Economic Strain Deteriorates While Education Fails to Protect Black Older Adults Against Depressive Symptoms, Pain, Self-rated Health, Chronic Disease, and Sick Days

Economic Strain Deteriorates While Education Fails to Protect Black Older Adults Against Depressive Symptoms, Pain, Self-rated Health, Chronic Disease, and Sick Days Shervin Assari1*, Sharon Cobb2, Mohammed Saqib3, Mohsen Bazargan1,4 1Departments of Family Medicine, Charles R Drew University of Medicine and Science 2School of Nursing, Charles R Drew University of Medicine and Science, Los Angeles, CA, United States 3University of Michigan School of Public Health, Ann Arbor, MI, United States 4Departments of Family Medicine, University of California, Los Angeles (UCLA), Los Angeles, CA, United States

Various SES indicators may have joint and unique health implications for racial minorities. As suggested by the Minorities' Diminished Returns (MDRs) theory 45 , education shows smaller health effects for marginalized people such as Blacks [20][21][22][23][24]46 . That is highly educated Blacks remain at high risk of poor health 20,24,25,[46][47][48][49] . Such patterns are observed in Black youth, adults, and older adults. These studies cumulatively suggest that objective SES indicators such as education may have "less than expected" effects on the well-being and health of Black older adults, while economic strain may have an exacerbated effect for the very same population 25, 51 . However, we are only aware of one study simultaneously comparing the effects of educational attainment and economic strain on the well-being and health of Black older adults.
In such conditions, while education shows weaker effects 45 , economic strain may be particularly impactful on populations that have low access to buffers such as social support, which is common in old age 33 . As a result, research has documented major adverse health effects of economic strain for Blacks 40,52,53 and older adults [54][55][56] . For example, in one study in a Black community, economic strain at baseline predicted future heart disease 42 . These adverse effects of economic strain in the elderly may be compounded by unemployment, social isolation, lack of access to transportation, and various chronic diseases.
In a recent study 57 , economic strain was positively associated with chronic diseases, chronic pain, self-rated health, depressive symptoms, and sick days. Although this study suggested that economic strain seems to be linked to self-rated health, sick days, chronic pain, chronic diseases, and depressive symptoms, this study did not compare the role of economic strain and educational attainment. The authors of that study advocated for the evaluation of social determinants of health in providing health care of Black older adults. However, the study did not control for several confounders. In fact, the results were based on zero-order or partial correlations and multivariable analysis that can investigate the joint effects of educational attainment and economic strain on mental and physical health outcomes while other confounders are controlled.

Aims
The current study investigated additive effects of educational attainment and economic strain on five health outcomes, namely depressive symptoms, pain intensity, chronic diseases, self-rated health, and sick days, for Black older adults in the underserved areas of South Los Angeles. The hypothesis was that educational attainment would have a small protective effect, while economic strain will show a considerable risk effect in our sample. We expected the very same pattern across health outcomes 58 for the combined effects of education and economic strain [59][60][61] . That is, we expect that in contexts with high economic constraints, economic strain will operate as a risk while education will fail to show considerable protective effects. This hypothesis is in line with the observation that racism limits the health gains that follow educational attainment 20,22,23,45,[47][48][49][62][63][64][65] .

Design
This cross-sectional household survey was conducted in South Los Angeles 2015-2018. The study was initially insurance (0/1). We considered being covered by Medicare, Medicaid, VA, non-federal health insurance, or private insurance as having health insurance. As only six individuals did not have health insurance, we did not include health insurance in our multivariable linear regression model.

Health Behaviors:
Participants reported if they were smoking cigarettes and whether they were drinking alcohol. These two variables were conceptualized as dichotomous variables.
Objective Socioeconomic Status (SES): Education attainment was the only measure of objective SES. Education was an interval variable (years of schooling) with a higher score indicating more educational attainment.
Economic Strain: Self-reported (perceived) economic strain was measured using three items that were in line with Pearlin's list of main chronic economic strains that are being experienced by low SES individuals 71-73 . These items assess whether self-reported money is sufficient to meet essential needs (food, clothes, mortgage, bills). The items also measured the frequency of not having enough money to buy enough food, clothing, and paying bills. Responses were on a 0 to 5 scale, with responses ranging from 'never' to 'always'(with 0 signifying "never" and five signifying "always"). A total score was calculated with a high score reflecting more economic strain. Reliability of the measure was excellent with a Cronbach alpha = .923.
Pain Severity: We measured the intensity of pain using four subscales of the McGill Pain Questionnaire-Short-Form 2 88 . During a face-to-face interview, respondents responded to 22 questions on the extent of experiencing various types of pain over the past seven days. Items were on an 11-point numeric rating scale from 0 to 10 (none to worst possible). Our total pain score was calculated based on mean score of all items [88][89][90][91][92][93][94] . A higher score is indicative of more pain intensity.

Chronic Diseases:
Participants were asked about 11 chronic diseases. Participants reported if a physician or a healthcare provider has ever told them that they have any of the following chronic diseases: asthma, chronic obstructive pulmonary disease, diabetes, hypertension, heart disease, lipid disorder, thyroid disorder, cancer, osteoarthritis, rheumatoid arthritis, and gastrointestinal disease. While self-reported chronic diseases is valid 95-97 , conducted to investigate medication-related challenges of Black older adults [66][67][68][69] . The interview collected data on demographic factors, SES characteristics (education and economic strain), health insurance, living arrangement, marital status, depressive symptoms, pain intensity, chronic diseases, self-rated health, and sick days. While the full evaluation took up to 4 hours, the structured faceto-face interview lasted about 2 hours. Interviews were performed by a physician or a nurse.

Participants and Sampling
The study recruited a convenience sample of Black older adults. Participants were recruited from predominantly Black housing units and senior centers that were located in South Los Angeles. Participants were Black or African American, non-institutionalized people, all age 65 years or older. All participants had hypertension.
The original study, which was conducted to examine medication-related challenges in Blacks with hypertension, included a total number of 740 Blacks aged 55 years and older. However, in the current study, we exclusively focused on participants that were 65+ years of age (n = 619). No institutionalized participant was enrolled from skilled nursing facilities. Considerable cognitive deficits or current enrollment in a clinical trial were exclusion criteria.

Institutional Review Board
The protocol of the current investigation was approved by the institutional review board at the Charles R. Drew University of Medicine and Science. All participants signed consent before enrollment in this study.

Study Measures
Study variables in this analysis included demographic factors (i.e., gender and age), objective SES (i.e., educational attainment), living arrangement, marital status, health insurance, economic strain, depressive symptoms, pain intensity, chronic diseases, self-rated health, and sick days.

Demographic Factors:
Gender and age were entered as demographic covariates. Age was an interval variable, operationalized as a continuous measure. Gender was a dichotomous variable.
Living Arrangement. Participants were asked if they were living alone or whether there were any other family members or a partner living with them. Living arrangement is a strong determinant of health among older adults 70 .
Marital Status: Participants' marital status was coded as non-married= 0 and married =1.

Health Insurance Status:
Participants were asked about their health insurance coverage. Health insurance status was a dichotomous variable that reflected health some bias should be expected due to underreporting of some conditions. Self-Rated Health: Participants' self-rated health was assessed by a single-item health measure with five options 64,[98][99][100][101][102][103][104][105] . The item reads, "In general, would you say your health is: "very good," "good," "fair," "bad," and "very bad." Responses were on an interval variable ranging from one to five, with a high score indicating poor selfrated health. Poor self-rated health predicts the risk of mortality 64,98-105 .
Sick Days: A single item which read "In the past 12 months, how frequently have you been sick?" was used to assess sick days in the past year. The responses ranged from never (coded as 1) to always (coded as 5). A higher score indicated a higher number of sick days. This measure is commonly applied in previous studies [106][107][108] .

Statistical Analysis
We used SPSS 23.0 for data analysis. To describe the sample, we reported frequency (n), relative frequency (%), mean, and standard deviation (SD). We used the Pearson correlation test for our bivariate analysis. We applied linear regression models with health outcomes as the dependent variable, educational attainment, and economic strain as the independent variables, and age, gender, marital status, and living arrangements were the covariates. Give that only six participants did not report having some types of health insurance; we did not include this variable in our multiple linear regression models. We reported unstandardized b, standard error (SE), 95% confidence intervals (95% CI), and level of p values. Table 1 shows a summary of all study variables in the pooled sample. Most participants were female (65.3%) and were living alone (59.9%). Only a small percentage were married (14.2%). Almost all participants reported having health insurance (99%). On average, participants were 74.0 years old.   was correlated with lower levels of economic strain (r = -.13, p < .05). Economic strain was positively correlated with worse status of all health outcomes (r ranging between .13 and .31, p < .05). Educational attainment was negatively correlated with the number of chronic diseases (r = -.09, p < .05) but not with any of the other four health outcomes (p > .05). All health problems were positively correlated with each other (r ranging between .33 and .44, p < .05). Table 3 shows the results of five linear regression models, one for each health outcome. In all these models, educational attainment and economic strain were the independent variables, a health outcome was the dependent variable, and age, gender, marital status, living arrangement, and health behaviors (smoking and drinking) were the covariates. These models suggest that economic strain, but not educational attainment, was associated with depressive symptoms, pain intensity, number of chronic diseases, self-rated health, and sick days. Similar patterns emerged regardless of health outcome.

Discussion
The current study explored how educational attainment and economic strain are associated with five health outcomes, namely depressive symptoms, pain intensity, number of chronic diseases, self-rated health, and frequency of sick days in Black older adults residing in economically disadvantaged areas of South Los Angeles. The results suggested that while educational attainment failed to protect them, high economic strain was associated with various poor health outcomes among Black older adults. Based on these results, we advocate for increased policies and programs that can stabilize incomes for Black populations across the life span.
As limitations should be in mind before the results are interpreted, we start our discussion with a review of the limitations of the current study. First, our cross-sectional design does not allow any causal inferences. Our results suggest association rather than causation. Second, there were some omitted SES indicators such as income and wealth. Given the sensitivity of economic and financial information and large amounts of missing data regarding income and wealth due to lack of disclosure 109 , and even low reliability of these variables in simple interviews 110 , we did not collect data on such sensitive information 111 . Future research should go beyond measuring education and economic strain and also measure income and wealth. Third, the sample was convenient. Thus, the results are not generalizable to all Black older adults in the US. Despite these limitations, the current study extends the limited existing knowledge on social determinants of physical and mental health of Black older adults in low-income urban settings.
In line with our hypotheses, educational attainment did not affect chronic diseases, self-rated health, sick days, depression, or pain. Similar smaller effects of SES were found on self-rated health 20,46 , depression 60,63 , and chronic diseases 46,47 . We, however, are not aware of any previous studies showing the same pattern for pain. Besides, most of the previous findings are in children 45,46,48,49 and adults 20,46,62 , and very few previous studies have shown the same patterns for older adults 50 . Thus, two innovations of these results are to expand the literature, which is mostly on youth and adults, to older adults, and also document the same patterns for pain intensity. Finally, while most previous findings are on national samples 20,47,62 , the current study suggests that similar patterns hold at a local level 46,65 . In the absence of a protective effect of education, the health and well-being of Black individuals seem to be strongly impacted by economic strain. Previous research has shown the role of economic strain as a significant threat to the health of the Black communities 40,52,53 , particularly for Black older adults [54][55][56] . While economic strain limits people's options for health-supporting behaviors and access to resources and services 32 , it also operates as a source of psychological stress 41 . Economic strain increases the risk of mortality 19,[112][113][114][115] . This increased risk may be due to an increased risk of a wide range of chronic diseases 42 including but not limited to heart disease 114 , diabetes 34 , cancer 113 , hypertension 42 , poor self-rated health 35 , or mental health problems such as anxiety 43,116 , depression 40,51 and suicide 39 . Economic strain also increases behavioral risk, such as poor diet 36 , smoking 37 , and alcohol use 38 .
Economic strain is shown to be one of the most influential social determinants of health 35,117-125 . This effect is shown for the general population 126 , people with chronic diseases 34 , and older adults 32 . Economic strain becomes a stronger threat for Black older adults who lack social support and other potential buffers 33 . Loneliness and lack of access to buffers seem to be a common element of aging in Black communities 33 .
In the US, Black older individuals with a more economic strain experience worse health outcomes across all domains. This finding is not new as economic strain is a reliable SES indicator impacting various aspects of health in multiple studies 40,112,127,128 . Also referred to as financial difficulty, stress, distress, or hardship 41,43,54,129 , economic strain reflects a lack of resources, particularly the absence of liquid expendable income. Individuals under economic strain are unable to access and use services that they need, maintain pro-health behaviors, or access resources that buffer stress and illness when they occur 32,55 .
Economic strain showing a harmful effect while education not showing a protective effect might be due to the economically constrained nature of low-income urban areas that are limited in resources. We found that economic strain shows consistent effects as identical patterns emerge across health outcomes. At the same time, highly educated and low educated Black older adults show the very same risk of depressive symptoms, pain intensity, chronic diseases, poor self-rated health, and sick days, which is in line with the MDRs of education in resourcelimited settings 130,131 .
Educational attainment's failure to protect Black older adults against health problems may be due to high segregation and concentration of poverty in urban Black communities. Lack of protective effects of educational attainment in marginalized people, also called MDRs, has been attributed to structural racism and social stratification 20,22,23,47,62,63 . This finding may also be attributed to five specific reasons: (a) poor quality of education in these communities, as discussed by Jenifer Manley 132-137 , (b) most Black older adults being out of the labor market (labor market carries some of the health gains of education attainment), (c) most Black older adults have some insurance, which covers their health needs regardless of their education, (d) absence of prestigious high paying jobs in urban settings, and (e) high education may increase psychological vulnerability of Blacks to racism, which reduces its protective effects. Some research has shown that high education may predict undesired mental health outcomes for Black individuals 26-30, 63,138,139 . Opposite to the pattern seen for Blacks, educational attainment always translates to better health for Whites 1-4 .
Economic strain is closely linked to perceived stress 33,112,140 . Thus, as shown by past research 25 , economic strain may have stronger effects on Black older adults than hard SES indicators such as income or education 25 . This might also be because perceived economic strain, depression, and perceived health have affective valence and component (i.e., tone/shade) 51,141-143 .
Although education is known as a strong determinant of health 1-4 , and as low education is believed to be a fundamental cause of racial health disparities 18 , education has differential implications across various social groups. While the most socially privileged groups gain the most amount of health from their education, this gain tends to zero for the least privileged groups 64,144 . That is, while education is commonly shown to improve health, crossracial variations exist in the health effects of education 1-4,6-14 . Overall, while educational attainment improves health through multiple mechanisms, including promoting lifestyle 16 for many populations including the elderly 17 , there is considerable evidence suggesting that educational attainment may lose some of its health effects for Black communities [20][21][22][23][24]46 .
In line with the MDRs theory, the inner strength of this manuscript is questioning the presumption that educational attainment should result in a protective (or positive) effect on a person's health, long-term. We argue that embedded in this presumption is the "bootstrap" philosophy that hard work (and investment on human capital) will pay off for all citizens, regardless of race or ethnicity, and translate into health as well as economic and financial security through educational pursuits that translate into job opportunities. Nevertheless, as shown by this study and the rest of the MDRs literature, these effects are not equitably distributed across all race/ethnic groups in our society. Socioeconomic factors do not similarly promote the health and wellbeing of Blacks and Whites 11,12 . The magnitude of the effects of each additional year of schooling on health outcomes is smaller for Blacks than Whites 11,12 . The threshold (i,e, stepwise) effect of educational credentials is also weaker for Blacks than Whites because the US labor market gives lower status jobs to Blacks than Whites 20,21 . In the US, it is race and SES, not race or SES, that shape health [145][146][147] . Aurora Jackson has conducted extensive work on economic strain, low paying jobs, unemployment, and how such conditions influence the daily lives of Blacks, particularly Black women 148,149 . So has work by Hamilton and Darity on the wealth gap and economic conditions of Black families. They have proposed reparation and baby bonds as potential solutions to the weatlh inequality of Blacks [150][151][152] . Lincoln has also conducted studies on the interplay between various sources of stress, such as economic strain in Blacks [153][154][155] . The work of Oliver and Shapiro also has helped us understand the wealth gap between Blacks and Whites 156,157 .
The work by Sherraden and Trina Shanks has helped us understand the life conditions associated with economic strain of low-income people, including Blacks 158-160 . Sherraden has tested innovations to improve the social and economic well-being of low-income families, including Blacks. He has proposed policies that may promote inclusion in asset building for marginalized people. One example of such asset-based policies and programs in the US is progressive child development accounts. In particular, they discuss the findings from the SEED for Oklahoma Kids study, which is a large-scale randomized statewide policy experiment that deposited $1,000 into state-owned Oklahoma 529 College Savings Plan (OK 529) accounts for 1,358 children. Although Sherraden's work is on children and not older adults, his work is still relevant to our finding as it addresses the critical aspect of the discussion and implication section, which is providing financial incentives. These lines of work collectively suggest safety nets and economic policies that may help Black families and communities accumulate wealth and buffer the effect of deep poverty [159][160][161][162][163] . Other studies have also shown positive health effects of reducing economic strain 164 .
As this study showed, economic strain and education have very different effects on the health of disadvantaged racial and ethnic groups such as urban Black older adults who struggle with poverty while affected by multiple chronic diseases. Similar patterns are shown by research for other sections of the Black communities [20][21][22][23][24]46 as well as Hispanics 62,65 , Native Americans 165 , and other marginalized groups 166 . The widespread and systemic nature of MDRs suggest MDRS are not innate or biological, but rather a function of racism, prejudice, and resource-scarce areas. Although MDRs of education are not specific to Blacks, but rather to any marginalized group, , given the harsh living conditions of Blacks in poor urban areas, Minorities' Diminished Returns 45 may be more pronounced in Blacks 20,24,25,46-49 .
Lack of wealth and economic reserve contribute to the vulnerability of Black older adults to economic strain and lack of cash 14,117,[167][168][169][170][171][172][173][174][175][176] . Being Black is associated with a hidden tax to the middle class. Due to some historical reasons, Blacks have extremely low levels of wealth. Slavery, redlining, Jim Crow laws, discriminatory lending, mass incarceration, war on crime, and many other policies have kept average Blacks poor [177][178][179][180][181][182][183] . Given such historical injustice, most Black families have not been able to accumulate wealth over time. Such wealth operates as a buffer in times of uncertainty when stress and hassles occur 172,174 .
Excessive financial necessities continue to hurt the health of Black older adults in urban settings. Our findings suggest that policies that provide cash supply at the time of need may be promising as a strategy to tackle health disparities in urban areas where deep and chronic poverty compounds the health effects of multiple chronic diseases. Policy solutions to health disparities should go beyond health policies by including public and economic policies that address economic strain of Black older adults. A promising policy may be lending at the time of emergency.
Interestingly enough, the results reported here are quite consistent with a recent study conducted on older adults residing in a Central and Eastern European country. The authors also found that higher education was not enough to protect individuals against depressive symptoms, when adult socioeconomic position was taken into account. These findings may suggest that the lack of protective effects of education may be a pattern that hold for some populations within and outside the US 184 .

Implications
Economic strain is a social factor modifiable through economic policies (e.g., tax, social welfare, and income redistribution) as well as on-ground programs that can be delivered to the communities (e.g., free tax preparation programs, financial empowerment services, and emergency funds). Reducing economic strain remains a hopeful public policy solution that can reduce or eliminate health inequalities of Blacks in urban areas [150][151][152] . urban areas is very limited [185][186][187][188][189][190][191] . In such a context, a lack of liquid assets (i.e., cash) becomes very detrimental. Economists Hamilton and Darity have proposed baby bonds as a potential solution to prevent economic strain among the next generation of Blacks. Baby bonds are trust accounts of up to $60,000 for every newborn, which are calibrated to the family's wealth. This bond can help generate the required seed money for buying a home, starting a business, or help at the time of highest financial need. Another proposed solution by the same scholars is reparations as a potential solution to the wealth inequality of Blacks. These economic policies are specific proposals that, if written as the law can impact society and have significant national health effects as they effectively increase economic security for the Black community 164 . These bold and innovative economic proposals (e.g., reparation, baby bonds, etc.) may be an effective way to reduce economic as well as health inequalities in the US. Writing these economic interventions as law and policies may be a stable solution to tackling health disparities in low-income Black communities.

What education can do in economically constrained
Although multiple policy solutions can be brought to the table, here we discuss a relevant economic policy that can potentially buffer the effect of economic strain on lowincome Black communities. Income redistribution policies that increase the access of two-income people to cash that can be used to pay bills and buy food are critical 115,192,193 . The challenge is that such policies are not viewed positively in the US, given the over-emphasis of American culture on meritocracy (i.e., relying on bootstraps).
To more effectively bring social justice and equity to the public discourse and law, Blacks should gain more political power, so they get a louder voice and their agenda can be written as law 115,192,193 . Although Blacks need to gain a stronger voice in the US political system, they are affected by poverty and related illnesses. That means the very same people who are supposed to solve the problem are disappearing faster and younger, because of poor healthrelated to poverty. In other words, the ones who need the change most do not get the chance to influence policy as they die earlier. Research shows that early mortality of Blacks results in a missed opportunity to correct the policies that can fix the problem of poverty among Blacks 115,192,193 .
Beyond the policies mentioned, such as reparations, baby bonds, and savings accounts for children, there is also a need for local services that are delivered at the community level to provide support for economic and financial security, especially for older Blacks. Although these policies are possible solutions, they are mainly early childhood methods that may ameliorate economic strain later in life. At the same time, there is a need for a range of solutions where we do not need to wait 70 years to see results of early childhood policy interventions. Another challenge is the likelihood of reparations in the current political climate. Although the need exists and the rationale clear, but these policy solutions may not be viable solutions for older adults, today.
To solve the current problems, safety net public healthcare systems should continue increasingly establishing co-located services (through Social Determinants of Health), and shared interagency centers should address nonmedical social needs, including financial and economic insecurity of individuals with multiple medical and social needs. Examples include free tax preparation programs that increase access to Earned Income Tax Credit and refunds and financial empowerment services that include onsite financial planning, budgeting, credit stabilization, under-banking, etc. In addition to emergency funds which have long existed through local government social service agencies, programs to address poverty concretely, and specifically for older adults, do exist in communities and should be mentioned as a viable way to serve older adults, particularly in underserved neighborhoods. Successful organizations that are on the ground and deliver services include Prosperity Now! and the NYC Department of Consumer Affairs, which is expanding its Financial Empowerment Centers across the city.

Future Research
It is still unknown whether elimination of economic strain through social and economic policies can be used as a sustainable policy to eliminate, or at least reduce, the existing racial gaps across health outcomes 115,192,193 . There are only a few previous attempts and social experiments to reduce the disparities in the health of Black communities by reducing economic strain. There is also a need to study why educational attainment has a "less than expected" health effects for Black older adults in general and particularly in low-income urban contexts.
There is a need for future research to differentiate the aspects that make all vs. Black older adults experience economic strain. Such research would require various ethnic groups, including Whites. Such research in diverse samples may be able to show how the effects of economic strain are compounded for all particularly, Black older adults. While most older adults are not working, and many are socially isolated and live alone, and a large proportion has health conditions. These effects may or may not differ between White and Black older adults, which needs more research. There is also a need to decompose the adverse health effects of economic strain from those of social isolation, aging, and comorbid health conditions. This research would suggest that if people are stabilized financially, their social isolation will still be a remaining risk operating in the background.
There are some other variables that have not been taken into account in the current analysis. Among the most important missing variables are markers of early life socioeconomic position. Studies, however, have shown considerble amount of complexties, nuances, and heterogeneities in this regard. In one study, higher education of the mother was associated with better mental health (e.g. lower depressive symptoms) in older adults, independent of socioeconomic status in adulthood 194 . On the contrary, higher education of the father was associated with lower depressive symptoms in adulthood, but this association is explained by socioeconomic status in adulthood 194 . In addition, no comparison group was used in this study, thus it is unknown whether the same results would be relevant to any non-Black group or not.

Conclusion
Economic strain may be more a more salient social determinant of the health of Black older adults than educational attainment, at least in economically constrained settings. This finding is in line with the literature on MDRs, which has repeatedly shown weaker than expected effects of educational attainment for Black people. Such diminishing returns may be due to structural factors such as social stratification, segregation, and systemic racism across levels and institutions which are embedded in U.S. society.

Funding
The mother study was supported by the Center for Medicare and Medicaid Services grant 1H0CMS331621. Additionally, Assari and Bazargan are supported by the NIH awards 5S21MD000103, 54MD008149, R25 MD007610, U54 TR001627, and 2U54MD007598,

Disclosures
The authors declare that they have no competing interests.