Health in the Neighborhood and Household Contexts Among Older Chinese Americans

Abstract

Objective:

This study examined the associations of neighborhood characteristics and living arrangements with physical and mental health among older Chinese Americans.

Method:

A sample of 3,159 community-dwelling Chinese older adults in the Greater Chicago area provided reports of health, socio-demographic characteristics, living arrangements, social cohesion, and neighborhood disorder. We used multinomial logistic, Poisson, and negative binominal regression analyses.

Results:

Neighborhood disorder was consistently associated with negative health indicators, including poor self-reported health, more chronic conditions, depressive symptoms, and anxiety symptoms. Findings about the relationships between social cohesion and health indicators were mixed. Social cohesion was more salient to mental health for those living with spouse, children, and/or grandchildren relative to those living with spouse only.

Discussion:

Policies and interventions are needed to improve the physical and social environments of neighborhoods and to promote healthy aging among Chinese older adults and in the general population as well.

Introduction

Chinese Americans are a large and rapidly increasing segment in the United States, and understanding of their health is an important population health objective. Older Chinese Americans have experienced a disproportionate burden of preventable disease, disability, and psychosocial distress as compared with non-Hispanic Whites (Dong, 2014). Health disparities may be related to the variations in neighborhoods and households. Living in some arrangements and residing in neighborhoods with threatening conditions are associated with higher levels of disability, morbidity, and mortality (e.g., Balfour & Kaplan, 2002; Beard et al., 2009; Chaix et al., 2006; Echeverria, Diez-Rouz, Shea, Borrell, & Jackson, 2008; Freedman, Grafova, Schoeni, & Rogowski, 2008; Pruchno, Wilson-Genderson, & Cartwright, 2012; White et al., 2010). By contrast, neighborhood belonging and social connectedness may contribute to positive health outcomes (e.g., Elliott, Gale, Parsons, Kuh, & The HALCyon Study Team, 2014; Freedman et al., 2008; Kawachi & Berkman, 2000; Osypuk, Diez Roux, Hadley, & Kandula, 2009). Embedded in the neighborhood context, living arrangements are salient to health in old age (e.g., Chen & Short, 2008; Chou, Ho, & Chi, 2006; Li, Zhang, & Liang, 2009), as older adults may maintain social relations and receive social support from coresident household members. However, little relevant research has focused on older Chinese Americans and health advantages and disadvantages of different living arrangements and neighborhoods. Significantly less is known about the interaction effect between neighborhood and household on the health of older adults. The relationships between health and living arrangements may vary by neighborhoods; for instance, older adults who live alone may maintain overall functioning through access to health resources and social connectedness in a cohesive neighborhood; whereas, coresidence may not have health protective effects if a neighborhood with threatening conditions truly affects individual and family life.

As an old saying goes, “A neighbor living nearby is better than a relative far away.” This proverb expresses the expectations for mutual aid and emergency assistance among neighbors, which may be especially important for older Chinese Americans in need of social support and immediate help from neighbors. However, previous research failed to simultaneously examine the relationships of health with neighborhood and household characteristics.

Drawing on the data from the largest community-based study of older Chinese Americans, this study investigates the associations of physical and mental health with living arrangements and neighborhood characteristics based on self-assessment reports.

Neighborhood and Health

A growing body of research in epidemiology and social gerontology has documented the relationship between neighborhood characteristics and the health of older adults (see Elliott et al., 2014). As reviewed by Yen, Yvonne, and Perdue (2009), previous research has applied the measures of both objective and subjective aspects of neighborhood, including socioeconomic composition, racial composition, neighborhood demographics, perceived resources and problems, physical and social environment. These aspects are linked to health outcomes in four categories: overall mortality, chronic condition mortality or disease prevalence, mental health, and health behaviors (Yen et al., 2009). In general, residing in socioeconomically and physically disadvantaged neighborhoods is associated with higher levels of disability, mental illness morbidity, and mortality (e.g., Balfour & Kaplan, 2002; Freedman et al., 2008; Pruchno et al., 2012; White et al., 2010). Both physical and social environments of neighborhoods are related to individual well-being. Physical environments such as green spaces have positive relationships with self-perceived health, longevity, chronic conditions, and psychiatric morbidity (Mass et al., 2009). Social environments, such as socioeconomic status (SES) and social capital, are well-known health indicators (e.g., Beard et al., 2009; Lochner, Kawachi, Brennan, & Buka, 2003). Altogether, low neighborhood SES, residential instability, racial/ethnic composition, and negative street characteristics are associated with higher prevalence of physical disability (Beard et al., 2009). As demonstrated in a large-scale randomized social experiment designed study, moving from a high-poverty to a lower poverty neighborhood is associated with long-term improvements in health and subjective well-being (Lugwig et al., 2012). Table 1 summarizes research findings about the relationships between neighborhood characteristics and various health indicators.

Table 1.

Literature Review of Relationships Between Living Environments and Health Among Older Adults.

Study	Neighborhood characteristics	Health outcomes
Balfour and Kaplan (2002)	Perceived problems: Crime, lighting, traffic, noise, trash and litter, public transportation	Multiple neighborhood problems were related to overall and lower extremity functional loss.
Beard et al. (2009)	SES, residential stability, racial/ethnic composition	Low neighborhood SES, residential instability, low proportions of foreign-born and high proportions of Black residents, and negative street characteristics were associated with higher prevalence of physical and functional disability.
Chaix et al. (2006)	Contextual deprivation: Mean income in quartiles, social disorganization	The risk of substance-related disorders increased with contextual deprivation and social disorganization. The risk of neurotic disorders only increased with contextual deprivation.
Echeverria, Diez-Rouz, Shea, Borrell, and Jackson (2008)	Perceived problems and cohesion	Less socially cohesive neighborhoods were associated with increased depression, smoking, and not walking for exercise. Results for neighborhood problems were robust but not for social cohesion.
Elliott, Gale, Parsons, Kuh, and The HALCyon Study Team (2014)	Perceived cohesion	Neighborhood cohesion was positively related to mental well-being.
Freedman, Grafova, Schoeni, and Rogowski (2008)	Connectivity, density, air pollution, immigration, residential stability, crime/Black segregation, economic advantage/disadvantage	Economic advantage was associated with a reduced risk of lower body limitations for both men and women. For men, living in a more connected area was associated with a lower risk of lADLs limitations, and living in an economically disadvantaged area was associated with an increased risk of ADLs limitations.
Lochner, Kawachi, Brennan, and Buka (2003)	Social capital: Reciprocity, trust, civic participation	Social capital was associated with lower mortality rates from all causes.
Lugwig et al. (2012)	Moving to opportunity, poverty rate, racial composition, safety, social processes	Moving from a high-poverty to lower poverty neighborhood was related to long-term improvement in physical, mental, and SWB.
Mass et al. (2009)	Green space	The annual prevalence rate of 15 of the 24 disease clusters was lower in living environments with more green spaces. The relation was strongest for anxiety disorder and depression.
Pruchno, Wilson- Genderson, and Cartwright (2012)	Availability of physicians, residential stability, violence, social vulnerability, wealth, storefronts	Supermarkets, availability of physicians, wealth and residential stability were associated with lower levels of disability; whereas, storefronts, violence, and social vulnerability were associated with higher levels of disability.
White et al. (2010)	parks, walking areas, handicap parking, public transportation	No parks or walking areas was associated with the reduced odds of engaging in fitness programs and social activities. Adequate handicap parking and presence of public transportation increased the odds of engagement in social, work, or leisure activities.
	Living arrangement
Chen and Short (2008)	Spouse, children, spouse and children, alone, nursing home, others	Living alone was associated with lower SWB. Coresidence with spouse and/or children was associated with positive SWB.
Chou, Ho, and Chi (2006); Mui (1999)	Alone or not	Living alone was related to more depressive symptoms, especially for older women and immigrants.
Hughes and Waite (2002)	Married (alone, children, or others), single (alone, with children, or others)	Married couples living alone or with children were the most advantaged; single women living with children were most disadvantaged on physical and mental health.
Kharicha et al. (2007)	Alone, others	Living alone was associated with increased risks for functional impairment, arthritis and/or rheumatism, glaucoma, and cataracts.
Li, Zhang, and Liang (2009)	Alone, spouse, children, spouse and children, others, institution	Living with a spouse lowered mortality risks. For men living in institutions was associated with lower mortality risks. Those living alone had fewer ADLs disabilities. Those living with children had better self-rated health.
Liu and Zhang (2004)	Alone, others, institution	The oldest-old Chinese living with family members had better self-rated health than those living in institutions.
Lund et al. (2002)	Alone or not	Living alone was associated with increased mortality.  Cohabitation was a stronger predictor of mortality than married status.
Russell (2009)	Spouse, alone, others	Those living alone or with others reported greater loneliness than those living with a spouse, especially among those with physical disabilities.
Samanta, Chen, and Vanneman (2015)	Alone, spouse, spouse and children, multi-generational, others	Older Indians living in multi-generational households had the lowest levels of short-term illness. Living alone was associated with the highest likelihood of short-term morbidity.
Sereny (2011)	Living arrangement preference, concordance (a match between actual and preferred living arrangement)	Independent living discordance was related to poor self-rated health than those with independent living concordance. Those who coresided were more likely to be disabled than those with independent living concordance.
Van Gelder et al. (2006)	Change in living situations, alone or others	Men losing a partner, unmarried, starting to live alone, or living alone during the 5-year period had stronger cognitive decline compared with men who were married or who lived with someone.
Wilmoth and Chen (2003)	Spouse (only or with family or others), alone, family or others only	Living alone and living with family/others were positively related to depressive symptoms compared with living with spouse only, especially among immigrants.
Ye and Chen (2014)	Spouse only, children or grandchildren, spouse and children, others, alone	Living with children was positively associated with mental health among urban Chinese elders.

Note. SES = socioeconomic status; IADLs = instrumental activities of daily life; ADLs = activities of daily life; SWB = subjective well-being.

Despite growing empirical evidence of the neighborhood–health associations, the pathways to individual health are not completely clear. Due to the aging-related decline in physical and cognitive functioning and reduction in social networks and social capital, older adults increasingly rely on the immediate residential neighborhood for services and amenities (Yen et al., 2009). Neighborhood disorder may affect residents’ daily lives and health through a number of pathways, for example, reduced levels of physical activity and social engagement, and increased distrust and social isolation (Balfour & Kaplan, 2002; Cagney et al., 2009; Elliott et al., 2014). Research also indicates that the relationship between neighborhood cohesion and mental health is stronger for adults above the age of 65, probably because a sense of neighborhood belonging or perceived cohesion provides a strong, positive identity for older adults (Elliott et al., 2014).

Subjective evaluations of neighborhood characteristics reflect social processes and dynamics that influence individual health (Cagney et al., 2009). Building on the collective efficacy and social disorganization theories, Cagney et al. (2009) created and confirmed two domains in the measures that capture key aspects of neighborhood characteristics: social cohesion and neighborhood disorder. In reliance on self-reported assessment of neighborhood quality, interactions, and attitudes that characterize neighborhood social processes, social cohesion indicates “neighborhood social resources in the form of mutual trust and solidarity and expectations for action” (p. 416); neighborhood disorder indicates that “visible signs of community decay and social declines contribute to fear of victimization and social withdrawal” (Cagney et al., 2009, p. 417). Both measures have relevance to the health of older adults. Social cohesion may affect health by means of improving health-related behaviors, increasing access to services and amenities, and providing affective support of self-esteem and mutual respect (Kawachi & Berkman, 2000). Neighborhood disorder may be a source of chronic stress that contributes to unhealthy coping behaviors, poor mental health outcomes, and decreased physical activity (Echeverria et al., 2008). Social cohesion and neighborhood disorder that reflect neighborhood conditions influence the lives and health of individual residents and their families.

Health in the Household Contexts

Intact family structure is one of important resources for promoting social cohesion and support among minority neighborhoods (Aranda, Ray, Snih, Ottenbacher, & Markides, 2011). In the Chinese culture of collectivism, spouse, kin, neighbors, friends, and coworkers are discrete entities but have experiences as intact social groups characterized by cohesiveness and interdependence (Brewer & Chen, 2007). As the primary units in everyday life, households and neighborhoods are positioned to provide opportunities for social integration and social support, thus promoting good health outcomes (Aranda et al., 2011; Chen & Short, 2008). Living arrangements are associated with older adults’ health, including overall mortality, chronic condition morbidity, mental health outcomes, and physical and cognitive function (see Table 1). Married couples are most advantaged, with the highest levels of physical, mental, and cognitive functioning; by contrast, those living alone are at greater risk of poor physical and emotional health, cognitive decline, and mortality (e.g., Chen & Short, 2008; Chou et al., 2006; Hughes & Waite, 2002; Kharicha et al., 2007; Li et al., 2009; Lund et al., 2002; Mui, 1999; Van Gelder et al., 2006). However, findings are not consistent, with some studies documenting that older adults living alone have better functional status relative to those living with others in China (Li et al., 2009; Liu & Zhang, 2004). Living alone and living with children have both health advantages and disadvantages in Chinese older adults, as shown in the studies conducted in China (Li et al., 2009; Ye & Chen, 2014).

The relationship between living arrangements and health may be mediated through social integration and support from household members (Hughes & Waite, 2002; Samanta et al., 2015). Living with a spouse encourages high levels of participation in social and recreational activities among older adults (Russell, 2009). Living with children or others provides the widowed emotional support, such as the feelings of being cared about, loved, and security (Ha, 2008). In addition, immigration experience may increase the need for coresidence, as a breakdown in traditional values and family support system exacerbates the aging and immigration stress process and leads to increasing reliance on family members to deal with various stresses (Wilmoth & Chen, 2003). Therefore, supportive family relationships and strong community ties are important to the health of older immigrants. Living in a cohesive neighborhood, coresidence may provide more opportunities for social integration and support that will further contribute to health advantages.

The Present Study

Based on the literature of health in the contexts of neighborhood and household, the current study aims to examine the relationships of physical and mental health with neighborhood characteristics and living arrangements. Although extensive research has shown that older adults’ physical and mental health vary across neighborhoods and households, no study has evaluated whether neighborhood characteristics moderate the relationship between living arrangements and health. Individual perceptions about social cohesion and neighborhood disorder play a role in health evaluations (Cagney et al., 2009). Living in a neighborhood with high levels of social cohesion may buffer the negative health effects of living alone, whereas living in disordered neighborhoods may limit the opportunities for social integration and support among household members. In addition, little is studied in older Chinese Americans as a group of racial minority and immigrants. A few studies with focus on mental health among older Chinese immigrants are limited in operationalizing living arrangement, which was measured as living alone versus living with others, while ignoring multiple arrangements (e.g., Mui, 1999).

Using a population-based, representative sample of Chinese older adults, we seek to address these gaps and ask the following questions:

Research Question 1:

What are the relationships between neighborhood characteristics, that is, social cohesion and neighborhood disorder, and overall health measured by self-rated health, chronic conditions, depressive symptoms, and anxiety symptoms?

Research Question 2:

What are the relationships between living arrangements and overall health?

Research Question 3:

Is there any interaction effect between neighborhood characteristics and living arrangements on individual health?

Method

Population and Settings

We used the data from the Population Study of Chinese Elderly in Chicago (PINE), a population-based, epidemiological study of Chinese older adults aged 60 and above. The PINE was conducted between 2011 and 2013 in the Greater Chicago area, with the main purpose of examining the key cultural determinants of health and well-being. The project was initiated by a community–academic collaborative team among the Rush Institute for Healthy Aging, Northwestern University Medical Center, and many community-based social services agencies and organizations (Dong, Wong, & Simon, 2014). A community-based participatory research approach was applied in the PINE study to carry out culturally and linguistically appropriate community recruitment strategies (Dong, Chang, Simon, & Wong, 2011). Neighborhoods were selected based on postal zip codes in large geographic areas (including north side, south west, west, south, suburbs, and others) and Chinatown.

Of 3,542 eligible older adults who were approached, 3,159 agreed to participate in the study, yielding a high response rate of 91.9%. Based on the available data drawn from the U.S. Census 2010 and a random block census project, the PINE study was representative of the Chinese aging population in the Greater Chicago area (Simon, Chang, Rajan, Welch, & Dong, 2014). The study was approved by the Institutional Review Board of the Rush University Medical Center.

Measurement

Social cohesion.

The index was composed of six items that were extracted from a series of questions used in the Chicago Neighborhood and Disability Study (CNSD; Cagney et al., 2009). Some items were designed to measure individual level of integration in the neighborhood; for example, respondents were asked “how often in your neighborhood do you see neighbors and friends talking outside in the yard or in the street?” “Do you see neighbors watching out for each other such as calling if they see a problem?” Other items were designed to evaluate the overall level of social cohesiveness that individuals perceived to exist in the neighborhood (Cagney et al., 2009), including the number of neighbors who the respondent knew by name, had a friendly talk at least once a week, or who could be called on for assistance (Cagney et al., 2009). The index had high internal consistency (Cronbach’s α = .86) in this study. Due to the different response scales used in six items, standardized scores were calculated and higher scores indicated more cohesion perceived in the neighborhood.

Neighborhood disorder.

The disorder index contained eight items that were used in the CNSD in evaluating the neighborhood’s physical and social disorders (Cagney et al., 2009). Respondents were asked whether they had observed the presence of potentially threatening or intimidating conditions, including strangers, speeding cars, vandalism, and unsafety of walking around the neighborhood. They were also asked about the state of disrepair or neglect of the built environment; for example, “How often in your neighborhood do you see poorly maintained sidewalks or broken curbs?” Responses were scaled from 0 (never) to 3 (often). The index had a high level of internal consistency (Cronbach’s α = .81) in this study. Standardized scores were used and higher scores indicated more disorders perceived.

Living arrangement.

Living arrangement was categorized into (a) living alone; (b) living with spouse only, (c) living with spouse and children; (d) living with spouse, adult children, and grandchildren, indicating residence with multi-generations; and (e) living with others (e.g., relative, friend, or roommate).

Self-rated health.

Respondents were asked to rate their health in general and the responses were given on a 4-point scale (1 = very good to 4 = poor). Due to a small number of respondents reporting very good health, they were combined with those self-rating good health (n = 1,215). In multinomial logistic regression analysis, those reporting fair health (n = 1,307) and those reporting poor health (n = 583) were compared with the good/very good health group, respectively.

Chronic conditions.

We used the summary score of medical conditions that had been diagnosed by health care providers to indicate morbidity. These conditions included heart disease, stroke or brain hemorrhage, cancer, high cholesterol, diabetes, high blood pressure, a broken or fractured hip, thyroid disease, osteoarthritis, inflammation, or problems with joints. It was used as a count variable (range = 0–5).

Depressive symptoms.

The Patient Health Questionnaire–9 (PHQ-9) scale instrument was used to measure depressive symptoms. The PHQ-9 scale is brief and validated with nine questions. It is appropriate for screening late-life depression in the general population (Arean & Ayalon, 2006; Nease & Malouin, 2003). In addition, the PHQ-9 scale contains the somatic domains that are common in depressed Asian older adults (Donnelly & Kim, 2008). Respondents were asked about the frequency of being bothered by such feelings as little interest in doing things, feeling down, trouble in sleeping, feeling tired, poor appetite, feeling bad about oneself, and trouble in concentrating on things during the past 2 weeks. Responses were scaled from 0 (not at all) to 3 (nearly every day). A summary score of nine items was used with higher scores indicating more symptoms (Cronbach’s α = .82).

Anxiety symptoms.

The 7-item Hospital Anxiety and Depression Scale–Anxiety (HADS-A) was used to measure anxiety symptoms. The HADS-A has been tested in Chinese populations and shown good inter-rater reliability (Lam, Pan, Chan, Chan, & Munro, 1995). Respondents were asked whether they had experienced the symptoms such as feeling tense or wound up. Responses were scaled from 0 (not at all) to 3 (most of the time). A summary score was used, and higher score indicated more symptoms (Cronbach’s α = .80).

Socio-demographic variables included age in years (range = 59–105), gender (male or female), education (measured by years in school, range = 0–26), personal income (range 1 = less than US$5,000 to 10 = US$45,000 or more), number of household members, and years living in the neighborhood. Less than 4% of the observations had missing values, including education (n = 19), income (n = 36), number of household members (n = 1), years living in the neighborhood (n = 10), social cohesion (n = 106), neighborhood disorder (n = 39), depressive symptoms (n = 18), and anxiety symptoms (n = 28). The small number of missing data does not raise concerns about biased results; thus, we did not impute the missing data. Table 2 presents the descriptive information of all variables under study.

Table 2.

Descriptive of the PINE Participants (N = 3,157).

Variable	M (SD)/%
Age (range = 59–105)	72.81 (8.30)
Female (%)	57.97
Education (range = 0–26)	8.72 (5.05)
Income (range = 1–10)	1.95 (1.14)
Years in the neighborhood (range = 0.1–80)	12.14 (11.04)
Number of household members (range = 0–10)	1.87 (1.89)
Living arrangement (%)
Living alone	21.48
Living with spouse only	37.88
Living with spouse and children	12.29
Three generations	22.49
Living with others	5.86
Neighborhood cohesion (range = −1.08–2.84)	0.00 (0.77)
Neighborhood disorder (range = −0.99–4.38)	0.00 (1.00)
Self-rated health (%)
Good/very good	39.13
Fair	42.09
Poor	18.78
Chronic conditions (range = 0–5)	1.42 (1.12)
Depressive symptoms (range = 0–27)	2.65 (1.13)
Anxiety (range = 0–21)	2.65 (3.28)

Note. PINE = Population Study of Chinese Elderly in Chicago.

Data Analysis

To examine the relationships of neighborhood characteristics and living arrangements with health indicators, multinomial logistic, Poisson, and negative binominal regression models were estimated. Multinomial logistic regression was applied in examining self-rated health, comparing respondents with poor and fair health against those with good/very good health, respectively. Poisson regression model was estimated to predict the count variable chronic conditions. Negative binominal regression models were estimated to test the relationships with depressive symptoms (skewness = 2.27, kurtosis = 5.95) and anxiety symptoms (skewness = 1.85, kurtosis = 4.10). Both measures were discretely distributed with a large proportion reporting zeros and only a few cases reaching the critical points (e.g., major depression). In this case, a continuous version of negative binomial model is appropriate to improve the model fit to the data and account for over-dispersion (Chandra & Roy, 2012). Negative binomial analysis results indicate the change in the incident rate of the outcome variable per unit change in the independent variable after controlling for covariates. In addition, interaction terms between living arrangements (with those living with spouse only used as the reference group) and neighborhood characteristics (i.e., social cohesion and neighborhood disorder) were created and tested, respectively. Statistical analyses were conducted using SAS version 9.2 (SAS Institute Inc., Cary, NC).

Results

Table 3 presents parameter estimates and standard errors of the multinomial logistic regression analysis on self-rated health and the Poisson regression analysis on chronic conditions. After controlling for socio-demographics, social cohesion was associated with decreased likelihood (B = −.28, SE = .08, p < .001), whereas neighborhood disorder was associated with increased likelihood of self-reporting poor health (B = .23, SE = .05, p < .001). Specifically, one standard deviation (SD) increase in the cohesion index was associated with 24% (e^b = e^−.28 = 0.76) decrease in the likelihood, whereas one SD increase in the disorder index was associated 26% (e^b = e^.23 = 1.26) increase in the likelihood of reporting poor health. Neither social cohesion nor neighborhood disorder was related to the likelihood of self-reporting fair health.

Table 3.

Multinomial Logistic and Poisson Regression Results of Self-Rated Health and Chronic Conditions.

	Poor healtha		Fair healtha		Chronic conditions
	B	SE	B	SE	B	SE
Age	.03***	.01	.02**	.01	.02***	.00
Female	.37**	.08	.16	.09	.14***	.03
Education	−.00	.01	−.00	.01	.01**	.00
Income	−.27***	.06	−.12**	.04	−.01	.02
Years in neighborhood	−.02***	.01	−.01***	.00	−.00	.00
Household members	−.11	.07	.07	.05	−.02	.02
Living arrangement (ref: spouse only)
Alone	−.26	.16	−.03	.12	.02	.05
SC	.10	.19	−.03	.15	−.00	.06
Multi-generations	.30	.27	−.33	.21	.01	.08
Others	.21	.27	−.03	.21	−.08	.08
Social cohesion	−.28***	.08	−.00	.05	.04*	.02
Neighborhood disorder	.23***	.05	.05	.04	.03*	.02
Living Alone × Social Cohesion	−.14	.19	−.24	.14	−.03	.05
SC × Social Cohesion	−.43	.29	−.08	.20	−.06	.08
Multi-Generation × Social Cohesion	−.33	.21	−.07	.15	−.01	.06
Others × Social Cohesion	−.14	.32	−.09	.23	−.10	.09
Living Alone × Neighborhood Disorder	−.19	.15	−.02	.12	.00	.04
SC × Neighborhood Disorder	−.18	.17	.17	.13	.03	.05
Multi-Generation × Neighborhood Disorder	.05	.13	.15	.11	.01	.04
Others × Neighborhood Disorder	−.09	.24	.17	.18	−.01	.07

Note. SC = spouse and children.

^aSelf-rated good/very good health was the reference group.

^*p < .05.

^**p < .01.

^***p < .001.

Social cohesion (B = .04, SE = .02, p < .05) and neighborhood disorder (B = .03, SE = .02, p < .05) were both positively related to chronic conditions. One SD increase in social cohesion was associated with 4% (e^b = e^.04 = 1.04) increase, and one SD increase in neighborhood disorder was related to 3% (e^b = e^.03 = 1.03) increase in the likelihood of having one more medical conditions. Living arrangements and interaction terms were not related to self-rated health or chronic conditions.

Table 4 shows the negative binominal regression results on depressive symptoms and anxiety symptoms. After controlling for socio-demographics, social cohesion was negatively related to the incident rate of depressive symptoms (B = −.23, SE = .04, p < .001) and anxiety symptoms (B = −.10, SE = .03, p < .01). That is, one SD increase in the cohesion index was related to 21% (e^b = e^−.23 = 0.79) less likely to increase depressive symptoms, and 10% (e^b = e^−.10 = 0.90) less likely to increase anxiety symptoms. By contrast, neighborhood disorder was positively related to the incident rate of depressive symptoms (B = .15, SE = .03, p < .001), and anxiety (B = .19, SE = .00, p < .001). Particularly, one SD increase was associated with 16% (e^b = e^.15 = 1.16) more likely to increase depressive symptoms, and 21% (e^b = e^.19 = 1.21) more likely to increase anxiety symptoms, respectively.

Table 4.

Negative Binominal Regression Results of Depressive Symptoms and Anxiety Symptoms.

	Depressive symptoms		Anxiety symptoms
	B	SE	B	SE
Age	.02***	.00	.00	.00
Female	.40***	.06	.45***	.05
Education	.01	.01	.01	.01
Income	−.13***	.03	−.07**	.02
Years in neighborhood	−.01***	.00	−.01**	.00
Household members	−.06	.04	−.04	.03
Living arrangement (ref: spouse only)
Alone	−.03	.09	−.01	.07
SC	−.14	.11	−.07	.09
Multi-generations	.09	.16	−.01	.12
Others	.11	.15	.16	.12
Social cohesion	−.23***	.04	−.10**	.03
Neighborhood disorder	.15***	.03	.19***	.02
Living Alone × Social Cohesion	.02	.10	−.15	.08
SC × Social Cohesion	−.31*	.15	−.30**	.11
Multi-Generation × Social Cohesion	−.21*	.10	−.34***	.09
Others × Social Cohesion	−.08	.17	−.15	.14
Living Alone × Neighborhood Disorder	.07	.09	−.07	.07
SC × Neighborhood Disorder	.10	.09	.01	.08
Multi-Generation × Neighborhood Disorder	.11	.08	.08	.06
Others × Neighborhood Disorder	.08	.13	.01	.10

Note. SC = spouse and children.

^*p < .05.

^**p < .01.

^***p < .001.

Social cohesion moderated the relationships of certain living arrangements with depressive symptoms and anxiety symptoms. Compared with those living with spouse only, respondents living with spouse and children were about 27% (e^b = e^−.31 = 0.73) and 26% (e^b = e^−.30 = 0.74) less likely to increase depressive symptoms and anxiety symptoms, respectively; those living with multiple generations were about 19% (e^b = e^−.21 = 0.81) and 29% (e^b = e^−.34 = 0.71) less likely to increase depressive symptoms and anxiety. The results suggested that social cohesion promoted mental wellness for those living with spouse, children, and/or grandchildren.

Generally, older age, female, and lower income were associated with health disadvantages, whereas more education was associated with increased number of chronic conditions. More years living in the neighborhood was related to decreased likelihood of self-reporting poor and fair health, and decreased incident rates of depressive symptoms and anxiety symptoms, indicating residential tenure contributed to general health status and mental health.

Discussion

The present study showed that neighborhood characteristics were associated with health indicators in a sample of Chinese older adults living in the Greater Chicago area. Consistent with previous research (e.g., Balfour & Kaplan, 2002), we found that the presence of threatening neighborhood conditions, such as unsafe traffic conditions and strangers, and the state of disrepair or neglect of built environment, such as poor lightening and excessive noise, was negatively related to physical and mental health among older Chinese Americans. As shown in previous research, disruptive living environments were associated with older adults’ functional health and overall self-perceived health by interfering with physical activity, community participation, safety, and self-care tasks (e.g., food shopping; Balfour & Kaplan, 2002). In addition, living in a disruptive and threatening neighborhood increased social isolation and loss of important relationships, which further contributed to the feelings of emptiness and depression (Singh & Misra, 2009).

In general, higher levels of social cohesion were related to less likelihood of self-reporting poor health, depressive symptoms, and anxiety symptoms. Due to reduced social roles and declined physical functioning, older adults may have restricted social and service interactions to the immediate environment; however, the close relationships with neighbors may provide opportunities for social integration and social support. Moreover, coresident household members, especially children and grandchildren may help older adults expand the social relationships through their engagement in a variety of activities and social networks. As suggested by the significant interaction effects, social cohesion was especially beneficial to the mental health of older Chinese Americans who lived with their children and/or grandchildren. Meanwhile, the richness of social networks with family and neighbors may promote the perceptions of social cohesion in the neighborhood, which thus related to older adults’ health.

An interesting finding of this study was that social cohesion was positively associated with diagnosed chronic conditions. It might be the case that older adults who had chronic conditions were likely to frequently talk with and socialize their neighbors for the purpose of seeking information, emotional support, or practical help. Also, some chronic conditions such as high blood pressure or high cholesterol do not have a direct link to mobility; older adults with some minor symptoms were still able to move around and participate in physical and social activities. Furthermore, social cohesion was less dependent on neighborhood SES than was disorder, or living in low SES neighborhoods may not prevent the development of social cohesion, exchange, and trust among residents (Cagney et al., 2009). That is, residents in such neighborhoods still have a high level of neighborhood feeling and cohesion despite declined physical functioning. Also, people with chronic conditions may choose to live in the neighborhood where they had access to social support.

Surprisingly, the study did not find sufficient evidence about the relationships between living arrangements and health, implying that the family may play a less significant role in maintaining the health of older adults relative to the neighborhood. In the Chinese culture that emphasizes family obligation and intergenerational support, older parents and grandparents often contribute to the care and well-being of descendant generations (Silverstein & Cong, 2013). Especially among immigrants, Chinese older adults may feel obligated to help their children and grandchildren to cope with various post-immigration challenges (Zhou, 2012). Serving as helpers or caregivers may help older adults to perceive their values and identity in the family life, but stress may increase along with care burden. Thus, residing in a cohesive neighborhood with access to social support and immediate help from neighbors is important to maintain family well-being and individual health as well. In addition, other factors, such as coresidence preference and selection, normative contexts, intergenerational conflicts, and socioeconomic environment may mediate the relationships between living arrangements and health (Samanta et al., 2015; Sereny, 2011).

One limitation of the present study is the lack of randomized social experiment design to ensure the causal relationship between living environment and individual health. Some people had choice over where to live and/or with whom to live in later life; thus, they were able to move in an aging-friendly environment with sufficient support and services to meet the old age needs. Therefore, social selection may exist in the neighborhood–health relationships. In addition, the cross-sectional study design did not allow us to track changes in neighborhood features and living arrangements and how they affected individual health change over time. Although we demonstrated the neighborhood–health relationship, we have not yet addressed the pathways through which neighborhood characteristics affected older adults’ health. A cohesive neighborhood providing social resources in forms of mutual trust, solidarity, and expectations for action may relate to the social control of health-related behaviors and conditions and the positive psychosocial processes that generate protective effects on health (Kawachi & Berkman, 2000). Future studies need to investigate these potential pathways that link neighborhood characteristics to health and examine both subjective and objective evaluations of neighborhoods, relying on individual- and neighborhood-level data. In addition, the lack of detailed measures of morbidity, such as the start of chronic diseases, may prevent accurate estimation of the relationship between living environment and health. It is important for future studies to include measures of chronic conditions mortality, disease prevalence, physical functioning such as activities of daily life (ADLs) and instrumental activities of daily life (IADLs).

Our findings speak to the importance of enhancing social cohesion and addressing neighborhood disorder in promoting older Chinese Americans’ well-being. Policy and practice interventions need to focus on how to improve neighborhood physical conditions, including street and road quality, yard and sidewalk quality, air quality (Schootman et al., 2006), green spaces (Mass et al., 2009), safety, and transportation. Clean, convenient, and pleasant environments can facilitate social interaction and connectedness with neighbors. Perhaps, more important, efforts are needed to provide opportunities for social integration and social participation with family members, friends, and neighbors, which further increase social resources that are important to the health of older adults. Besides, a variety of activities and services needs to be offered to meet the unique needs of native-born and immigrant populations. Notably, with the traditional culture of collectivism that emphasizes interdependence among in-group members, Chinese Americans likely form and rely on the community of their own. Although it is important to affirm cultural and ethnic identities, it is not necessary to separate from the broader social systems and promote ethnospecific communities that will lead to social exclusion (Sonn, 2002). In addition, continuing to provide support and financial funding at various government and community levels will not only help to build aging and minority-friendly livable neighborhoods but also help to promote aging in place and healthy aging.

To sum up, this study improves our knowledge about how neighborhood characteristics and living arrangements are related to older adults’ health in older Chinese Americans. Older adults face much higher risks of functional and cognitive decline than do younger adults, thus more relying on living environment and social support to maintain their functioning. In addition to individual factors, social and physical conditions of the living environment have significant impacts on older adults’ well-being. This study shows that neighborhood disorder is consistently associated with negative health aspects and that social cohesion has mental health protective effects among older Chinese Americans. Policies and practice interventions are needed to improve living environments and to facilitate activity participation, social engagement, self-care behaviors, and health promotion with the ultimate goal of healthy aging.