Stress Exposure and Physical Health among Older African American and Caribbean Black Women

Abstract

Objectives:

This study assessed whether multiple stress exposures and stress accumulation explained differences in physical health among Afro-Caribbean and African American women in older adulthood. Whether specific stressors uniquely influenced the health of African American and Afro-Caribbean women was also examined.

Methods:

Data were drawn from the National Survey of American Life (NSAL) (N=867; 50 years and older). Physical health was assessed by multiple chronic conditions, functional limitations, and self-rated health. Weighted binary logistic regression and ordered logistic regression analyses were conducted.

Results:

Compared to Afro-Caribbean women, African American women had worse physical health and greater stress exposure. Nonetheless, stress exposure did not explain ethnic differences in physical health. The association between specific measures of stress and physical health was dependent on the stressor and physical health measure.

Discussion:

Findings emphasize the importance of ethnic variation in health and stress exposure among older US Black women. Study implications are discussed.

“Conducting research with ethnically diverse populations is critical because of changing immigration patterns from the Caribbean and Africa, diverging socioeconomic conditions, shifting black family structure, and disparities in social indicators of health and well-being, including living conditions and access to services. The incorporation of blacks of Caribbean descent in the NSAL [National Survey of American Life] will facilitate comparative empirical analyses of issues never before addressed due to limited conceptualizations of heterogeneity of experiences across ethnic groups within the black population and the limited sample sizes of previous research.” (Jackson et al., 2012, pp. 187–188).

As one of the principal investigators of the National Survey of American Life, Dr James S. Jackson profoundly influenced the development of research examining ethnic distinctions in various social outcomes within the US Black population. The current study heeds the call for research to investigate ethnic heterogeneity among Black Americans¹ by focusing specifically on stress exposure and physical health among older Black women, as this population is particularly vulnerable with regards to physical health. Among Black women 50 years of age and older, nearly two-thirds have multiple chronic conditions (MCC) (i.e., two or more chronic physical conditions) (Ward & Schiller, 2013). Compared to other race-gender groups, Black women experience relatively worse health on virtually all dimensions (Brown et al., 2016). For example, older Black women report higher functional limitations (Lin, 2020; Warner & Brown, 2011) and lower self-rated health (Brown et al., 2016) relative to White women, White men, and Black men.

Though Black women remain one of the most health disadvantaged race–gender groups in the US, ethnic variation in Black women’s physical health is understudied. Nonetheless, the Black population is becoming increasingly more diverse due to significant migration streams from Caribbean area nations such as Haiti, Jamaica, and Trinidad/Tobago. In fact, 18% of the US Black population is either an immigrant or the child of an immigrant (Anderson & López, 2018). Furthermore, more than 25% of Black Americans living in the New York City, Miami, and Ft. Lauderdale metropolitan areas are of Caribbean descent (Logan, 2007). Moreover, 55% of Caribbean immigrants residing in the US are female (Thomas, 2012).

In addition to the concern for understanding ethnic variation in physical health among Black women, this study focuses on how stress exposure influences their health. The stress process model (SPM), a widely utilized framework in sociological health research, grounds the current study (Pearlin & Bierman, 2013; Pearlin, Menaghan, Lieberman, & Mullan, 1981). Three core tenets of SPM are germane. First, compared to those occupying historically advantaged social statuses, individuals with disadvantaged social statuses are exposed to a disproportionately higher volume of stressors (Pearlin & Bierman, 2013; Turner, 2013); thus, an examination of stress exposure among older Black women is particularly relevant given their occupation of (at least) three societally devalued statuses (i.e., age, race, and gender). Second, given that individuals from disadvantaged social statuses experience greater exposure to social stress, SPM proposes that stress exposure may explain status distinctions in health (Turner, 2013). Third, because individuals navigate myriad social contexts in their daily lives, stressors do not occur in a social vacuum; instead, stressors tend to accumulate, particularly among individuals who have experienced a lifetime of cumulative disadvantage (Pearlin et al., 2005). Therefore, above and beyond a single stressor, accumulative stress, an indicator of multiple co-occurring stressors, may have a greater magnitude of negative consequences for health (Pearlin & Bierman, 2013; Pearlin, Menaghan, Lieberman, & Mullan, 1981).

The pathways linking social stress to physical health are complex: physiological responses to prolonged chronic stress exposure include increased blood pressure, a reduction in insulin sensitivity, and, over time, a general wear and tear on the body that accumulates to produce physical disease (Felix et al., 2019; Geronimus, 1992; Geronimus et al., 2006). Though stressors tend to be studied in isolation, assessing the health effects of multiple stressors provides a more comprehensive assessment of how stress exposure influences health (Cuevas et al., 2020; Wheaton & Montazer, 2017). This study makes use of national data to assess how various stress exposures influence the physical health of older US Black women with particular attention to ethnic variation within this population.

Older Black Women’s Stress Exposure and Physical Health

As a marginalized race, gender, and age group (Nguyen et al., 2013), older Black women face social disadvantages attributable to the accumulation of compounded stress experienced over the life course (D. Carr, 2019; Ensminger & Juon, 2001). The myriad stressors to which older Black women are exposed span social contexts and life domains, ranging from being the recipients of discriminatory treatment to enduring material hardship and residing in unsafe neighborhoods. One of the most common stressors examined among Black Americans is perceptions of discrimination (Lewis et al., 2015). Chronic exposure to discrimination is associated with poor physical health among Black women (Bey et al., 2019; Lewis et al., 2006). Because they accumulate significantly less wealth and have lower marriage rates than their White female peers (Addo & Lichter, 2013; Hogan & Perrucci, 2007), Black women are disproportionately exposed to financial strain and other stressful life events (e.g., problems with children, at work, caregiving for aging parents) in late-life. These stressors, in turn, increase risk for physical health problems such as cardiovascular disease (Felix et al., 2019) and functional limitations (Kasper et al., 2008). Beyond interpersonal challenges, the neighborhood context is an ecological domain in which stress occurs. Neighborhood stress is important to consider, as older adults spend a significant proportion of time in their immediate residential area (Cornwell & Cagney, 2014). Compared to their White counterparts, Black Americans are more likely to live in highly segregated neighborhoods characterized by overrepresentation of unhealthy food options, little opportunity to engage in physical activity, and concerns about safety (Cornwell & Cagney, 2014; D. R. Williams & Collins, 2001). Furthermore, the presence of neighborhood crime is associated with increased body mass index and coronary heart disease risk for women (Mobley et al., 2006).

Though the association between stress and physical health has been confirmed among older Black women, stressors tend to be studied in isolation. Thus, the relative health impact of stressors vis-à-vis one another remains unknown. Answering such an inquiry can aid in determining points of prevention and intervention with regards to reducing stressors that most strongly influence older Black women’s physical health.

Ethnicity, Stress Exposure, and Physical Health

In research on older Black women, ethnic distinctions in the association between stress exposure and health is rarely a topic of investigation. Research on ethnic nuances in stress exposure among older Black Americans indicates points of convergence and divergence. For instance, some research finds that older African Americans and Afro-Caribbeans experience similar exposure to perceived discrimination (G. Marshall & Rue, 2012), but other research reports higher discrimination levels among African Americans (Marshall-Fabien & Miller, 2016). Older African Americans also report higher levels of material hardship than their Afro-Caribbean counterparts (Marshall-Fabien & Miller, 2016).

Prior research using National Survey of American Life data has shown that the social determinants of health for African American and Afro-Caribbean women differ (Barrington et al., 2021), and that the association between stress and health is nuanced by ethnicity among Black Americans (Lincoln, 2019; Marshall-Fabien & Miller, 2016). For example, Lincoln (2019) found that stressful life events (e.g., marriage problems, having been a victim of a crime, problems with children) were associated with obesity for Afro-Caribbean but not African American women. Though focused on a broader age range, Lincoln’s (2019) study provides suggestive evidence that the association between stress and physical health could operate in distinct ways for older African American and Afro-Caribbean women. Marshall-Fabien and Miller (2016) found a positive association between material hardship and depressive symptoms for older African Americans, but not Afro-Caribbeans, suggesting that material hardship may be a stronger predictor of health for African Americans. Nevertheless, this proposition has yet to be empirically tested among US older Black women and for physical health outcomes.

The Current Study

To address the dearth of research on ethnically diverse older Black women, this study is guided by four research questions. First, do older Afro-Caribbean and African American women experience different rates of physical health problems and levels of stress exposure? Though past research suggests a health advantage for Afro-Caribbean women (Erving, 2011), ethnic variation in stress exposure is rarely investigated within the Black population. Second, does differential exposure to stress explain ethnic differences in physical health between the groups? While stress theory proposes that stress exposure may explain status distinctions in health (Turner, 2013), this SPM tenet has yet to be confirmed among older Black women. Third, how do multiple stress exposures influence Black women’s physical health? Individual stressors (e.g., everyday discrimination) tend to be the focus of empirical research, yet an assessment of multiple stressors and their accumulative effects more holistically capture the extent to which stress contributes to physical health in older adulthood (Brown & Hargrove, 2018). Fourth, does ethnicity condition the association between stress exposure and physical health? Given differences in their lived experience in the US, it is plausible that the health consequences of stress could also differ for older Afro-Caribbean and African American women.

More generally, this study provides much-needed nuance to the broader research landscape of minority aging by examining heterogeneous health profiles of older Black women, a population often studied monolithically. In addition, this study assesses the heterogeneous influence of four distinct stressors on physical health; such stressors tend to be studied in isolation, precluding the research community from identifying the relative influence of specific stress exposures on health. Such an endeavor has practical implications in that such findings may provide guidance on prioritizing prevention and intervention strategies for improving the health of older Black women. Theoretically, this study breaks new ground in the SPM research tradition which has tended to rely on mostly White samples, preventing in-depth understanding of the extent to which SPM tenets are relevant to the stress experiences of and concomitant health implications for older Black women.

Data and Methods

Data were drawn from the National Survey of American Life: Coping with Stress in the 21st Century (NSAL). NSAL included a nationally representative sample of African Americans, Afro-Caribbeans, and Non-Hispanic Whites. Data for the NSAL were collected by the Program for Research on Black Americans at the University of Michigan’s Institute for Social Research (Jackson et al., 2004). The survey population included African American, Caribbean Black, and non-Hispanic White non-institutionalized adults aged 18 years and older who resided in households in the coterminous United States. In the NSAL, African Americans were persons who self-identified as Black but did not report Caribbean ancestry. Caribbean Blacks were persons who self-identified as Black and answered affirmatively to any of the following inclusion criteria: (1) of West Indian or Caribbean descent, (2) born within a Caribbean area country, or (3) had parents or grandparents who were born in a Caribbean area country. Individuals living on military bases and non-English speakers were excluded from the study. To increase comfort level, respondents were matched with interviewers of the same racial background. Interviews lasted an average of 2 hours 20 minutes. The final overall response rate was 71.5%. A total of 6082 face-to-face interviews were completed and consisted of 3570 African Americans, 1621 Afro-Caribbeans, and 891 Non-Hispanic Whites. For this analysis, the sample was restricted to individuals who identified as female, age 50 years and older, and either African American or Afro-Caribbean. Based on these exclusion criteria, the sample was composed of 709 African American women and 249 Afro-Caribbean women (total sample = 958). After conducting listwise deletion, the restricted sample included 644 African Americans and 223 Afro-Caribbeans (total sample = 867), thus retaining 91% of the sample.

Dependent Measures

Three health indicators were included as dependent measures. First, multiple chronic conditions (MCC) refer to the cooccurrence of two or more physical conditions (American Geriatrics Society Expert Panel on the Care of Older Adults with Multimorbidity, 2012; US Department of Health and Human Services [DHHS], 2010). The MCC measure was constructed based on respondents’ self-report of whether they had been professionally diagnosed with 15 physical health conditions: arthritis/rheumatism, ulcer, cancer, hypertension, diabetes, liver problems, kidney problem, stroke, asthma, chronic lung disease, blood circulation problems, sickle cell disease, heart trouble/heart attack, glaucoma, and osteoporosis. These conditions had high prevalence in the US population (e.g., hypertension), were associated with aging (e.g., osteoporosis), and/or were associated with disability and mortality risk (e.g., cardiovascular disease). Second, functional limitations were assessed by items that queried respondents about the amount of difficulty they experienced performing the following tasks: standing for long periods such as 30 minutes, moving around inside their homes, walking a long distance such as a kilometer or half a mile, washing their bodies, getting dressed, and staying alone for a few days. Responses to the six items were combined, and the comparison was between those who reported one or more functional limitations and those who reported no functional limitations (Erving, 2011; Gorman & Read, 2006). Third, self-rated health asked respondents to evaluate their health in general on a five-point scale (1 = poor, 2 = fair, 3 = good, 4 = very good, 5 = excellent) (Brown et al., 2016; Erving, 2011).

Independent Measures

Ethnicity distinguished between African American (=0) and Afro-Caribbean (=1) women.² Five stressors were included. Major discrimination was assessed using a count of nine lifetime discriminatory events experienced across various domains (e.g., neighborhood, education, and workplace). Respondents were asked whether they experienced each event (e.g., “stopped, searched, questioned, physically threatened or abused by the police” and “unfairly discouraged by a teacher or advisor from continuing your education”) (Kessler et al., 1999; D. R. Williams, Yu, Jackson, & Anderson, 1997). Items were summed and higher values were indicative of greater discrimination exposure. The everyday discrimination measure summed 10 items that captured unfair treatment in day-to-day life (e.g., “being treated with less respect than others,” “people act as if they are afraid of you,” and “you are called names or insulted”) (Kessler et al., 1999; Mouzon et al., 2020). Response categories included never (0), less than once a year (1), a few times a year (2), a few times a month (3), at least once a week (4), and almost every day (5). The scale is the average reported discrimination across all items, with a range of 0–5 (Cronbach’s alpha = .89), with higher values reflecting more frequently experiencing everyday discrimination (Essed, 1991; Keith et al., 2010; D. R. Williams et al., 1997). With regards to both discrimination measures, as opposed to solely focusing on discrimination experiences pertaining to race or ancestry, these measures did not constrain the attribution of unfair treatment, as the intersectional identities of older Black women would have been constrained to single status attribution (e.g., solely race, gender, or age).

Stressful life events were assessed using 10 questions that queried respondents whether they had experienced the following in the past month: health problems, money problems, job problems, problems with children, family or marriage problems, been (or family been) the victim of a crime, problems with police, problems with love life, and treated badly because of race, and difficulty with gambling. Response options were yes (=1) or no (=0), and affirmative answers were used to construct a count of stressful life events. Higher values reflected a greater number of stressful life events (Assari & Lankarani, 2015; Glass et al., 1997; Lincoln, 2019).

Material hardship was a seven-item scale which asked respondents whether they experienced the following in the past year: (1) did not meet basic expenses; (2) did not pay full rent or mortgage; (3) were evicted for non-payment; (4) did not pay full gas, electric, or oil; (5) had gas or oil disconnected; (6) had telephone disconnected; and (7) could not afford leisure activities. Responses were either no (0) or yes (1). All responses were summed to create a total composite score, with higher scores reflecting greater material hardship (Hughes et al., 2014; G. L. Marshall, Thorpe, & Szanton, 2017). Neighborhood crime was assessed by the question: “how often are there problems with muggings, burglaries, assaults, or anything else like that in your neighborhood?” Response options included: Never (1), hardly ever (2), not too often (3), fairly often (4), and very often (5) (Erving & Cobb, 2021; Simning et al., 2012).

To assess the health effects of stress accumulation, a count measure was created using the five stress exposures. To establish a count of stress exposures, dichotomous cutoffs for each stress exposure were established. The cutoff points for major discrimination, everyday discrimination, stressful life events, and material hardship were assessed by using the highest-risk quartile (Brown & Hargrove, 2018; Sternthal et al., 2011). The neighborhood safety stress measure was counted as “1 ” if individuals reported “fairly often” or “very often” crime frequency in their neighborhood. The number of stressors was top-coded at four, as only seven respondents reported five stressors. The resulting stress accumulation measure included the following categories: no (reference), one, two, three, and four or more stress exposures.

Controls

Age (measured in years; age range: 50–93 years), relationship status, parental status, socioeconomic status (i.e., educational attainment, annual household income, and employment status), and region were included as controls. Relationship status distinguished among those who were married/cohabiting (reference), never married, and divorced/separated/widowed. Parental status distinguished between women who reported having biological children (=1) and those who did not (=0). Educational attainment categories included less than high school (reference), high school/some college, and college educated. Annual household income was measured in $10,000 increments ($0 to $200,000 or more). Employment status distinguished among those who were employed (reference category), unemployed, and not in the labor force. Region distinguished among residents living in the South (reference), Midwest, Northeast, and West.

Models also adjusted for psychological health and health-related behavior. Serious psychological distress was assessed using the 6-item version of the Kessler scale (K6) designed to assess non-specific psychological distress within the past 30 days (Kessler et al., 2002, 2003). Scale items included feeling “nervous,” “restless or fidgety,” “hopeless,” “that everything was an effort,” “blue,” and “worthless”. Each item was measured on a five-point Likert scale, with response categories including “none of the time” (0), “a little of the time” (1), “some of the time” (2), “most of the time (3), and “all of the time” (4). Higher scores reflected higher levels of serious psychological distress. Health behaviors included physical activity (Lincoln, 2019; Ray, 2014) and smoking status (Gebreab et al., 2012). With regards to physical activity, respondents were queried regarding the frequency with which they do work in the garden, engage in active sports or exercise, and take walks. Response categories included “never” (0), “rarely” (1), “sometimes” (2), and “often” (3). The three items were indexed to create a summative measure of physical activity, with a range of 0–9. Smoking status distinguished between current smokers (1) and non-smokers (0). Health insurance status included the following categories: No insurances (reference), government-sponsored insurance, employer-based (i.e., respondent’s own employer or a family member’s employer) insurance, and multiple sources of insurance or purchased insurance directly.

Analytic Strategy

The analysis proceeds by reporting means, proportions, and—where appropriate—standard deviations for the study measures for African American and Afro-Caribbean women. Adjusted Wald tests were used to assess differences between the two groups. In the multivariate analysis, binary logistic regression analysis was used to analyze MCC and functional limitations while ordered logistic regression analysis was used for self-rated health. The regression models begin with ethnicity and adjustments for control measures. To assess the association between stress exposure and health, the second model added the five stressors. The third model examines stress accumulation by including the number of stressors. Last, to assess whether ethnicity moderated the association between stress exposure and physical health, statistical interactions between ethnicity and each individual stress exposure measure and stress accumulation were run. Statistically significant interactions are reported in the regression analysis tables as well as presented as figures. Because of the complex sampling strategy used to collect the NSAL data, survey procedures (i.e., sampling weights) were applied to correct for unequal probabilities of selection, non-response, and design effects in the sample. All analyses were conducted using STATA 16.1.

Results

Descriptive Statistics

Table 1 includes means, proportions, and (where appropriate) standard deviations for the study measures by ethnicity. With regards to health, African Americans experienced higher rates of MCC and functional limitations as well as lower average self-rated health relative to Afro-Caribbeans. Nearly two-thirds of African American women and 46% of Afro-Caribbean women had MCC. One-third of African American women reported functional limitations; on the other hand, 12% of Afro-Caribbean women reported functional limitations. Self-rated health means were closest to the “good” category, at 3.02 (SD = 1.09) for African Americans and 3.42 (SD = 1.08) for Afro-Caribbeans. With regards to stress exposure, African Americans reported more stressful life events (mean = 1.41, SD = 1.27) compared to Afro-Caribbeans (mean = 1.09, SD = 1.16). African Americans were also more than four times likely (9%) to report four or more stressors compared to Afro-Caribbeans (2%).

Table 1.

Weighted Descriptive Statistics by Ethnicity.

	African American Women	Afro-Caribbean Women
	Mean (SD)	Mean (SD)	Sig. Test	Range
Physical health measures
Multiple chronic conditions (MCC)	.65	.46	***	0.1
Functional limitations	.33	.12	***	0.1
Self-rated health	3.02 (1.09)	3.42 (1.08)	***	1–5
Stress exposure
Major discrimination	1.05 (1.42)	1.04 (1.38)		0–8
Everyday discrimination	.85 (.78)	.91 (.81)		0–5
Stressful life events	1.41 (1.27)	1.09 (1.16)	*	0–7
Material hardship	.64 (1.28)	.53 (1.15)		0–7
Neighborhood crime	2.38 (1.16)	2.22 (1.05)		1–5
Stress accumulation
No stressor (reference)	.34	.31		0.1
One stressor	.25	.32		0.1
Two stressors	.19	.19		0.1
Three stressors	.12	.15		0.1
Four or five stressors	.09	.02	***	0.1
Controls
Age (years)	62.75 (9.76)	62.89 (10.08)		50–93
Relationship status
Married/cohabitating (reference)	.32	.37		0.1
Never married	.07	.09		0.1
Divorced/separated/widowed	.61	.54		0.1
Parental status	.92	.94		0.1
Education
Less than high school (reference)	.35	.41		0.1
High school/some college	.50	.44		0.1
College graduate	.15	.15		0.1
Household income (in $10,000s)	2.99 (2.53)	3.92 (3.29)		0–20+
Employment status
Employed (reference)	.42	.60	**	0.1
Unemployed	.06	.02	*	0.1
Not in labor force	.52	.38	*	0.1
Region
South (reference)	.56	.26	***	0.1
Midwest	.18	.07	**	0.1
Northeast	.17	.61	***	0.1
West	.09	.06		0.1
Serious psychological distress	2.95 (3.88)	2.45 (3.32)		0–19
Physical activity	4.46 (2.59)	5.36 (2.22)	**	0–9
Current smoker	.20	.08	*	0.1
Health insurance status
No insurance (reference)	.13	.09		0.1
Government insurance	.31	.15	***	0.1
Employer-based insurance	.29	.45		0.1
Multiple insurance sources	.28	.30		0.1

Source: National Survey of American Life, 2001–2003.

^*p < .05

^**p < .01

^***p < .001.

Sample size for African American women is 606 for functional limitations and 644 for all other measures. Sample size for Afro-Caribbean women is 209 for functional limitations, 222 for self-rated health, and 223 for all other measures.

For the controls, several significant differences emerged. First, Afro-Caribbeans had higher employment rates; in contrast, African Americans had higher unemployment rates and were more likely to not be in the labor force. African Americans were more concentrated in the South and Midwest while Afro-Caribbeans had higher representation in the Northeast. Afro-Caribbeans had higher physical activity and lower smoking rates. With regards to health insurance, African American women had higher rates of government-sponsored insurance (31%) compared to Afro-Caribbean women (15%).

Regression Analysis

The goal of the regression analysis was to assess the following questions: Does stress exposure explain health differences between African American and Afro-Caribbean women? What are the effects of stress exposure and stress accumulation on physical health? And, last, does ethnicity condition the stress-physical health association? The results for multiple chronic conditions (MCC) are presented and discussed first, followed by functional limitations and self-rated health.

Multiple Chronic Conditions

In Table 2, odds ratios from the weighted binary logistic regression analysis are reported for MCC. In Model 1, after adjusting for socioeconomic and health-related behaviors, Afro-Caribbean women had marginally lower odds of MCC than African American women (OR = .61, p = .07). Model 2 introduced the individual stressors. Major discrimination (OR = 1.29, p < .01) and stressful life events (OR = 1.46, p < .001) were associated with higher odds of MCC. In Model 3, stress accumulation heightened MCC risk, as experiencing one (OR = 1.92, p < .05), two (OR = 2.17, p < .05), three (OR = 3.86, p < .001), and four or more stressors (OR = 3.38, p < .01) were progressively associated with higher odds of MCC compared to no stress exposure.

Table 2.

Weighted Odds Ratios from Binary Logistic Regression Models for Multiple Chronic Conditions (MCC) (N = 867).

	Model 1	Model 2	Model 3	Model 4
Afro-Caribbean	.61 (.36, 1.04)	.66 (.35, 1.23)	.62 (.36, 1.04)	.30**(.13, .71)
Stress exposure
Major discrimination		1.29**(1.11, 1.50)		1.28***(1.11, 1.47)
Everyday discrimination		1.09 (.76, 1.57)		1.02 (.72, 1.45)
Stressful life events		1.46***(1.26, 1.69)		1.49***(1.29, 1.71)
Material hardship		1.03 (.84, 1.26)		1.03 (.84, 1.25)
Crime in neighborhood		.99 (.80, 1.22)		.99 (.80, 1.22)
Stress accumulation
No stressor (reference)			1.00
One stressor			1.92*(1.13, 3.27)
Two stressors			2.17*(1.03, 4.60)
Three stressors			3.86***(2.13, 7.02)
Four or more stressors			3.38**(1.55, 7.37)
Significant statistical interaction
Afro-Caribbean × everyday discrimination				2.51*(1.16, 5.45)

Source: National Survey of American Life, 2001–2003.

95% confidence intervals in parentheses. Multiple Chronic Conditions (MCC) is defined as reporting two or more physical conditions. All models control for age, relationship status, parental status, educational attainment, household income, employment status, region, serious psychological distress, physical activity, smoking status, and health insurance.

^*p < 0.05

^**p < 0.01

^***p < 0.001.

To assess whether the association between stress and MCC differed for African Americans compared to Afro-Caribbeans, statistical interactions between ethnicity and each stress exposure were modeled. As shown in Model 4, one interaction was significant: everyday discrimination. The interaction between ethnicity and everyday discrimination is modeled in graphical form in Figure 1. Predicted probabilities of MCC are reported. Among Afro-Caribbean women, as everyday discrimination increased, the predicted probability of MCC increased. For instance, for those who reported no everyday discrimination (= 0), MCC was approximately 0.38, while this predicted probability approached .61 for those who reported “less than once a year” (=1) average everyday discrimination. For those who reported the most frequent discrimination (i.e., “almost every day”), the predicted probability of MCC was .98. For African American women, however, everyday discrimination did not increase the predicted probability of MCC.

Figure 1.

MCC Predicted Probabilities by Everyday Discrimination. Source: National Survey of American Life, 2001–2003. Predicted Probabilities are based on the interaction between ethnicity and everyday discrimination shown in Table 2, Model 4. All other covariates are set at their means.

Functional Limitations

In Table 3, odds ratios from a binary logistic regression analysis are reported for functional limitations. In Model 1, which included ethnicity and the controls, Afro-Caribbean women experienced 70% lower odds of functional limitations compared to African American women (OR = .30, p < .001). Model 2 introduced stress exposures and the health advantage for Afro-Caribbean women persisted (OR = .34, p < .01). Stressful life events (OR = 1.94, p < .001) were associated with higher odds of functional limitations. In Model 3, stress accumulation increased risk for functional limitations, as experiencing two (OR = 4.05, p < .001), three (OR = 5.22, p < .01), and four or more stressors (OR = 6.59, p < .01) were progressively associated with higher odds of functional limitations compared to no stress exposure.

Table 3.

Weighted Odds Ratios from Binary Logistic Regression Model for Functional Limitations (N = 815).

	Model 1	Model 2	Model 3	Model 4
Afro-Caribbean	.30***(.16, .59)	.34**(.17, .70)	.34**(.16, .71)	1.40 (.36, 5.39)
Stress exposure
Major discrimination		1.13 (.90, 1.41)
Everyday discrimination		1.10 (.71, 1.70)
Stressful life events		1.94***(1.63, 2.31)
Material hardship		1.11 (.88, 1.42)
Crime in neighborhood		1.00 (.81, 1.24)
Stress accumulation
No stressor (reference)			1.00	1.00
One stressor			1.49 (.85, 2.60)	1.57 (.89, 2.78)
Two stressors			4.05***(1.96, 8.37)	4.37***(2.07, 9.24)
Three stressors			5.22**(1.98, 13.74)	5.53**(2.04, 14.95)
Four or more stressors			6.59**(2.03, 21.39)	6.79**(2.04, 22.57)
Significant statistical interactions
Afro-Caribbean × one stressor				.16 (.02, 1.13)
Afro-Caribbean × two stressors				.05**(.01, .47)
Afro-Caribbean × three stressors				.16*(.03, .85)
Afro-Caribbean × four or more stressors				.87 (.05, 14.24)

Source: National Survey of American Life, 2001–2003; 95% confidence intervals in parentheses;

^*p < 0.05

^**p < 0.01

^***p < 0.001.

All models control for age, relationship status, parental status, educational attainment, household income, employment status, region, serious psychological distress, physical activity, smoking status, and health insurance.

To assess whether the association between stress and functional limitations differed for African Americans compared to Afro-Caribbeans, statistical interactions between ethnicity and each stress exposure were modeled. Though no individual stress exposure interactions were significant, the association between stress accumulation and functional limitations differed for Afro-Caribbean and African American women as shown in Model 4: specifically, the effect of two and three stressors differed for the two groups. The interaction between stress accumulation and ethnicity is presented in graphical form in Figure 2 which reports predicted probabilities. For African Americans, the predicted probability of functional limitations increased in incremental fashion as stress accumulated. Among Afro-Caribbean women, however, the effects of two and three stressors on predicted probability of functional limitations were less pronounced compared to African American women. Nonetheless, women experiencing the highest stress exposure (i.e., stress accumulation = 4 or more) had the highest predicted probability of functional limitations for both groups.

Figure 2.

Functional Limitations Predicted Probabilities by Stress accumulation. Source: National Survey of American Life, 2001–2003. Predicted Probabilities are based on the interaction between ethnicity and stress accumulation shown in Table 3, Model 4. All other covariates are set at their means.

Self-Rated Health

In Table 4, odds ratios from an ordered logistic regression analysis are reported for self-rated health. In Model 1, which included ethnicity and the controls, Afro-Caribbean and African American women did not significantly differ from one another. Model 2 introduced the individual stressors. Stressful life events (OR = .73, p < .001) were associated with lower self-rated health. In Model 3, stress accumulation was associated with lower self-rated health, as experiencing three (OR = .51, p < .05) or four or more stressors (OR = .50, p = .07) were associated with lower self-rated health compared to no stress exposure.

Table 4.

Weighted Odds Ratios From Ordered Logistic Regression Model for Self-Rated Health (N = 866).

	Model 1	Model 2	Model 3
Afro-Caribbean	1.35 (.86, 2.11)	1.16 (.72, 1.87)	1.328 (.83, 2.11)
Stress exposure
Major discrimination		.98 (.85, 1.12)
Everyday discrimination		1.15 (.89, 1.50)
Stressful life events		.73***(.63,.84)
Material hardship		.97 (.84, 1.13)
Crime in neighborhood		.86*(.73, 1.00)
Stress accumulation		.86*(.73, 1.00)
No stressor (reference)			1.00
One stressor			.73 (.45, 1.20)
Two stressors			.74 (.47, 1.18)
Three stressors			.51*(.29, .88)
Four or more stressors			.50 (.23, 1.07)

Source: National Survey of American Life, 2001–2003. 95% confidence intervals in parentheses.

^*p < 0.05

^**p < 0.01

^***p < 0.001.

All models control for age, relationship status, parental status, educational attainment, household income, employment status, region, serious psychological distress, physical activity, smoking status, and health insurance.

Discussion

Using data from the National Survey of American Life, this study had four aims. First, I assessed the extent to which Afro-Caribbean and African American women experienced different physical health profiles and stress exposures in older adulthood. Second, I investigated whether differential exposure to stress explained ethnic differences in physical health. Third, I examined how multiple stressors and stress accumulation influenced physical health. And, last, I ascertained whether the association between specific stressors and physical health differed for older African American and Afro-Caribbean women. Study results revealed several interesting patterns.

First, Afro-Caribbean women experienced a physical health advantage vis-à-vis African American women with regards to MCC, functional limitations, and self-rated health. After adjustments for sociodemographic controls, however, the health advantage of Afro-Caribbean women only persisted for functional limitations. Given that 85% of the sample of older Afro-Caribbean women was foreign-born, this finding is consistent with the immigrant health literature which reports a health advantage for foreign-born Black immigrants relative to their US-born Black counterparts (Hamilton, 2019; Hamilton & Hummer, 2011). With regards to stress exposure, African American and Afro-Caribbean women experienced similar levels of most stressors, with one exception: African American women reported more stressful life events. With regards to stress accumulation, the groups differed at the upper end of the stress accumulation distribution: while only 2% of Afro-Caribbean women reported four to five stressors, 9% of African American women did. This finding conveys that, on average, older African American women experience greater levels of stress exposure than their Afro-Caribbean female counterparts. Alternatively, older Afro-Caribbean women may experience stressors not captured here. Acculturative stress is common among US Caribbean immigrants (Livingston et al., 2007), and could be especially salient for foreign-born Afro-Caribbean women as they adapt to US culture during their transition to older adulthood. More research is needed to understand acculturative stress for US Caribbean immigrant populations in general (Fanfan & Stacciarini, 2020), and older Afro-Caribbean women in particular. It is equally critical to assess heterogeneity in immigrants’ orientation and exposure to the US; for instance, age of immigration, length of stay in the US, and the timing of immigration (e.g., immigrants arriving during a restrictive immigration era may have an especially stressful transition to the US) have distinct influences on health (Alegria et al., 2007; Jackson et al., 2007).

Second, stress exposure did not explain ethnic differences in functional limitations. As opposed to focusing on micro-level processes such as perceptions of stress, differences in health between Afro-Caribbean and African American older women could be attributable to macro-level structural processes. For instance, highly selective immigration policies facilitate US entry for the healthiest individuals and their families who migrate from Caribbean area nations (Model, 2008). Another hypothesis that has received little empirical attention for Caribbean immigrants is that physically ill immigrants are more likely to return to their country of origin than those who remain in good health (i.e., the salmon bias effect) (Riosmena et al., 2013). In sum, a myriad of immigration processes could collectively contribute to the health advantage of this mostly foreign-born group of Afro-Caribbean women.

Third, when examining the effects of specific stressors on health, some stressors were never prominent contenders (e.g., material hardship) while others were only associated with one health indicator. For instance, major discrimination was only associated with higher risk for MCC. Major discrimination involves significant barriers preventing Black women from accessing employment, education, and economic capital (i.e., being denied a bank loan), thus constricting opportunities for upward economic mobility. Furthermore, problematic interactions with police could induce prolonged stress activation that, in turn, negatively impacts physiological systems in the body (Sewell & Jefferson, 2016). Unfairness at the neighborhood level is critical from a physical health standpoint, as Black women may be actively excluded from accessing high-resource neighborhoods and even when living in well-resourced neighborhoods, neighbors could make life difficult (Ray, 2014; Sewell, 2016). The effect of major discrimination on MCC risk reflects a cumulative impact of major instances of unfair treatment, with each additional unfair event inducing greater risk for diagnosable health problems.

Though the association between major discrimination and health was outcome-dependent, stressful life events and cumulative stress were associated with higher risk for MCC risk and functional limitations as well as lower self-rated health. These robust effects across health measures suggest that reducing stressful life events could be insightful foci for health promotion and prevention programs. Stress accumulation’s robust health effects forcefully convey that the volume and proliferation of stress are important to reduce above and beyond focusing on a single stressor. Accordingly, multifaceted prevention and stress reduction efforts will be more effective than targeting a single stress domain.

Fourth, everyday discrimination was associated with MCC risk among Afro-Caribbean women, but not African American women. Perhaps discrimination is a jarring experience for Afro-Caribbean women, as many Black immigrants arrive in the US with an optimistic perspective characterized by little anticipation of racialized unfair treatment (Vickerman, 1999; Waters, 1999). Afro-Caribbean immigrant women, having hailed from majority-Black countries (e.g., Jamaica, Haiti), may have been less likely to encounter racial discrimination during their formative years in their countries of origin (Read & Emerson, 2005; Read et al., 2005). Upon encountering an unwelcoming racialized context in the US where they are perceived as African American, Afro-Caribbean immigrants perhaps become less optimistic over time. Relatedly, Afro-Caribbeans who perceive discrimination are more likely to identify as “Black” than those who do not perceive discrimination; therefore, awareness of US racism seems to catalyze a shift in the salience of race as a part of their US identity construction (Jones & Erving, 2015). Hence, discrimination could be devastating for Afro-Caribbean immigrants which, in turn, induces problematic health behaviors and maladaptive coping that could heighten risk for chronic conditions. On the other hand, for African American women who have been socialized in the US and witnessed US-style racial inequality over their entire life course, daily assaults like those captured by the everyday discrimination scale (e.g., hyper-surveillance in stores, being treated with less courtesy) may be less influential for their physical health in comparison to other ongoing stressors in their lives. Older African American women who came of age during the pre-Civil Rights era may have also developed psychological resilience to everyday discrimination because of their repeated exposure to such treatment earlier in life.

Last, the discrepant findings for everyday discrimination versus accumulative stress exposure in terms of their influence on different health outcomes and across ethnicity suggest that distinct mechanisms link stressors (individually and cumulatively) to specific health outcomes. On one hand, study results revealed that the association between everyday discrimination and MCC was strong for Afro-Caribbean women but non-existent for African American women. As described earlier, African American women may have developed some psychological and physiological resilience against everyday discrimination because of its ubiquitous nature and their racial socialization experiences that primed them to anticipate discrimination early on in life. On the other hand, stress accumulation was associated with higher risk for functional limitations, in incremental fashion, for African American women but not for Afro-Caribbean women. In this regard, findings for African American women are consistent with the proposition that accumulative stress will be more harmful to health than exposure to a single stressor. For Afro-Caribbean women, experiencing four or more stressors was the condition under which they faced heightened risk for functional limitations. Afro-Caribbean women experiencing fewer stressors may be able to draw upon psychosocial resources (e.g., social support from relatives) to address a more contained volume of stressors in their lives. The stress mechanisms underlying health may also be health outcome-dependent for Afro-Caribbean women. These somewhat counterintuitive findings suggest further investigation of stress processes and health are needed to better understand these patterns.

Despite its strengths, this study has some limitations. First, though this study captured multiple stress exposures, other stressors such as negative interactions with family members and caregiving strain were not accounted for here. Caregiving strain is especially relevant, as provision of care for aging parents, an ailing spouse, or younger grandchildren disproportionately falls on the shoulders of Black women (G. F. Carr, Hayslip, & Gray, 2012; Thorne, 2020). Second, health measures were self-reported, which could present challenges with validity and potentially underestimate illness prevalence (Idler & Cartwright, 2018); nonetheless, this study’s use of multiple health indicators, and consistency in ethnic patterns across the health outcomes, aids in addressing potential biases that a single health status measure could introduce. Last, much of the Afro-Caribbean sample was foreign-born, making it methodologically untenable to ascertain nativity distinctions within this population. Future research should investigate differences in stress exposure between older US-born and foreign-born Afro-Caribbean women.

Despite these limitations, study findings revealed complex relationships among ethnicity, stress exposure, and physical health. Ethnic distinctions in stress and health patterns reflect the reality that older Black women and their experiences are far from homogenous, and future research should be attuned to within-group variation. Aging research more broadly should examine status heterogeneity among Black Americans, as there is much to learn about the diverse experiences of the US aging Black population.