Toolbox: Results From the ARMADA Study. Arch Clin Neuropsychol. 2024;39:535–46. doi: 10.1093/arclin/acad105. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.National Institutes of Health . Bethesda (MD): NIH; 2025. Sadness/depression – NIH toolbox.https://nihtoolbox.org/test/sadness-depression/ Available. [Google Scholar]
- 27.Schalet BD, Pilkonis PA, Yu L, et al. Clinical validity of PROMIS Depression, Anxiety, and Anger across diverse clinical samples. J Clin Epidemiol. 2016;73:119–27. doi: 10.1016/j.jclinepi.2015.08.036. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Esses VM. Prejudice and Discrimination Toward Immigrants. Annu Rev Psychol. 2021;72:503–31. doi: 10.1146/annurev-psych-080520-102803. [DOI] [PubMed] [Google Scholar]
- 29.Portes A, Zhou M. The New Second Generation: Segmented Assimilation and its Variants. Ann Am Acad Pol Soc Sci. 1993;530:74–96. doi: 10.1177/0002716293530001006. [DOI] [Google Scholar]
- 30.Safi M. Immigration theory between assimilation and discrimination. J Ethn Migr Stud. 2024;50:173–202. doi: 10.1080/1369183X.2023.2207250. [DOI] [Google Scholar]
- 31.Waters MC, Tran VC, Kasinitz P, et al. Segmented Assimilation Revisited: Types of Acculturation and Socioeconomic Mobility in Young Adulthood. Ethn Racial Stud. 2010;33:1168–93. doi: 10.1080/01419871003624076. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Garcia MA, Ortiz K, Arévalo SP, et al. Age of Migration and Cognitive Function Among Older Latinos in the United States. JAD. 2021;76:1493–511. doi: 10.3233/JAD-191296. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Guo M, Li M, Xu H, et al. Age at Migration and Cognitive Health Among Chinese Older Immigrants in the United States. J Aging Health. 2021;33:709–20. doi: 10.1177/08982643211006612. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Du C, Dong X, Katz B, et al. Source of perceived social support and cognitive change: an 8-year prospective cohort study. Aging Ment Health. 2023;27:1496–505. doi: 10.1080/13607863.2022.2126433. [DOI] [PubMed] [Google Scholar]
- 35.Jung M, Kim H, Loprinzi PD, et al. Age-varying association between depression and cognitive function among a national sample of older U.S. immigrant adults: the potential moderating role of physical activity. Aging Ment Health. 2023;27:653–62. doi: 10.1080/13607863.2022.2056139. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
Data may be obtained from a third party and are not publicly available.