Racial/Ethnic Variations in Social Determinants of Mental Health Among Medicare Advantage Beneficiaries

Abstract

Objectives:

We examine associations between social determinants and mental health and assess how the associations vary by race/ethnicity using a large, diverse sample of older adults.

Method:

A retrospective study of 444,057 older adults responding to the Medicare Health Outcomes Survey in 2015–2017 was conducted. Using a multilevel linear regression, we examined the associations between the self-reported number of unhealthy days due to mental health and social determinants, stratified by race/ethnicity.

Results:

Health factors were most strongly associated with unhealthy days across all racial/ethnic groups. Strength of other factors varied by race/ethnicity. Social/economic factors had stronger associations among Whites, Asians, and multiracial individuals, while such factors were not significant for American Indians/Alaska Natives and Native Hawaiians/Other Pacific Islanders.

Discussion:

We found varying degrees of associations between social determinants and poor mental health by racial/ethnic groups. These results suggest that homogeneous interventions may not meet the mental health needs of all.

Keywords

race/ethnicity mental health unhealthy days social determinants of health disparities Medicare advantage

Background and Objectives

Mental health plays a critical role in the well-being, functional and cognitive independence, and health of older adults. Approximately 20% of adults 55 years and older report some type of mental health issue (Centers for Disease Control and Prevention [CDC] & National Association of Chronic Disease Directors, 2008). Poor mental health is associated with higher health care costs (Figueroa et al., 2020), reduced quality of life (Barnes et al., 2012), increased risk of disability and morbidity (Fiske et al., 2009), more frequent hospitalizations and nursing home placements (Sheeran et al., 2010), and higher rates of mortality (Whooley & Browner, 1998). The mental health needs of older adults are not adequately met and are more likely to be underestimated, underdiagnosed, and undertreated in comparison with middle-aged and younger adults (Karlin et al., 2008). Indicators of perceived poor mental health, such as the number of unhealthy mental health days, are imperative for estimating mental health needs and are often the first step in the process of initiating mental health care services (Assari, 2018).

Racial/ethnic differences in poor mental health among older adults have been observed previously (Alegria et al., 2008; Arévalo et al., 2015; Chen et al., 2019; Glassgow et al., 2019; Gonzales et al., 2021; Hooker et al., 2019; J. H. Ng et al., 2014; Williams & Mohammed, 2009). Many of these studies identify Black–White disparities, and to a lesser extent, Latinx–White disparities, with Black and Latinx populations experiencing the greatest burden of poor mental health in comparison with Whites (Arévalo et al., 2015; Chen et al., 2019; Glassgow et al., 2019; Gonzales et al., 2021; González et al., 2010). Examining racial/ethnic differences in the mental health status of older adults across other racial/ethnic groups such as Asians, American Indians/Alaska Natives, and Native Hawaiians/Other Pacific Islanders is limited. One recent study showed that American Indians/Alaska Natives and Native Hawaiians/ Other Pacific Islanders had worse mental health outcomes in comparison with other racial/ethnic groups (J. Ng et al., 2017). Although these analyses, inclusive of racial/ethnic groups that are typically excluded due to small sample size, provide an opportunity for a more complete understanding of disparities in mental health, there is a dearth of studies examining the distribution of social determinants and their role in patterning mental health across diverse older adults.

A broad body of empirical evidence links the social determinants of health—the conditions in which people are born, live, work, play, and age—with the development, incidence, prevalence, severity, and treatment outcomes of poor mental health across the life course (Alegría et al., 2018; Finegan et al., 2018; Kim, 2008; Sun et al., 2020; Wight et al., 2013; Wilson-Genderson & Pruchno, 2013). One conceptual framework summarizes the social determinants of mental health into key domains, including demographic (e.g., age, gender, race/ethnicity), economic (e.g., income, employment, income inequality), neighborhood (e.g., area-level deprivation, residential segregation, accessibility to services), environmental (e.g., air pollution, natural disasters), and social/cultural (e.g., family and peer relationships, social isolation, discrimination; Lund et al., 2018). Notably, several models have illuminated the connections between the social determinants of mental health and life course theories and situated these determinants within a multilevel (individual and neighborhood) framework (Alegría et al., 2018; Allen et al., 2014; Lund et al., 2018). Although these frameworks serve as a critical organizing tool to elucidate mental health inequalities, we have yet to fully understand the extent to which social determinants affect mental health outcomes across racially and ethnically diverse older adults.

Recent recommendations to advance the dissemination of interventions and policies aimed at addressing the social determinants of mental health suggest examining the differential impact of social determinants across race/ethnicity (Alegría et al., 2018). In a recent study comparing differences in the association between self-rated mental health and the social determinants across 10 racial/ethnic groups, the findings revealed that factors associated with poor mental health status varied across ethnic groups (Moghani Lankarani & Assari, 2017). For example, education was protective against poor mental status for some groups, but not for all groups. The observation of how mental health outcomes are differently shaped by social determinants across racial/ethnic groups may be a function of differences in the cultural interpretation of perceived mental health, experiences of stress that may be dependent upon cultural and community norms, or racial/ethnic variations in the economic, environmental, and social opportunity structures that reflect the unequal distribution of resources (Assari, 2018; Cook et al., 2019; Villatoro et al., 2018). Investigating the variations in the patterns of the social determinants of mental health may provide clues to the mechanisms underlying poor mental health and racial/ethnic disparities. However, data to examine these determinants across a racially/ethnically diverse population of older adults have been limited.

Interest in addressing the mental health needs of older adults is increasing and signals a shift toward embracing culturally tailored strategies that account for the unique needs and experiences of an increasingly diverse population. Yet, the role of social factors in influencing mental health among older adults disaggregated by race/ethnicity and inclusive of groups, such as Asian, American Indians/Alaska Natives, and Native Hawaiians/Other Pacific Islander populations, has not been fully investigated. Although prior studies have documented racial and ethnic disparities in mental health status, to the best of our knowledge, there are no studies that have investigated racial/ethnic differences in the distribution of the social determinants of mental health among a diverse sample of older adults. To address this gap, the aim of this research was to explore the association between social determinants and poor mental health among a large sample of racially and ethnically diverse Medicare Advantage beneficiaries and to assess variations by race and ethnicity.

Research Design and Methods

Data

We conducted a retrospective study using the Medicare Health Outcomes Survey (HOS) baseline surveys fielded in 2015–2017. This survey is required by the Centers for Medicare & Medicaid Services to be conducted annually by Medicare Advantage plans and was available in English and Spanish, and starting in 2016 in Chinese. Among plans with more than 1,200 individuals, the HOS survey sample is a random sample of 1,200 individuals in each Medicare Advantage contract (CMS Health Services Advisory Group, 2016).

A total of 1,697,111 were eligible for the baseline survey in 2015–2017, and 803,357 answered the survey at least one question (response rate: 47.3%). Among these, those who were younger than 65 years (n = 135,716) and those who were residents of Puerto Rico, Virgin Islands, and Guam (n = 15,884) were excluded. We also excluded the respondents who were enrolled in plans with less than 1,200 beneficiaries (n = 31,227) to avoid overrepresentation of individuals who could be sampled every year, unlike larger plans whose beneficiaries are randomly selected for the survey (CMS Health Services Advisory Group, 2016). Finally, we excluded those who did not report race or ethnicity (n = 79,381); our dependent variable, unhealthy days due to mental health (n = 29,049); and missing predictors (n = 68,043). Our final sample included 444,057 respondents. This study was considered exempt from institutional review board approval as it is a secondary analysis of a limited data set (1292614-2).

Dependent Variable—Number of Unhealthy Days Due to Mental Health

The dependent variable was the number of unhealthy days due to mental health during the past 30 days, collected by a survey item: “Now, thinking about your mental health, which includes stress, depression, and problems with emotions, for how many days during the past 30 days was your mental health not good? Please enter a number between ‘0’ and ‘30’ days” (CMS Health Services Advisory Group, 2018). This survey measure was developed and validated by the CDC with the goal of developing a brief and valid measure (Moriarty et al., 2003).

We focus on the unhealthy days measure because of its interpretability. We also conducted a sensitivity test with VR-12 Mental Health Component Summary score as an outcome and found generally similar results. According to a CDC report, previous studies that translated the unhealthy days measure to Spanish language, Asian languages used in the United States, and Norwegian language have found that the unhealthy days measure “have good respondent acceptability (low nonresponse), normal test–retest reliability, good concurrent validity, and good responsiveness to change over time” (CDC, 2000).

Race/Ethnicity

According to the respondents’ self-report, each respondent was categorized into seven racial/ethnic groups: non-Hispanic White (White), non-Hispanic Black/African American (Black), Hispanic, Asian, American Indian/Alaska Native, Native Hawaiian/Other Pacific Islanders, and multiracial. We first classified respondents who reported Hispanic or Latino origin or descent as Hispanic. Among those not reporting Hispanic, we then classified race as White, Black or African American, or American Indian or Alaska Native. Individuals identifying as Indian, Chinese, Filipino, Japanese, Korean, Japanese, Vietnamese, and Other Asian were categorized as Asian. Individuals identifying as Native Hawaiian, Guamanian or Chamorro, Samoan, and other Pacific islanders were categorized as Native Hawaiian/Other Pacific Islander. Finally, those who reported multiple races were classified as multiracial.

Independent Variables—Social Determinants

Economic factors included education and homeownership. Education was grouped into three categories: did not graduate high school, high school graduate or general educational development (GED), 2- or 4-year college, or more. Respondents were categorized as a homeowner if they reported living in a house or apartment that is owned or being bought by themselves, and those who answered to other choices (home being owned by someone in the family, rented, or living without payment) were identified as nonhomeowners.

Neighborhood-level factors were measured using an area measure of neighborhood disadvantage from the Area Deprivation Index (ADI). ADI is a composite measure that identifies neighborhood socioeconomic status at the census block group level. The index is derived from factors such as neighborhood poverty, employment, and housing quality (Kind & Buckingham, 2018; Kind et al., 2014). Each respondent’s nine-digit zip code was linked to the census block group to identify their neighborhood’s ADI. The neighborhood disadvantage variable was dichotomized as most disadvantaged neighborhoods whose ADI scores are above the 85th percentile in the national distribution (Jung et al., 2018; Kind et al., 2014).

Social and cultural factors were measured using language spoken at home and whether someone lived alone. Language at home was identified using a survey item asking, “What language do you mainly speak at home?” The respondents were asked to choose among English, Spanish, Chinese, and some other language, and then those who answered “some other language” were asked to specify their main language spoken at home with a short answer response. We grouped these responses into the 10 most spoken languages defined by the U.S. Census Bureau (2013; English, Spanish, Chinese, French, Tagalog, Vietnamese, Korean, German, Russian, and Italian). Answers with minor typos (e.g., “Veitnamese” was classified as Vietnamese) were included in the classification. Also, answers with subcategories of each language were classified according to the language classification from the U.S. Census Bureau (2015, 2019). For example, “Cantonese” or “Mandarin” was classified as Chinese, and “Filipino” was categorized as Tagalog. The nine non-English languages were then aggregated as non-English in the stratified model analysis. The distribution of the 10 languages is presented in Online Appendix Figure 1 to examine whether the relationship between the language spoken at home and mental health affected our overall results. Individuals responding that they lived alone to the survey item, “Do you live alone or with others?” were coded as living alone, and otherwise coded as not living alone.

Covariates

Demographic factors such as age and gender were controlled in the model as predisposing factors. Age was categorized as 65 to 70, 70 to 75, 75 to 80, 80 to 85, and 85 years or older. Gender was classified as male and female. Age and gender were derived from the database of the Center for Medicare & Medicaid Services (CMS Health Services Advisory Group, 2018).

Health needs were measured with the number of chronic conditions and difficulties in activities of daily living (ADL). A total of 14 chronic conditions were included (high blood pressure, angina pectoris/coronary artery disease, congestive heart failure, myocardial infarction/heart attack, other heart conditions, stroke, emphysema/asthma/chronic obstructive pulmonary disease, Crohn’s disease/ulcerative colitis/inflammatory bowel disease, arthritis hip/knee/hand/wrist, osteoporosis, sciatica, diabetes, depression, and cancer) and grouped as a categorical variable: none, one to two conditions, three to five conditions, and six or more conditions. Difficulties in ADL were identified as 1 if the respondent reported one or more difficulties or inabilities in bathing, dressing, eating, getting in and out of chairs, walking, and using the toilet.

To address potential differences in respondents’ access to care and survey year, we also accounted for differences in health plans, geography, and time. Plan type (health maintenance organization [HMO], preferred provider organization [PPO], other), plan’s tax status (for-profit, nonprofit), and survey year (2015, 2016, 2017) were included in the model. Urban/rural status was defined from the rural–urban commuting area (RUCA) codes. Each respondent’s five-digit zip code was linked to the RUCA code and grouped into metropolitan, micropolitan, and rural areas according to the RUCA code.

Statistical Analysis

Descriptive statistics were used to examine differences between race/ethnicity groups with respect to individual-level and neighborhood-level factors.

To identify protective and risk factors for self-reported mental health, we used multilevel linear regression to account for the clustering of individuals within state and health plans. We report the overall model with all eligible individuals and stratified by race/ethnicity. As this study was designed to be exploratory to identify individual- and neighborhood-level factors associated with mental health in a diverse population of older adults, we did not test a specific hypothesis.

Number of unhealthy days was not normally distributed, and so we tested whether a multilevel negative binomial model that better accounts for right-skewed dependent variables affected our primary analysis. The sensitivity analysis is reported in Online Appendix Table 1. We examined whether adding the disaggregated measure of languages spoken at home changed our results in Online Appendix Table 2. Also, to investigate whether our results are sensitive to item missingness, we conducted multiple imputation on the predictor variables, and the regression results after multiple imputation are shown in Online Appendix Table 3. The analysis was conducted using R v3.6.0 (R Core Team, Vienna, Austria).

Results

On average, the 444,057 older adults in the sample reported 3.5 unhealthy days due to mental health (SD = 7.42). More than 75% of the sample identified as non-Hispanic White, followed by 9.7% identifying as Black or African American and 8.5% identifying as Hispanic. Asians (3.1%), American Indians/Alaska Natives (0.3%), Native Hawaiians/Other Pacific Islanders (0.2%), and individuals reporting multiple racial/ethnic groups (2.5%) accounted for just over 6% of the sample (Table 1). Hispanics (M = 5.5, SD = 9.27) and American Indians/Alaska Natives (M = 5.3, SD = 9.02) reported the most unhealthy days due to mental health, and Whites (M = 3.1, SD = 6.95) reported the fewest unhealthy days (p < .001).

Table 1.

Sample Characteristics of Older Adults Enrolled in Medicare Advantage: The Medicare Health Outcomes Survey, 2015–2017 (N = 444,057).

Variables	Overall	White	Black	Hispanic	Asian	AI/AN	NH/OPI	Multiracial	p
n	444,057	336,381	42,932	37,641	13,583	1,416	783	11,321
(%)		75.8	9.7	8.5	3.1	0.3	0.2	2.5
Number of unhealthy days, M (SD)	3.50 (7.42)	3.10 (6.95)	4.41 (8.28)	5.51 (9.27)	3.58 (7.40)	5.30 (9.02)	4.92 (8.59)	4.94 (8.62)	<.001
Social determinants
Education (%)									<.001
HS+	219,937 (49.5)	182,419 (54.2)	15,075 (35.1)	8,920 (23.7)	7,314 (53.8)	539 (38.1)	243 (31.0)	5,427 (47.9)
HS	144,225 (32.5)	114,311 (34.0)	14,567 (33.9)	8,562 (22.7)	2,574 (19.0)	441 (31.1)	264 (33.7)	3,506 (31.0)
<HS	79,895 (18.0)	39,651 (11.8)	13,290 (31.0)	20,159 (53.6)	3,695 (27.2)	436 (30.8)	276 (35.2)	2,388 (21.1)
Homeownership (%)									<.001
Homeowner	308,770 (69.5)	255,346 (75.9)	20,865 (48.6)	17,947 (47.7)	6,809 (50.1)	745 (52.6)	311 (39.7)	6,747 (59.6)
Nonhomeowner	135,287 (30.5)	81,035 (24.1)	22,067 (51.4)	19,694 (52.3)	6,774 (49.9)	671 (47.4)	472 (60.3)	4,574 (40.4)
ADI (%)									<.001
ADI 0–84	389,161 (87.6)	309,075 (91.9)	27,713 (64.6)	28,606 (76.0)	12,650 (93.1)	732 (51.7)	689 (88.0)	9,364 (82.7)
ADI 85+	54,896 (12.4)	27,306 (8.1)	15,219 (35.4)	9,035 (24.0)	933 (6.9)	352 (24.9)	94 (12.0)	1,957 (17.3)
Language (%)									<.001
English	410,483 (92.4)	333,414 (99.1)	42,440 (98.9)	15,412 (40.9)	6,093 (44.9)	1,331 (94.0)	662 (84.5)	11,131 (98.3)
Non-English	33,574 (7.6)	2,967 (0.9)	492 (1.1)	22,229 (59.1)	7,490 (55.1)	85 (6.0)	121 (15.5)	190 (1.7)
Living status (%)									<.001
Not alone	306,684 (69.1)	234,903 (69.8)	24,865 (48.6)	26,849 (71.3)	11,203 (82.5)	886 (62.6)	583 (74.5)	7,395 (65.3)
Alone	137,373 (30.9)	101,478 (30.2)	18,067 (42.1)	10,792 (28.7)	2,380 (17.5)	530 (37.4)	200 (25.5)	3,926 (34.7)
Covariates
No. of chronic conditions (%)									<.001
None	38,790 (8.7)	31,284 (9.3)	2,104 (4.9)	2,868 (7.6)	1,631 (12.0)	105 (7.4)	62 (7.9)	736 (6.5)
1 to 2	170,378 (38.4)	132,890 (39.5)	15,041 (35.0)	12,626 (33.5)	5,647 (41.6)	428 (30.2)	253 (32.3)	3,493 (30.9)
3 to 5	183,567 (41.3)	136,807 (40.7)	19,673 (45.8)	16,080 (42.7)	5,125 (37.7)	610 (43.1)	326 (41.6)	4,946 (43.7)
6 or more	51,322 (11.6)	35,400 (10.5)	6,114 (14.2)	6,067 (16.1)	1,180 (8.7)	273 (19.3)	142 (18.1)	2,146 (19.0)
Difficulties in ADL (%)									<.001
None	287,630 (64.8)	225,300 (67.0)	24,392 (56.8)	21,001 (55.8)	9,337 (68.7)	732 (51.7)	442 (56.4)	6,426 (56.8)
1 or more	156,427 (35.2)	111,081 (33.0)	18,540 (43.2)	16,640 (44.2)	4,246 (31.3)	684 (48.3)	341 (43.6)	4,895 (43.2)
Age (%)									<.001
65–69	138,067 (31.1)	100,318 (29.8)	15,669 (36.5)	12,765 (33.9)	4,278 (31.5)	515 (36.4)	279 (35.6)	4,243 (37.5)
70–74	122,686 (27.6)	92,794 (27.6)	11,836 (27.6)	10,426 (27.7)	3,745 (27.6)	408 (28.8)	225 (28.7)	3,252 (28.7)
75–79	85,106 (19.2)	65,184 (19.4)	7,803 (18.2)	7,134 (19.0)	2,565 (18.9)	266 (18.8)	143 (18.3)	2,011 (17.8)
80–84	54,914 (12.4)	43,167 (12.8)	4,463 (10.4)	4,232 (11.2)	1,694 (12.5)	139 (9.8)	87 (11.1)	1,132 (10.0)
85+	43,284 (9.7)	34,918 (10.4)	3,161 (7.4)	3,084 (8.2)	1,301 (9.6)	88 (6.2)	49 (6.3)	683 (6.0)
Gender (%)									<.001
Male	186,060 (41.9)	144,426 (42.9)	14,666 (34.2)	15,280 (40.6)	6,303 (46.4)	542 (38.3)	322 (41.1)	4,521 (39.9)
Female	257,997 (58.1)	191,955 (57.1)	28,266 (65.8)	22,361 (59.4)	7,280 (53.6)	874 (61.7)	461 (58.9)	6,800 (60.1)
Plan type (%)									<.001
HMO	306,255 (69.0)	216,865 (64.5)	35,377 (82.4)	33,258 (88.4)	10,798 (79.5)	1,125 (79.4)	^a	8,278 (73.1)
PPO	112,711 (25.4)	96,772 (28.8)	6,337 (14.8)	3,936 (10.5)	2,553 (18.8)	255 (18.0)	^a	2,633 (23.3)
Other	25,091 (5.7)	22,744 (6.8)	1,218 (2.8)	447 (1.2)	232 (1.7)	36 (2.5)	^a	410 (3.6)
Plan tax status (%)									<.001
For-profit	295,316 (66.5)	216,137 (64.3)	33,042 (77.0)	27,448 (72.9)	9,268 (68.2)	933 (65.9)	512 (65.4)	7,977 (70.5)
Not-for-profit	148,741 (33.5)	120,244 (35.7)	9,890 (23.0)	10,193 (27.1)	4,316 (31.8)	483 (34.1)	271 (34.6)	3,344 (29.5)
Urban/rural (%)									<.001
Metropolitan	371,685 (83.7)	272,588 (81.0)	40,275 (93.8)	34,942 (92.8)	13,015 (95.8)	1,062 (75.0)	646 (82.5)	9,157 (80.9)
Micropolitan	38,785 (8.7)	33,475 (10.0)	1,603 (3.7)	1,781 (4.7)	429 (3.2)	182 (12.9)	97 (12.4)	1,218 (10.8)
Rural	33,587 (7.6)	30,318 (9.0)	1,054 (2.5)	918 (2.4)	139 (1.0)	172 (12.1)	40 (5.1)	946 (8.4)
Survey year (%)									<.001
2015	161,441 (36.4)	122,615 (36.5)	15,766 (36.7)	13,556 (36.0)	4,720 (34.7)	501 (35.4)	263 (33.6)	4,020 (35.5)
2016	145,849 (32.8)	110,753 (32.9)	13,910 (32.4)	12,431 (33.0)	4,367 (32.2)	433 (30.6)	279 (35.6)	3,676 (32.5)
2017	136,767 (30.8)	103,013 (30.6)	13,256 (30.9)	11,654 (31.0)	4,496 (33.1)	482 (34.0)	241 (30.8)	3,625 (32.0)

Note. AI/AN = American Indian/Alaska Native; NH/OPI = Native Hawaiian/Other Pacific Islander; HS = high school; ADI = Area Deprivation Index; ADL = activities of daily living; HMO = health maintenance organization; PPO = preferred provider organization; CMS = Centers for Medicare & Medicaid Services.

Intentionally left blank to follow CMS Data Usage Agreement requirements.

Across all groups, more than half were female. The age distribution skewed older for Whites (10.4% were 85 years and older, and 29.8% were 65–69 years) compared with other groups such as Blacks with 7.4% of age 85 years and older and 36.5% of age 65 to 69 years and American Indians/Alaska Natives with 6.2% of age 85 years and older and 36.4% of age 65 to 69 (p < .001) years. Asians and Whites reported the highest rates of not living alone (82.5% Asians and 69.8% Whites). Hispanics and Asians had higher proportions of respondents who reported speaking a language other than English at home.

With respect to individual-level socioeconomic status, Whites reported higher education and homeownership than other groups. Asians reported similar education levels to Whites, but substantially lower homeownership rates (Asians: 50.1% vs. Whites: 75.9%). Asians (6.9%) and Whites (8.1%) also had the lowest proportion of respondents living in one of the most disadvantaged neighborhoods; Blacks had the highest proportion of respondents living in the most disadvantaged neighborhoods (35.4%). More than 15% of Hispanics, multiracial individuals, and American Indians/Alaska Natives reported living with six or more chronic conditions. More than 40% of individuals in these same groups and Blacks also reported living with one more difficulty with ADL.

Table 2 reports the association of individual- and neighborhood-level factors with the number of unhealthy days due to mental health by race and ethnicity group. Across all groups, we find that health factors, number of chronic conditions, and difficulties with ADL are strongly associated with more unhealthy days. Between groups, the magnitude of the association varies from as many as 7.5 unhealthy days (95% confidence interval [CI] = [5.6, 9.4]) among American Indians/Alaska Natives reporting six or more chronic conditions to an average of 4.8 more unhealthy days for Whites reporting six or more chronic conditions (95% CI = [4.7, 4.9]). Having more than six chronic conditions or having difficulties in ADL was associated with more unhealthy days among Blacks, Hispanics, and Asians compared with Whites.

Table 2.

Multilevel Linear Regression on the Number of Unhealthy Days: Overall model and Stratified Models by Race and Ethnicity, the Medicare Health Outcomes Survey, 2015–2017 (N = 444,057).

Variables	Overall	White	Black	Hispanic	Asian	AI/AN	NH/OPI	Multiracial
Variables	Coefficients (95% CI)	Coefficients (95% CI)	Coefficients (95% CI)	Coefficients (95% CI)	Coefficients (95% CI)	Coefficients (95% CI)	Coefficients (95% CI)	Coefficients (95% CI)
Social determinants
Education (Ref: HS+)
HS	0.29*** [0.25, 0.34]	0.31*** [0.26, 0.36]	0.17* [−0.01, 0.34]	0.31** [0.06, 0.56]	0.39** [0.09, 0.69]	0.50 [−0.53, 1.52]	−1.57** [−2.88, −0.25]	0.50*** [0.16, 0.83]
<HS	0.93*** [0.86, 0.99]	1.02*** [0.95, 1.10]	0.62*** [0.43, 0.81]	0.68*** [0.45, 0.91]	0.95*** [0.66, 1.24]	0.59 [−0.49, 1.66]	0.35 [−1.03, 1.73]	1.07*** [0.67, 1.46]
Homeownership (Ref: Homeowner)
Nonhomeowner	0.88*** [0.83, 0.93]	0.98*** [0.92, 1.03]	0.68*** [0.52, 0.83]	0.81*** [0.62, 0.99]	0.73*** [0.48, 0.98]	1.06** [0.15, 1.97]	0.90 [−0.27, 2.07]	1.02*** [0.70, 1.33]
ADI (Ref: 0–84)
ADI 85+	0.17*** [0.11, 0.24]	0.22*** [0.13, 0.30]	0.16* [−0.0005, 0.32]	0.08 [−0.14, 0.29]	0.44* [−0.003, 0.88]	−0.77 [−1.77, 0.23]	−1.13 [−2.80, 0.54]	0.10 [−0.29, 0.50]
Language (Ref: English)
Non-English	0.54*** [0.45, 0.63]	0.40*** [0.17, 0.64]	0.09 [−0.60, 0.78]	0.45*** [0.25, 0.66]	0.30** [0.05, 0.56]	−0.30 [−2.16, 1.57]	0.51 [−1.02, 2.03]	1.36** [0.22, 2.50]
Living status (Ref: Not alone)
Alone	0.09*** [0.04, 0.13]	0.07*** [0.02, 0.12]	−0.01 [−0.16, 0.14]	0.28*** [0.09, 0.48]	0.57*** [0.27, 0.87]	−0.35 [−1.23, 0.53]	0.51 [−0.73, 1.75]	0.17 [−0.15, 0.48]
Covariates
No. of chronic conditions (Ref: None)
1 to 2	0.35*** [0.27, 0.43]	0.35*** [0.27, 0.43]	0.35* [−0.003, 0.70]	0.42** [0.08, 0.76]	0.31* [−0.05, 0.67]	0.72 [−0.98, 2.43]	−0.24 [−2.31, 1.83]	0.50 [−0.12, 1.13]
3 to 5	1.68*** [1.60, 1.76]	1.57*** [1.49, 1.65]	1.93*** [1.57, 2.28]	2.40*** [2.05, 2.74]	1.90*** [1.52, 2.28]	1.98** [0.28, 3.68]	1.71 [−0.38, 3.80]	1.99*** [1.36, 2.61]
6 or more	5.13*** [5.03, 5.22]	4.77*** [4.66, 4.87]	5.71*** [5.31, 6.11]	6.34*** [5.94, 6.74]	6.14*** [5.62, 6.66]	7.47*** [5.56, 9.39]	5.05*** [2.68, 7.42]	5.89*** [5.20, 6.59]
Difficulties in ADL (Ref: None)
1 or more	3.56*** [3.51, 3.61]	3.33*** [3.28, 3.39]	3.77*** [3.61, 3.94]	4.69*** [4.49, 4.88]	4.55*** [4.27, 4.82]	4.17*** [3.21, 5.13]	5.60*** [4.41, 6.80]	3.87*** [3.54, 4.20]
Age (Ref: 65–69)
70–74	−0.68*** [−0.73, −0.62]	−0.67*** [−0.73, −0.61]	−0.70*** [−0.88, −0.51]	−0.83*** [−1.04, −0.61]	−0.32** [−0.61, −0.04]	−1.15** [−2.19, −0.11]	−0.96 [−2.26, 0.35]	−0.52*** [−0.88, −0.16]
75–79	−1.01*** [−1.07, −0.95]	−1.03*** [−1.09, −0.97]	−0.97*** [−1.19, −0.76]	−0.96*** [−1.20, −0.71]	−0.36** [−0.68, −0.03]	−1.52** [−2.71, −0.32]	−0.32 [−1.84, 1.20]	−1.29*** [−1.71, −0.87]
80−84	−1.22*** [−1.28, −1.15]	−1.25*** [−1.33, −1.18]	−1.14*** [−1.40, −0.88]	−1.10*** [−1.40, −0.81]	−0.33* [−0.71, 0.04]	−1.42* [−2.94, 0.09]	−0.55 [−2.37, 1.28]	−1.43*** [−1.95, −0.91]
85+	−1.18*** [−1.25, −1.10]	−1.31*** [−1.39, −1.23]	−0.58*** [−0.88, −0.28]	−0.56*** [−0.89, −0.22]	−0.14 [−0.56, 0.29]	−0.71 [−2.56, 1.13]	−1.07 [−3.39, 1.25]	−1.00*** [−1.65, −0.36]
Gender (Ref: Male)
Female	0.40*** [0.36, 0.44]	0.48*** [0.43, 0.52]	−0.24*** [−0.39, −0.08]	0.25*** [0.07, 0.42]	0.26** [0.03, 0.48]	0.28 [−0.58, 1.15]	−0.54 [−1.61, 0.54]	0.53*** [0.23, 0.83]
Plan type (Ref: HMO)
PPO	−0.31*** [−0.43, −0.20]	−0.29*** [−0.40, −0.18]	−0.48*** [−0.71, −0.24]	−0.38** [−0.72, −0.03]	−0.22 [−0.59, 0.14]	−1.04* [−2.23, 0.14]	−0.74 [−2.36, 0.88]	−0.13 [−0.50, 0.25]
Other	−0.43*** [−0.67, −0.19]	−0.41*** [−0.64, −0.18]	−0.09 [−0.63, 0.44]	−0.99** [−1.86, −0.12]	−0.65 [−1.62, 0.31]	−1.08 [−3.96, 1.79]	−0.65 [−8.03, 6.73]	0.49 [−0.34, 1.32]
Plan tax status (Ref: For-profit)
Not-for-profit	−0.12** [−0.22, −0.02]	−0.13** [−0.23, −0.03]	−0.24** [−0.45, −0.03]	0.07 [−0.20, 0.34]	−0.16 [−0.45, 0.14]	−0.76 [−1.73, 0.21]	0.11 [−1.26, 1.47]	−0.09 [−0.44, 0.26]
Urban/Rural (Ref: Metropolitan)
Micropolitan	0.07* [−0.01, 0.15]	0.06 [−0.02, 0.14]	0.03 [−0.36, 0.43]	0.11 [−0.31, 0.52]	−0.67** [−1.33, −0.01]	0.10 [−1.22, 1.42]	0.46 [−1.40, 2.32]	0.05 [−0.43, 0.53]
Rural	0.05 [−0.04, 0.13]	0.01 [−0.08, 0.09]	0.77*** [0.28, 1.25]	0.05 [−0.52, 0.62]	−0.34 [−1.45, 0.77]	0.45 [−0.92, 1.83]	−0.49 [−3.15, 2.18]	0.14 [−0.41, 0.68]
Survey year (Ref: 2015)
2016	0.03 [−0.02, 0.08]	0.01 [−0.05, 0.06]	0.01 [−0.17, 0.19]	0.15 [−0.05, 0.36]	0.13 [−0.14, 0.41]	0.48 [−0.55, 1.51]	−1.10* [−2.38, 0.18]	0.24 [−0.12, 0.59]
2017	0.08*** [0.03, 0.14]	0.09*** [0.03, 0.14]	−0.07 [−0.25, 0.11]	0.24** [0.03, 0.45]	0.04 [−0.23, 0.31]	−0.10 [−1.11, 0.90]	−0.11 [−1.45, 1.23]	0.33* [−0.03, 0.68]
Constant	0.78*** [0.67, 0.90]	0.86*** [0.73, 0.98]	1.13*** [0.74, 1.52]	0.30 [−0.11, 0.71]	−0.14 [−0.59, 0.30]	1.21 [−0.75, 3.16]	2.12* [−0.32, 4.56]	0.35 [−0.34, 1.03]
Observations	444,057	336,381	42,932	37,641	13,583	1,416	783	11,321

Note. The p-values indicate the significance of the coefficient in comparison with the reference group of each variable (e.g., age: 65–69, survey year: 2015). CI = confidence interval; AI/AN = American Indian/Alaska Native; NH/OPI = Native Hawaiian/Other Pacific Islander; HS = high school; ADI = Area Deprivation Index; ADL = activities of daily living; HMO = health maintenance organization; PPO = preferred provider organization.

p < .1. **p < .05. ***p < .01.

Across individual-level sociodemographic factors, in general, we found that individuals reporting less education were statistically significantly more likely to report more unhealthy days compared with those with more education. The relationship between having a high school degree or less and unhealthy days ranged from 0.6 more unhealthy days (95% CI = [0.43, 0.81]) among Black older adults to 1.1 more unhealthy days (95% CI = [0.67, 1.46]) among multiracial older adults. Individuals who reported not owning their homes, nonhomeowners, were more likely to report more unhealthy days for all racial/ethnic groups, except that it was not statistically significant among Native Hawaiian/Other Pacific Islanders.

Living in one of the most disadvantaged neighborhoods was associated with statistically significantly more unhealthy days among Whites, Blacks, and Asians. The relationship between neighborhood disadvantage and mental health, however, was not strong—on average less than 0.5 days.

We find that increasing age was negatively associated with the reported number of unhealthy days, although not all groups had statistically significant associations.

Results of the sensitivity test using a negative binomial model were generally consistent with the main findings. The full results are reported in Online Appendix Table 1. We also examined whether the findings differed using a disaggregated measure of language (Online Appendix Table 2) and multiple imputation (Online Appendix Table 3). These sensitivity tests did not meaningfully differ from the main analysis.

Discussion and Implications

We examined how the associations between social determinants and poor mental health vary by race/ethnicity in a nationally representative sample of Medicare Advantage beneficiaries.

Although we expected individual and neighborhood socioeconomic status indicators to be strongly related to mental health, we found that these factors were less salient among American Indians/Alaska Natives and Native Hawaiians/Other Pacific Islanders than other racial/ethnic groups. Similarly, we found that speaking a language other than English at home was associated with unhealthy days in different directions: more than one additional unhealthy day among individuals reporting multiple racial/ethnic groups; less than one additional unhealthy days among Whites, Hispanics, and, Asians; and not significant among Blacks, American Indians/Alaska Natives, and Native Hawaiians/Other Pacific Islanders.

These results are consistent with previous studies examining differences in mental health among racial/ethnic subgroups using the Medicare HOS. J. H. Ng et al. (2014) found that Blacks, Hispanics, American Indians/Alaska Natives, and Native Hawaiians/Pacific Islanders have poor health using the VR-12 Mental Component Summary score and were more likely to report depressed mood. J. H. Ng et al. (2014) also found that the degree of poor mental health varied by racial/ethnic groups in the same direction as we found, with the highest risk of poor mental health among Hispanics and the lowest among Whites. Similarly, Hooker et al. (2019) found Mexicans, Puerto Ricans, Cubans, African Americans, Asian Indians, Filipinos, and Native Hawaiians/Pacific Islanders were more likely to be screened positively for depression compared with non-Hispanic Whites. Our study adds to the literature by examining which neighborhood-level and social and cultural factors (primary language at home) could be associated with mental health across a broad range of racial and ethnic older adults.

While the individual effects of each social determinant of health and neighborhood-level factor were relatively small (less than 1 unhealthy day), the cumulative effect of multiple economic, neighborhood, and social/cultural factors can be substantial. For example, holding all other factors constant, a Black woman aged 65, living alone in a rural area in a disadvantaged neighborhood, and with less than high school education would have an average of 2.6 unhealthy days in a month. With these same characteristics, an Asian would have an average of 1.7 unhealthy days, a Native Hawaiian/Other Pacific Islander would have an average of 0.8 unhealthy days, and a multiracial woman would have an average of 2.4 unhealthy days in a month.

In addition to using area-level measures and race/ethnicity to identify individuals at risk of poor mental health, health plans could consider routinely reporting patient-level information on social determinants of health using the International Classification of Diseases, Tenth Revision (ICD-10) Z codes (Allsopp & Kinderman, 2017). Laura Gottlieb and colleagues compared ICD-10 Z codes with many important domains of social well-being but did not find codes that addressed neighborhood safety (Gottlieb et al., 2016). Medicare Advantage plans have robust systems in place to capture diagnosis codes used in the Medicare Advantage risk adjustment payments, including chart reviews and home visits, which could be leveraged for tracking social determinants of health.

This study has several limitations. First, the sample of this study was limited to Medicare Advantage beneficiaries aged 65 years and older who answered to race, ethnicity, and mentally unhealthy days. Our results may not reflect the experience of those enrolled in traditional fee-for-service Medicare, Medicare Advantage beneficiaries aged 65 years and younger, and those living in U.S. territories. Second, this study was not able to investigate causal relationships due to its cross-sectional research design. Third, this study did not consider the diversity within each racial/ethnic group. There might be different contexts that affect mental health among each subgroup; for example, among Asian subgroups, Chinese, Filipino, Japanese, Korean, and Vietnamese have different immigration experiences, cultural backgrounds, and community resources affecting mental health. Fourth, this study relies on individual survey responses, and it is possible that individuals may underreport mental health issues due to respondent bias. Finally, it is possible that the respondents answering the survey in English versus Chinese or Spanish may have interpreted the questions differently.

The results of this study suggest that the strength of the relationship between physical health and mental health varies by racial/ethnic group and that commonly used socioeconomic status indicators to identify individuals at high risk of poor health outcomes may not work well among American Indians/Alaska Natives and Native Hawaiians/Other Pacific Islanders. This exploratory study cannot address the causal relationship among the economic, neighborhood, and social and cultural factors associated with mental health among different racial/ethnic groups, and how health care and public health interventions can mediate these relationships. Recent Medicare Advantage initiatives have focused on value-based insurance designs to target program benefits to those who need them most. To address mental health in a diverse population of Medicare Advantage beneficiaries, health plans may want to consider multiple approaches for defining high-risk groups.

Supplemental Material

sj-pdf-1-jag-10.1177_07334648211039311 – Supplemental material for Racial/Ethnic Variations in Social Determinants of Mental Health Among Medicare Advantage Beneficiaries

Supplemental material, sj-pdf-1-jag-10.1177_07334648211039311 for Racial/Ethnic Variations in Social Determinants of Mental Health Among Medicare Advantage Beneficiaries by Taehyun Kim, Kellee White and Eva DuGoff in Journal of Applied Gerontology

Footnotes

Declaration of Conflicting Interests

The authors declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: E.D. received grant support from the Commonwealth Fund, speaking fees from Zimmer Biomet, and fees for consulting for the Berkeley Research Group.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

IRB Approval

This study was approved by the institutional review board (IRB) at the University of Maryland College Park (UMCP), which determined it not to be human subject research and to be exempt from IRB review according to federal regulations.

ORCID iD

Taehyun Kim

Supplemental Material

Supplemental material for this article is available online.

References

Alegria

Chatterji

Wells

Cao

Chen

C.-n.

Takeuchi

Jackson

Meng

X.-L.

(2008). Disparity in depression treatment among racial and ethnic minority populations in the United States. Psychiatric Services, 59(11), 1264–1272. https://doi.org/10.1176/appi.ps.59.11.1264

Alegría

NeMoyer

Falgàs Bagué

Wang

Alvarez

(2018). Social determinants of mental health: Where we are and where we need to go. Current Psychiatry Reports, 20(11), Article 95. https://doi.org/10.1007/s11920-018-0969-9

Allen

Balfour

Bell

Marmot

(2014). Social determinants of mental health. International Review of Psychiatry, 26(4), 392–407. https://doi.org/10.3109/09540261.2014.928270

Allsopp

Kinderman

(2017). A proposal to introduce formal recording of psychosocial adversities associated with mental health using ICD-10 codes. The Lancet Psychiatry, 4, 664–665. https://doi.org/10.1016/S2215-0366(17)30318-8

Arévalo

S. P.

Tucker

K. L.

Falcón

L. M.

(2015). Beyond cultural factors to understand immigrant mental health: Neighborhood ethnic density and the moderating role of pre-migration and post-migration factors. Social Science and Medicine, 138, 91–100. https://doi.org/10.1016/j.socscimed.2015.05.040

Assari

(2018). Ethnic groups differ in how poor self-rated mental health reflects psychiatric disorders. Journal of Racial and Ethnic Health Disparities, 5(4), 728–736. https://doi.org/10.1007/s40615-017-0417-2

Barnes

A. L.

Murphy

M. E.

Fowler

C. A.

Rempfer

M. V.

(2012). Health-related quality of life and overall life satisfaction in people with serious mental illness. Schizophrenia Research and Treatment, 2012, Article 245103. https://doi.org/10.1155/2012/245103

Centers for Disease Control and Prevention. (2000). Measuring healthy days: Population assessment of health-related quality of life. https://www.cdc.gov/hrqol/pdfs/mhd.pdf

Centers for Disease Control and Prevention, & National Association of Chronic Disease Directors. (2008). The State of Mental Health and Aging in America Issue Brief 1: What do the data tell us? National Association of Chronic Disease Directors.

10.

Chen

Kessler

R. C.

Sadikova

NeMoyer

Sampson

N. A.

Alvarez

Vilsaint

C.L.

Green

J. G.

McLaughlin

K. A.

Jackson

J. S.

Alegría

Williams

D. R.

(2019). Racial and ethnic differences in individual-level and area-based socioeconomic status and 12-month DSM-IV mental disorders. Journal of Psychiatric Research, 119, 48–59. https://doi.org/10.1016/j.jpsychires.2019.09.006

11.

CMS Health Services Advisory Group. (2016). Medicare HOS 2015 Cohort 18 baseline data user’s guide. https://www.hosonline.org/globalassets/hos-online/dugs/qio-baseline/hos_dug_c18b.pdf

12.

CMS Health Services Advisory Group. (2018). Medicare health outcomes survey limited data sets file specifications for Cohorts 14, 15, 16, 17, and 18. https://www.hosonline.org/globalassets/hos-online/hos_cohorts-14-18-lds-file-specifications.pdf

13.

Cook

B. L.

Hou

S. S. Y.

Lee-Tauler

S. Y.

Progovac

A. M.

Samson

Sanchez

M. J.

(2019). A review of mental health and mental health care disparities research: 2011–2014. Medical Care Research and Review, 76, 683–710. https://doi.org/10.1177/1077558718780592

14.

Figueroa

J. F.

Phelan

Orav

E. J.

Patel

Jha

A. K.

(2020). Association of mental health disorders with health care spending in the Medicare population. The Journal of the American Medical Association Network Open, 3(3), Article e201210. https://doi.org/10.1001/jamanetworkopen.2020.1210

15.

Finegan

Firth

Wojnarowski

Delgadillo

(2018). Associations between socioeconomic status and psychological therapy outcomes: A systematic review and meta-analysis. Depression and Anxiety, 35(6), 560–573. https://doi.org/10.1002/da.22765

16.

Fiske

Wetherell

J. L.

Gatz

(2009). Depression in older adults. Annual Review of Clinical Psychology, 5, 363–389. https://doi.org/10.1097/01.NAJ.0000422251.65212.4b

17.

Glassgow

A. E.

Gerges

Atkins

Martin

Caskey

Sanders

Mirza

Voorhees

B. V.

Kim

(2019). Exploring racial disparities in mental health diagnoses and neighborhood disorganization among an urban Cohort of children and adolescents with chronic medical conditions. Health Equity, 3(1), 604–611. https://doi.org/10.1089/heq.2019.0085

18.

Gonzales

Lee

Y. J.

Marchiondo

L. A.

(2021). Exploring the consequences of major lifetime discrimination, neighborhood conditions, chronic work, and everyday discrimination on health and retirement. Journal of Applied Gerontology, 40(2), 121–131. https://doi.org/10.1177/0733464819892847

19.

González

H. M.

Tarraf

Whitfield

K. E.

Vega

W. A.

(2010). The epidemiology of major depression and ethnicity in the United States. Journal of Psychiatric Research, 44(15), 1043–1051. https://doi.org/10.1016/j.jpsychires.2010.03.017

20.

Gottlieb

Tobey

Cantor

Hessler

Adler

N. E.

(2016). Integrating social and medical data to improve population health: Opportunities and barriers. Health Affairs, 35(11), 2116–2123. https://doi.org/10.1377/hlthaff.2016.0723

21.

Hooker

Phibbs

Irvin

V. L.

Mendez-Luck

C. A.

Doan

L. N.

Turner

Choun

(2019). Depression among older adults in the United States by disaggregated race and ethnicity. The Gerontologist, 59(5), 886–891. https://doi.org/10.1093/geront/gny159

22.

Jung

Kind

Robert

Buckingham

DuGoff

(2018). Linking neighborhood context and health in community-dwelling older adults in the Medicare Advantage program. Journal of the American Geriatrics Society, 66(6), 1158–1164. https://doi.org/10.1111/jgs.15366

23.

Karlin

B. E.

Duffy

Gleaves

D. H.

(2008). Patterns and predictors of mental health service use and mental illness among older and younger adults in the United States. Psychological Services, 5(3), 275–294. https://doi.org/10.1037/1541-1559.5.3.275

24.

Kim

(2008). Blues from the neighborhood? Neighborhood characteristics and depression. Epidemiologic Reviews, 30, 101–117. https://doi.org/10.1093/epirev/mxn009

25.

Kind

A. J. H.

Buckingham

W. R.

(2018). Making neighborhood-disadvantage metrics accessible—The neighborhood atlas. New England Journal of Medicine, 378, 2456–2458. https://doi.org/10.1056/NEJMp1802313

26.

Kind

A. J. H.

Jencks

Brock

Bartels

Ehlenbach

Greenberg

Smith

(2014). Neighborhood socioeconomic disadvantage and 30-day rehospitalization: A retrospective cohort study. Annals of Internal Medicine, 161(11), 765–774. https://doi.org/10.7326/M13-2946

27.

Lund

Brooke-Sumner

Baingana

Baron

E. C.

Breuer

Chandra

Haushofer

Herrman

Jordans

Kieling

Medina-Mora

M. E.

Morgan

Omigbodun

Tol

Patel

Saxena

(2018). Social determinants of mental disorders and the Sustainable Development Goals: A systematic review of reviews. The Lancet Psychiatry, 5, 357–369. https://doi.org/10.1016/S2215-0366(18)30060-9

28.

Moghani Lankarani

Assari

. (2017). Demographic and socioeconomic determinants of physical and mental self-rated health across 10 ethnic groups in the United States. International Journal of Epidemiologic Research, 4(3), 185–193. https://doi.org/10.15171/ijer.2017.02

29.

Moriarty

D. G.

Zack

M. M.

Kobau

(2003). The Centers for Disease Control and Prevention’s healthy days measures—Population tracking of perceived physical and mental health over time. Health and Quality of Life Outcomes, 1, Article 37. https://doi.org/10.1186/1477-7525-1-37

30.

Zhang

Scholle

S. H.

Guerino

(2017). Racial and ethnic disparities in mental health among diverse groups of Medicare Advantage beneficiaries. CMS OMH Data Highlight No. 11. Baltimore, MD. 2017. https://www.cms.gov/About-CMS/Agency-Information/OMH/research-and-data/information-products/data-highlights/vol-11-dec-2017

31.

J. H.

Bierman

A. S.

Elliott

M. N.

Wilson

R. L.

Xia

Scholle

S. H.

(2014). Beyond Black and White: Race/ethnicity and health status among older adults. American Journal of Managed Care, 20(3), 239–248.

32.

Sheeran

Byers

A. L.

Bruce

M. L.

(2010). Depression and increased short-term hospitalization risk among geriatric patients receiving home health care services. Psychiatric Services, 61(1), 78–80. https://doi.org/10.1176/ps.2010.61.1.78

33.

Sun

Hua

C. L.

Qiu

Brown

J. S.

(2020). Urban and rural differences in trajectories of depressive symptoms in later life in the United States. Journal of Applied Gerontology. Advance online publication. https://doi.org/10.1177/0733464820972527

34.

U.S. Census Bureau. (2013, February 14). Top languages other than English spoken in 1980 and changes in relative rank, 1990–2010. https://www.census.gov/dataviz/visualizations/045/508.php

35.

U.S. Census Bureau. (2015, October). Detailed languages spoken at home and ability to speak English. https://www.census.gov/data/tables/2013/demo/2009-2013-lang-tables.html

36.

U.S. Census Bureau. (2019, May 19). About language use in the U.S. population.

37.

Villatoro

A. P.

Mays

V. M.

Ponce

N. A.

Aneshensel

C. S.

(2018). Perceived need for mental health care: The intersection of race, ethnicity, gender, and socioeconomic status. Society and Mental Health, 8(1), 1–24. https://doi.org/10.1177/2156869317718889

38.

Whooley

M. A.

Browner

W. S.

(1998). Association between depressive symptoms and mortality in older women. Archives of Internal Medicine, 158(19), 2129–2135. https://doi.org/10.1001/archinte.158.19.2129

39.

Wight

R. G.

Aneshensel

C. S.

Barrett

Chodosh

Karlamangla

A. S.

(2013). Urban neighbourhood unemployment history and depressive symptoms over time among late middle age and older adults. Journal of Epidemiology and Community Health, 67(2), 153–158. https://doi.org/10.1136/jech-2012-201537

40.

Williams

D. R.

Mohammed

S. A.

(2009). Discrimination and racial disparities in health: Evidence and needed research. Journal of Behavioral Medicine, 32(1), 20–47. https://doi.org/10.1007/s10865-008-9185-0

41.

Wilson-Genderson

Pruchno

(2013). Effects of neighborhood violence and perceptions of neighborhood safety on depressive symptoms of older adults. Social Science and Medicine, 85, 43–49. https://doi.org/10.1016/j.socscimed.2013.02.028

Supplementary Material

Please find the following supplemental material available below.

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

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

0.00 MB

0.23 MB