Health Risks Posed by Social and Linguistic Isolation in Older Korean Americans

Abstract

Objectives:

Inspired by the notion of double jeopardy positing the dual disadvantage in the coexistence of different sources of vulnerability, the present study examined health risks posed by social and linguistic isolation in older Korean Americans.

Design:

A cross-sectional study, using a four-cell classification to compare isolation types (no isolation, social isolation only, linguistic isolation only, and dual isolation) and to examine their impacts on physical (self-rated health), mental (mental distress), and cognitive (cognitive performance) health.

Setting, Participants:

Data are from the Study of Older Korean Americans (SOKA) survey of Korean immigrants aged 60 or older (n = 2,032), conducted in five states in the U.S. (California, New York, Texas, Hawaii, and Florida).

Measurement:

Social isolation was indexed by the Lubben Social Network Scale−6; linguistic isolation, by a question on English-speaking ability. Physical, mental, and cognitive health were indicated by a single-item self-rating of health, the Kessler Psychological Distress Scale 6, and the Mini-Mental State Examination, respectively.

Results:

The dual-isolation group exhibited notable sociodemographic and health disadvantages. The odds of having fair/poor health, mental distress, and cognitive impairment were 2.21−3.17 times higher in the dual-isolation group than in the no-isolation group.

Conclusion:

Our findings confirm that both social relationships and language proficiency are key elements for older immigrants’ social connectedness and integration, the deprivation of which puts them at heightened risk in multiple dimensions of health.

INTRODUCTION

As a critical social determinant of health and a major component of successful aging, social connectedness plays an integral role in the lives of older adults.^1,2 A sizable body of literature has documented the detrimental effects of being disengaged from family, friends, and communities in individuals’ later years, effects that include higher rates of morbidity and mortality, increased mental health problems, deterred use of health and social services, and diminished quality of life.^3–7 Given that social isolation is a major public health concern that one in five adult Americans face,^3,8 its impact on multiple dimensions of health in diverse groups of older adults deserves attention.

Older immigrants are especially prone to social isolation. About 14% of the U.S. adults age 65 or older are immigrants, a proportion expected to double by 2060.^9,10 What has been called the “broken convoy” effect¹¹ indicates that the experience of immigration makes it difficult for older individuals to maintain their existing social networks and build new relationships in their new host country. Owing to these disrupted and restricted social relationships, older immigrants demonstrate a higher rate of social isolation than do their non-immigrant counterparts, and these social vulnerabilities contribute to their disparities in health.^7,12–15

Another important source of social disconnectedness in the older immigrant population is linguistic isolation. Over 65 million Americans do not speak English as their primary language and have a limited ability to read, speak, write, or understand English.^16,17 English-speaking ability, a key indicator of acculturation, is often considered a personal asset linked with higher socioeconomic status and better access to resources and services.^18,19 Limited English proficiency is a major life stressor that presents challenges to older immigrants’ daily functioning, information-seeking and service use, sense of autonomy and independence, social engagement, and health and well-being.^18–20 The negative impacts of linguistic isolation on health are well documented,^12,21,22 and studies suggest that the significance of one’s English-speaking ability is pronounced among those who reside in areas with a low population density of their own ethnic group (i.e., non-ethnic enclaves).^23–25 It is important to note that ethnic population density in a residential area is a critical contextual factor that influences older immigrants’ needs, desires, and opportunities for acquiring the language of the host society and building social relationships with members of that society.²³ Variations in ethnic population density therefore should be considered when addressing social and linguistic isolation.

Recognizing that older Asian immigrants are both a fast-growing population and encompass more than two dozen ethnic subgroups,¹⁰ we focused on older Korean Americans as a target group. Koreans are the 5th largest Asian American subgroup, and the current population of older Korean Americans consists predominantly of foreign-born first-generation immigrants.²⁷ Valuing collectivism and familism,²⁸ strong social ties and support are important resources for older Korean Americans; a lack of such resources leads to adverse health outcomes.^11,13,14 It is also noteworthy that Korean is the 4th most common language spoken by those in the U.S. population who have limited English proficiency.^16,17 This language barrier is widely known to be a major source of health vulnerabilities among older Korean Americans.^29–31

Despite solid evidence of the health risks posed by social isolation^3–7 and linguistic isolation,^21,22 these two sources of social disconnectedness have rarely been considered together. Here we use a four-cell typology (no isolation, social isolation only, linguistic isolation only, and dual isolation) to evaluate the structural pattern of social disconnectedness and quantify health risks linked to the combination of social and linguistic isolation. With respect to the double jeopardy hypothesis²⁶ that posits a dual health disadvantage accruing from being both older and a member of racial/ethnic minority groups, the intersection of two sources of vulnerability (social isolation and linguistic isolation) faced by older immigrants might intensify these health disadvantages. Given that the construct of health encompasses multiple dimensions² and that studies often focus on either physical or mental health,^6,11 it is imperative to consider a wide range of health domains.

In the present study, we examine how our typology of social and linguistic isolation in older Korean Americans may be associated with physical, mental, and cognitive health, as indicated by self-rated health, mental distress, and global cognitive performance. Building on the independent lines of literature on social isolation and language as barriers,^3–7,21,22 we hypothesize that older immigrants who are isolated both socially and linguistically are at heightened risks for physical, mental, and cognitive health. We conduct this investigation by considering sociodemographic characteristics (age, gender, marital status, and education), health indicators (chronic conditions and functional disability), immigration-related characteristics (length of stay in the U.S.), and region (areas according to Korean population density).

METHODS

Participants

Data for the present investigation are from the Study of Older Korean Americans (SOKA), a multi-state survey of Korean immigrants aged 60 and older. The states included in the SOKA are California, New York, Texas, Hawaii, and Florida, which represent a wide range of the proportion of the total Korean population residing in the U.S. (29.3%−2.2%).³² In each state, the survey focused on a primary metropolitan statistical area with a representative proportion of Korean Americans: Los Angeles, New York City, Austin, Honolulu, and Tampa. Combined, these sites represent a continuum of Korean population densities. The inclusion of multiple sites is intended to address geographic variations and increase generalizability.

The SOKA’s community-based samples were recruited by investigators who shared the language and culture of the target population. At each of the five SOKA sites, surveys took place at multiple locations and events (e.g., churches, temples, grocery stores, small group meetings, and cultural events) from April 2017 to February 2018. The SOKA questionnaire was in Korean, developed using back-translation and reconciliation. Major instruments were selected for the questionnaire on the basis of their psychometric qualities in the original and Korean-translated versions. The questionnaire was designed to be self-administered, but trained interviewers were onsite for anyone who needed assistance. Upon completion of the questionnaire, each participant was also assessed for cognitive function using the Mini-Mental State Examination (MMSE).³³ Data collection for the project was approved by the Institutional Review Board at the University of Texas at Austin. A total of 2,176 individuals participated in the survey. After removal of those with more than 10% data missing on study variables or whose cognitive status suggested severe impairment (MMSE score <10), the final sample for the present study consisted of 2,032 participants.

Measures

Social isolation.

The Lubben Social Network Scale−6 (LSNS−6)³⁴ was used to indicate social isolation. The scale includes three items on family and a similar set of three items on friends (“How many relatives/friends do you see or hear from at least once a month?” “How many relatives/friends do you feel at ease with such that you can talk with them about private matters?” “How many relatives/friends do you feel close to such that you could call on them for help?”). Respondents answer each question on a 6-point scale (0 = none to 5 = nine or more), with total scores ranging from 0 to 30. The LSNS−6 has been translated into Korean, and its psychometric properties and cut-off scores have been validated.^35,36 Internal consistency of the scale in the present sample was high (Cronbach’s α = .88). Using the suggested cut-off score,³⁴ those who scored lower than 12 were identified as being socially isolated.

Linguistic isolation.

English proficiency was assessed with a question on how well participants spoke English, answered on a 4-point scale ranging from 1 (not at all) to 4 (very well). Responses were dichotomized as no linguistic isolation (0 = well/very well) or linguistic isolation (1 = not at all/a little).

Physical health.

Self-rated health was assessed with a single question: “How would you rate your overall health?” The original response coded on a 5-point scale was dichotomized as either positive (0 = excellent/very good/good) or negative (1 = fair/poor).

Mental health.

As an indicator of mental health, the level of mental distress was measured by the Kessler Psychological Distress Scale 6 (K6).^37,38 Participants are asked to report how often, over the past 30 days, they have experienced such symptoms as “so depressed that nothing could cheer you up,” “hopeless,” and “everything was an effort.” Each item is rated on a 5-point scale ranging from 0 (none of the time) to 4 (all of the time), with responses summed for a composite score from 0 to 24. A score of 6 or greater is indicative of mental distress.^37,38 The K6 has been translated into Korean, and its psychometric properties have been validated in samples of Koreans and Korean Americans.³⁹ Cronbach’s alpha for the present sample was .91. In the present analysis, a dichotomized score was used: 0 = no mental distress (K6 <6) or 1 = mental distress (K6 ≥6).

Cognitive health.

The MMSE³³ was used as an index of global cognitive function. The MMSE includes items on orientation to time and place, word registration and recall, attention and calculation, language, and visual construction. Responses for each item are scored as 0 (incorrect) or 1 (correct); total scores can range from 0 to 30. A score of 24 or below indicates cognitive impairment. The psychometric properties of the Korean version of the MMSE and its cult-off scores have been validated.⁴⁰ Internal consistency of the scale was satisfactory (Cronbach’s α = .73). In the present analysis, a dichotomized score was used: 0 = normal cognition (MMSE score >24) or 1 = cognitive impairment (MMSE score ≤24).

Covariates.

Sociodemographic variables included age (in years), gender (0 = male, 1 = female), marital status (0 = married, 1 = not married), and education (0 = >high school graduation, 1 = ≤high school graduation). Chronic conditions were assessed with a checklist of 10 diseases and conditions common in older populations (hypertension, heart disease, stroke, diabetes, cancer, arthritis, hepatitis, kidney problem, asthma, chronic obstructive pulmonary disease); a total count was used in the analysis. Functional disability was assessed with a composite measure,⁴¹ including activities of daily living (ADL) and instrumental activities of daily living (IADL). The scale included 16 activities (e.g., walking, bathing, dressing, managing medication), and participants were asked to indicate how well they could perform each activity. Responses were coded as 0 (without help), 1 (with some help), or 2 (unable to do). Total scores could thus range from 0 (no functional disability) to 32 (severe functional disability). Internal consistency of the scale in the present sample was high (Cronbach’s α = .89). Length of stay in the U.S. (in years) and region were also included as covariates. Region was coded into three Korean density areas: high (California and New York), medium (Texas and Hawaii), and low (Florida).

Analytical Strategy

After review of the descriptive characteristics, the overall sample was divided into four isolation types: no isolation, social isolation only, linguistic isolation only, and dual isolation. Group comparisons were conducted using F or χ² tests. Separate logistic regression models were examined for the health measures: fair/poor ratings of health, mental distress, and cognitive impairment. For each measure, the odds of reporting health risks associated with each isolation type were examined with adjustment for covariates (age, gender, marital status, education, chronic conditions, functional disability, length of stay in the U.S., and region). The group with no isolation, hypothesized to be least vulnerable, was used as a reference group. All analyses were conducted using IBM SPSS Statistics 27 (IBM Corp., Armonk, NY).

RESULTS

Descriptive Characteristics of the Sample

Characteristics of the overall sample are summarized in Table 1. The mean age of the sample was 73.2 years (SD = 7.96; range, 60 to 100). Over two thirds were women, about 40% were unmarried, and 60% had received an education of high school or less. The scores for chronic conditions and functional disability averaged 1.57 (SD = 1.40) and 1.67 (SD = 3.43), respectively. All participants were foreign-born, and the length of stay in the U.S. ranged from 0.17 to 80 years, with an average of 31.4 (SD = 12.1). More than half of the sample (55%) came from California and New York (high Korean density areas), 30% from Texas and Hawaii (medium Korean density areas), and 15% from Florida (low Korean density areas). Regarding health risks, 36% rated their health fair or poor, about 30% had mental distress, and 18% fell into the category of cognitive impairment.

Table 1.

Descriptive Characteristics of the Sample

	Total Sample (N = 2,032)	No Isolation (n = 419)	Social Isolation Only (n = 102)	Linguistic Isolation Only (n = 1,119)	Dual Isolation (n = 392)	F (χ²)
Social isolation, %	24.3	──	──	──	──
Linguistic isolation, %	74.4	──	──	──	──
Isolation type
No isolation, %	20.6	──	──	──	──
Social isolation only, %	5.0	──	──	──	──
Linguistic isolation only, %	55.1	──	──	──	──
Dual isolation, %	19.3	──	──	──	──
Health risks
Fair/poor rating, %	36.3	13.9	20.6	40.0	53.7	(159.6***)
Kessler 6 ≥6, %	29.6	19.7	32.7	28.2	43.9	(59.0***)
MMSE ≤24, %	18.4	5.7	14.7	18.7	31.9	(93.5***)
Covariates
Age, M±SD	73.2±7.96	71.5±7.45	73.2±7.29	73.3±7.91	74.8±8.43	11.8***
Female, %	66.6	60.6	65.7	69.7	64.5	(12.4**)
Not married, %	39.2	26.7	40.6	38.2	55.0	(68.7***)
≤high school graduation, %	60.1	28.9	38.6	68.7	74.8	(258.1***)
Chronic conditions, M±SD	1.57±1.40	1.12±1.13	1.38±1.18	1.68±1.44	1.76±1.48	20.3***
Functional disability, M±SD	1.67±3.43	0.73±1.40	0.79±2.17	1.69±3.33	2.88±4.57	28.9***
Years in the U.S., M±SD	31.4±12.1	39.3±10.5	39.3±10.0	28.7±11.5	28.6±11.2	117.3***
Region (area Korean density)
High (CA and NY), %	55.1	40.3	41.2	60.6	59.2	(81.9***)
Medium (TX and HI), %	30.0	33.9	37.3	28.0	29.6
Low (FL), %	14.9	25.8	21.6	11.4	11.2

^*p < .05.

^**p < .01.

^***p < .001.

About a quarter of the sample were socially isolated, and three quarters were linguistically isolated. Percentages for the four isolation groups were as follows: no isolation, 20.6%; social isolation only, 5.0%; linguistic isolation only, 55.1%; and dual isolation, 19.3%. Table 1 also shows comparisons across the four isolation groups, and statistically significant group differences were found for all variables. The dual-isolation group demonstrated notably high risks in all health domains: about 54% for fair/poor ratings of health, 44% for mental distress, and 32% for cognitive impairment. The dual-isolation group was also more likely to be older, unmarried, and less educated, with more chronic conditions and greater functional disability than in the other groups. Individuals living in high Korean density areas were more likely to be represented in the groups with linguistic isolation only (60.6%) or dual isolation (59.2%).

Logistic Regression Models of Health Risks

Table 2 summarizes logistic regression models for the three health measures. In the model for physical health, the odds of reporting fair or poor health were 1.99 times higher in the linguistic-isolation-only group and 3.17 times higher in the dual-isolation group than in the no-isolation group. Higher odds were also associated with female gender, unmarried status, lower education, more chronic conditions, and greater functional disability. The odds for having mental distress were 1.89 times higher in the social-isolation-only group, 1.43 times higher in the linguistic-isolation-only group, and 2.69 times higher in the dual-isolation group than those in the reference group. Higher odds of mental distress were also associated with younger age, unmarried status, more chronic conditions, and greater functional disability. In the model for cognitive health, the odds of having cognitive impairment were 2.12 times higher in the social-isolation-only group and 2.21 times higher in the dual-isolation group than those in the reference group. Higher odds of cognitive impairment were also associated with advanced age, female gender, unmarried status, lower education, and greater functional disability.

Table 2.

Logistic Regression Models for Health Risks

	Odds Ratio (95% Confidence Interval)
	Self-rated Health (fair/poor)	Mental Distress	Cognitive Impairment
Isolation type
No isolation	[reference]	[reference]	[reference]
Social isolation only	1.33 (0.71, 2.47)	1.89* (1.14, 3.13)	2.12* (1.01, 4.44)
Linguistic isolation only	1.99*** (1.37, 2.87)	1.43* (1.04, 1.96)	1.38 (0.83, 2.28)
Dual isolation	3.17*** (2.08, 4.83)	2.69*** (1.87, 3.89)	2.21** (1.28, 3.79)
Covariate
Age	1.01 (0.99, 1.02)	0.98** (0.96, 0.99)	1.10*** (1.08, 1.12)
Female	1.39* (1.07, 1.81)	1.06 (0.83, 1.34)	1.65** (1.18, 2.30)
Not married	1.32* (1.04, 1.69)	1.30* (1.04, 1.63)	1.35* (1.01, 1.79)
≤high school	2.24*** (1.72, 2.89)	0.82 (0.64, 1.04)	3.46*** (2.41, 4.96)
Chronic conditions	1.94*** (1.75, 2.14)	1.21*** (1.12, 1.31)	1.05 (0.95, 1.16)
Functional disability	1.12*** (1.08, 1.18)	1.08*** (1.04, 1.11)	1.06** (1.02, 1.10)
Years in the U.S.	.99 (0.98, 1.00)	1.00 (0.99, 1.01)	.99 (0.97, 1.00)
Region (area Korean density)
High	[reference]	[reference]	[reference]
Medium	0.89 (0.69, 1.17)	0.84 (0.66, 1.07)	1.07 (0.79, 1.45)
Low	0.83 (0.59, 1.18)	1.00 (0.74, 1.37)	0.63 (0.39, 1.03)

^*p < .05.

^**p < .01.

^***p < .001.

DISCUSSION

Responding to the growth of the older Asian immigrant population in the U.S.^9,10 and their susceptibility to social disconnectedness,^7,12–15 we have examined the health risks posed by social and linguistic isolation in older Korean Americans. Of particular interest was to identify the pattern of social disconnectedness and quantify its associated risks for physical, mental, and cognitive health. Inspired by the notion of the double jeopardy hypothesis,³¹ we hypothesized that older immigrants who were isolated both socially and linguistically would have a heightened health risk, and our findings provide supportive evidence.

Almost a quarter of the sample fell into the category of social isolation based on the LSNS−6 scale, which is higher than the 11−22% reported in studies using the same scale with European and North American older adult samples.^34,42 Reflecting the high level of limited English proficiency in the Korean population in the U.S.^16,17 and the particular linguistic challenge to its older members who are predominantly foreign-born immigrants,^28,29 almost three quarters of the present sample reported that they spoke English less than well. Using a four-cell typology, we divided the sample into groups with no isolation (20.6%), social isolation only (5.0%), linguistic isolation only (55.1%), and dual isolation (19.3%). The social-isolation-only group was small, in part because many of those who were socially isolated were linguistically isolated as well. The fact that more than half of the sample was in the linguistic-isolation-only group underscores the pervasiveness of language barriers in older Korean Americans. It is also noteworthy that the rate of limited English proficiency or linguistic isolation was higher in the geographic areas with higher Korean densities. This is in line with previous findings that older immigrants living in ethnic enclaves with abundant resources and services tend to have lowered needs and desires for learning English.²³

About 20% of the participants were both socially and linguistically isolated, and they demonstrated notable social disadvantages and lack of resources (e.g., advanced age, unmarried status, low education, comorbidity, and functional disability). They also presented high rates of health risks in all three domains, with about 54% reporting fair or poor health, 44% having mental distress, and 32% having cognitive impairment.

Logistic regression models showed increased health adversities associated with either social or linguistic isolation or with both. Compared with the reference group without any isolation, those who were both socially and linguistically isolated demonstrated 2.21−3.17 times higher odds of having fair/poor ratings of health, mental distress, and cognitive impairment. Having social-isolation-only or linguistic-isolation-only was associated with 1.43−1.89 times higher odds of having mental distress. Linguistic-isolation-only was a significant predictor only in the model for self-rated health (OR = 1.99, 95% CI = 1.37, 2.87), and social-isolation-only was significant only in the model for cognitive impairment (OR = 2.12, 95% CI = 1.01, 4.44). Although the covariates were differently associated with the health measures, unmarried status and functional disability were common predictors in all models. It is noteworthy that the length of stay in the U.S. and the regional variations in Korean population densities were not significantly related to health risks. This may suggest that the effects of social or linguistic isolation cancel out the influences of immigration- and geography-related characteristics.

To reiterate, our goal was to move beyond the independent lines of literature on social isolation or language as barriers^3–7,21,22 and to consider both interpersonal relations and language proficiency as major sources of social integration in older immigrants. We wished to quantify the health risks associated with isolation from both sources. Our finding that being both socially and linguistically isolated exerts heightened risks in all health domains accords with the notion of the double jeopardy.³¹ It should also be noted that the health risks are not limited to physical or mental health but also include cognitive health. Interpersonal resources and English skills not only keep older immigrants physically and emotionally healthy but also serve as cognitive reserves that stimulate cognitive health and well-being. Given that social network and language proficiency are also critical enablers for health management and service utilization,^12,22,30 health care providers should be alert to older immigrants who are both socially and linguistically disconnected.

Some limitations of our investigation should be noted. First, the cross-sectional nature of the investigation limits our ability to draw causal inferences. The possibility that poor physical, mental, and cognitive health might interfere with older immigrants’ social and linguistic integration should not be precluded. Although the research team made efforts to recruit participants in various settings and geographic areas, the study may have excluded hard-to-reach older immigrants such as those who are homebound and disconnected from the community. In addition, our assessment of physical, mental, and cognitive health was based on self-reports. Future studies might use performance-based measures and diagnostic tools for objective health assessment and include psychosocial and cultural variables to contextualize the link between social connectedness and health. Future studies should also consider community-level social capital, such as the availability and accessibility of social resources in the ethnic community.

Despite these limitations, however, the present study contributes to our understanding of the roles of social and linguistic isolation in multiple dimensions of health among older Korean immigrants. The four-cell typology of social and linguistic isolation can be used to identify vulnerable subgroups within any immigrant population and to develop targeted interventions to meet their needs.

Key Points:

• Both social relationships and language proficiency are key elements for older immigrants’ social connectedness and integration.

• The coexistence of social and linguistic isolation puts older immigrants at heightened risks in multiple dimensions of health.

Why Does This Paper Matter?

By considering the coexistence of social and linguistic isolation, the paper identified the pattern of social disconnectedness and quantified its associated risks for physical, mental, and cognitive health in older immigrants.