Socioeconomic Disadvantage, John Henryism and Incident Heart Failure in the Jackson Heart Study

Abstract

Background:

For reasons not fully understood, Black adults experience more socioeconomic disadvantages than their White counterparts, as well as earlier onset and greater mortality from heart failure (HF). The John Henryism hypothesis predicts that repeated high-effort coping, i.e. John Henryism (JH), with socioeconomic adversity can accelerate cardiovascular aging thus increasing the risk for HF.

Methods:

The analysis sample consisted of participants from the Jackson Heart Study (JHS), a cohort of Black adults from the Jackson, Mississippi Metropolitan area. The analyses sample included participants with no cardiovascular disease (CVD) at baseline (2000–2004) with complete socioeconomic disadvantage (SED) and JH scores. Indicators of SED included low household income, low educational attainment, and low maternal educational attainment scores. JH scores (0–36) were categorized as low (<28), moderate (29–32), and high (>32). Effect moderation of JH in the SED and incident HF association was assessed using interaction terms and stratification. Proportional hazards regression determined the hazard ratio (HR) and 95% CI and models adjusted for age, sex, and established lifestyle risk factors.

Results:

Among 1704 participants (mean age: 52.15 years, 64.3% female), 100 HF events occurred by 2016 (mean follow-up 10 years). A statistically significant interaction between income and JH was observed for incident HF (p=0.04). For every 1-unit increase in income disadvantage score, the risk of HF increased two-fold among those with high JH (HR 2.00, 95% CI 1.39, 2.86) after full adjustment. Among participants with low JH, the corresponding unadjusted association was HR 1.40, 95% CI 1.04, 1.90, but this association attenuated after adjusting for age and sex (HR 1.19, 95% CI 0.87, 1.63).

Conclusions:

The association between income disadvantage and HF differed by JH level. Because SED and high JH tend to co-occur, both should be considered in future research aiming to decrease the burden of HF in Black Americans.

Introduction

Heart failure (HF) is a chronic disease characterized by the inability of the heart to pump or fill properly.¹ The disease takes years, if not decades, to develop,^2,3 but it manifests earlier in adulthood for Black Americans than White Americans.^4,5 Without effective management,^6–8 HF also progresses more rapidly for Black Americans⁹ resulting in a greater need for hospital-based care compared to White Americans with the same diagnosis.^10–12 Hence, the combination of an earlier HF onset, and a more rapid advance if inadequately managed,^6–8,13 likely accounts for the striking 2–3 fold excess HF death rate for 35–64 year old Black Americans compared to their White counterparts.¹⁴

Large scale, population-based biracial¹⁵ and multi-ethnic^16,18 cohort studies document the important role of hypertension,^17,18 cardiovascular morbidities (e.g., obesity, dyslipidemia, type 2 diabetes, metabolic syndrome) and behavioral risk factors (e.g., physical inactivity, cigarette smoking, alcohol consumption) play in the excess risk for HF among Black Americans. Increasingly, however, clinicians and researchers^3,11–13,19 acknowledge that many of these comorbidities and behavioral risk factors for HF also emerge earlier in life for US Black adults, thereby contributing to their faster progression to HF. One plausible explanation for this racially patterned complex of health disparities is that Black Americans, as a group, are disproportionately exposed, from an early age,²⁰ to poverty^21,22 and other adverse social conditions^23–26 that can accelerate cardiometabolic aging^26,27 through the frequent activation²⁸ of neurophysiological stress pathways.^29,30

In the US, as in other high-income countries, the inverse gradient between socioeconomic disadvantage (SED) and HF risk^{1,3,22,31–33} is a well-documented finding for all segments of society. Statistical adjustments for age, obesity, and standard behavioral risk factors (smoking, physical inactivity, etc.) usually attenuate the inverse association, but the increased risk for HF observed for individuals with less education or less personal/household income persists.^9,22,32–34 One possible explanation for this robust inverse SED/HF association is that most multivariable analytic models fail to include behavioral and lifestyle risk factors that adequately reflect the neurophysiological stress burdens associated with living under conditions of chronic SED. Analytic models of HF risk that take into consideration, however indirectly, the frequency with which the hypothalamic-pituitary adrenal (HPA) axis is likely to be activated under conditions of SED seem especially relevant for understanding HF in Black Americans given their well-documented greater exposure to stress-inducing environments^26–29 from both racial discrimination²⁰ and economic hardship.²¹ The John Henryism Hypothesis can provide such an analytic framework.

The construct, “John Henryism,” is defined as a strong behavioral disposition to confront systemic adversity (e.g., poverty, or unfair treatment based on race/ethnicity or gender) with “high-effort coping” and a determination to succeed despite the odds.³⁵ The John Henryism construct evokes the legend of John Henry, a Black, late 19^th century, unskilled manual laborer³⁶ who allegedly defeated a machine in a steel-driving contest but dropped dead, immediately thereafter, from complete exhaustion. Hence, the John Henryism hypothesis posits that repeated high-effort coping, within a context of hard-to-overcome adversity, accelerates physiological wear and tear,^26,28 especially on the cardiac-renal-metabolic system.^37–43 More specifically, the John Henryism hypothesis predicts that, other things being equal, the expected inverse gradient for SED and HF risk will be steeper for individuals who score “high” on John Henryism than for those who score “low.” Figure 1 shows the conceptual model of this relationship.

Figure 1:

Conceptual model of the John Henryism Hypothesis. As shown in the figure, as socioeconomic disadvantage increases so does the risk of incident heart failure and increasing John Henryism active coping “acts on” or moderates the relationship.

To our knowledge, the John Henryism hypothesis has not been used to test, prospectively, differential risk for HF in a cohort of Black Americans. Although low educational attainment and low income are common and relevant indicators of SED, childhood SED has been largely overlooked in HF research. We address these gaps by analyzing prospective cohort data collected from participants in the Jackson Heart Study (JHS). Our study tests the hypothesis that the association between SED (using adult education, adult income, and mother’s education) and incident HF will be steeper for Black adults who score high on John Henryism, relative to those who score low.

Methods

The data used for this study can be requested for purposes of reproducing results. Request to access this data set (or other data in the JHS) may be directed to the qualified researchers trained in human subject confidentiality within the JHS Coordinating Center at jhsccrc@umc.edu.

Sample Population

The JHS recruited adults (ages 20–95 years old) who self-identified as Black/African American in the greater Jackson, Mississippi (MS) metropolitan area at baseline (2000–2004). The purpose of the prospective cohort study was to assess risk factors for cardiovascular disease (CVD). Enrolled participants (n=5,306) consisted of 1) shared participants of the Atherosclerosis Risk in Communities (ARIC) study (31%), 2) family members of existing participants (22%), 3) a random selection of adults from the Mississippi Driver’s License list (17%), and 4) community volunteers (30%). Data were collected annually and at later exams: exam 2 (2005–2009) and exam 3 (2010–2013). Written informed consent was provided by all participants, and the study was approved by the institutional review boards of participating institutions’ (Tougaloo College, Jackson State University, and the University of Mississippi Medical Center). Further details of the JHS have been published elsewhere.⁴⁴

Socioeconomic Disadvantage (Exposure)

Socioeconomic Disadvantage (SED), the primary exposure, included 2 variables from exam 1: education and household income. Mother’s education, a proxy of childhood SES, was collected at the first annual follow-up visit to exam 1. Categorical and continuous forms were utilized. Education categories were derived from self-reported responses: <High School, High School/GED, and Some College or more (defined as attended vocational/trade school/attended college). A score was derived for education disadvantage by reverse-coding the number of years of school completed (0–19). Therefore, completing 0 years of school represented the greatest disadvantage for education, and completing 19 years of school represented the least disadvantage. Household income was a JHS derived variable that included reported family income, household size, and the 2000 poverty thresholds from the Census Bureau. Based on reported family income, household size and poverty threshold, the JHS categorized income in 4 categories: “poor, lower-middle, upper-middle and affluent”. Income disadvantage scores were also calculated by reverse-scoring ordinal income categories, with “poor” as a score of greatest disadvantage (3) and “affluent” as a score of least disadvantage (0), representing a linear trend across ordered income levels. Mother’s education, which represents childhood circumstances by the time the participant was 16 years of age, was utilized due to there being fewer missing observations compared to father’s education. The categories were: <High School, High School/GED, Some College (defined as vocational/trade/attended college, college degree or more) or College Degree or more. A score for childhood disadvantage was derived by reverse coding years of mother’s education (0–19), where 0 represented greater disadvantage and 19 years represented the least disadvantage in childhood.

John Henryism (Effect Modifier)

John Henryism was measured during the third follow-up after exam 1 and was operationalized as the sum of scores on the 12-item John Henryism Active Coping Scale (range 0–36).³⁵ The internal consistency of the scale in JHS was good (Cronbach’s α = 0.78). Three sample items from the scale are: (1) Once I make up my mind to do something, I stay with it until the job is completely done; (2) When things don’t go the way I want them to, that just makes me work even harder; and (3) It’s not always easy, but I usually find a way to do the things I really need to get done.” Each item had four response options ranging from 3 to 0, respectively: completely true, somewhat true, somewhat false, and completely false. A higher John Henryism score suggests greater propensity by the individual to engage in high-effort coping with social or economic adversity. Because there are no established clinical or population-based cut points for the John Henryism scale, and we sought to have more than two categories for John Henryism, scores were categorized into tertiles based on their empirical distribution in the JHS. Thus, scores were divided into three categories: low (<28), moderate (29–32), and high (>32).

Incident Heart Failure (Health Outcome)

Trained JHS personnel monitored and adjudicated the surveillance of HF events, which began in 2005. Participants were contacted by phone each year to identify events, and then events were verified by medical record abstraction.⁴⁵ Details on the classification and ascertainment of CVD events have been previously published.⁴⁶ Incident HF was defined as the first inpatient or outpatient diagnosis of HF using the International Classification of Diseases, Ninth Revision code 428, and/or an underlying cause of death code of 150. The classification of HF also included radiographic findings, increased venous pressure >16 mmHg, dilated ventricle/left ventricular function <40% by echocardiography/multiple gated acquisition, or an autopsy finding of pulmonary edema. Event data were available up to December 2016.

Covariates

Baseline covariates included age (continuous) and sex [male/female (referent)]. The American Heart Association’s Life’s Simple 7 categorization (non-ideal vs ideal)⁴⁷ was used to define smoking, weekly physical activity, and body mass index (from in-person measurements). Smoking was categorized as ideal (referent) if the participant reported they were not a current smoker or if they had never smoked and were categorized as non-ideal if they were a current smoker or had quit less than 1 year prior to the baseline examination. Physical activity was categorized as non-ideal if they reported 0 minutes of moderate/vigorous activity, < 150 minutes of moderate activity, < 75 minutes of vigorous activity or <150 minutes of combined moderate and vigorous activity over a week. Participants had ideal (referent) physical activity if they reported ≥150 minutes of moderate activity, ≥ 75 minutes of vigorous activity, or ≥ 150 minutes of combined moderate/vigorous activity. Body mass index (BMI) was considered non-ideal if values were > 25, and BMI was considered ideal (referent) if values were 25 or less.

Based on previous literature,^30,47–49 we also accounted for diabetes, hypertension, and depressive symptoms as potential confounding risk factors for HF. Type 2 Diabetes Mellitus status was defined as: no diabetes (referent) HbA1c < 5.7% and fasting plasma glucose (FPG)< 100 mg/dL; prediabetes: HbA1c ≥ 5.7% but < 6.5% and FPG ≥ 100 mg/dL but < 126 mg/dL, and diabetes: HbA1c ≥ 6.5%, FPG ≥ 126 mg/dL or use of diabetes medication (actual or self-reported) within 2 weeks prior to the clinic visit, or self-reported diabetes. Hypertension status was defined as systolic blood pressure/diastolic blood pressure > 140/90 mmHg, or use of blood pressure-lowering medication, or physician-diagnosed hypertension; no hypertension was defined as having a blood pressure <140/90 and no hypertensive medication use. Depressive symptoms were defined using the Centers for Epidemiologic Studies Depression (CES-D) scale,⁵⁰ which includes 20 items that capture the presence or absence of clinically significant depressive symptoms over one week (e.g., feeling sad, loss of interest, sleep disturbance, low appetite, fatigue). Missing values were imputed to the mean. The scale has a range of 0–60, where a score ≥ 16 indicates a high risk for clinical depression.

Statistical Analysis

The statistical analysis excluded participants with incomplete SED indicators (n=2,531), John Henryism scores (n=283) and any covariates (n=261), resulting in a sample size of 1,987. After removing individuals with a history of CVD at baseline, the sample size was 1828. Those with missing HF status were also removed, thus, the final analytic sample size was 1704. Individuals omitted from the sample due to missing data (vs. those who remained in the sample) were 63% (vs. 65%) women, and 15% (vs. 10%) non-ideal smoking, 60% (vs. 50%) had hypertension, 54% (vs. 54%) had non-ideal BMI, 7% (vs. 5%) had incident HF, with a mean age of 56 (vs. 52) years old. Based on Chi Square and Kruskal-Wallis tests, statistically significant differences in missing data versus non-missing data were observed for age, hypertension status, smoking status, and HF incidence.

The distribution of participant characteristics at baseline was calculated in the full sample and by levels of John Henryism. Chi-square and Kruskal-Wallis test was used to examine differences in characteristics by John Henryism. To evaluate associations between SED indicators, John Henryism, and incident HF, Cox proportional hazards regression models were used to obtain hazard ratios (HRs) and 95% confidence intervals (95% CI), using the continuous disadvantage scores. Deaths unrelated to HF were censored (n=175). The John Henryism hypothesis was evaluated using two criteria prior to stratification by John Henryism: 1) A positive and significant association between SED indicators and incident HF and 2) a significant (p-value <0.05) interaction term between John Henryism and SED indicators in the association with incident HF. Model 1 was unadjusted. Model 2 adjusted for age and sex, and model 3 added smoking, physical activity, BMI, hypertension, and diabetes. The final model added depressive symptoms. A sensitivity analysis formally evaluated effect measure modification in the fully adjusted model. The proportional hazards assumption was evaluated by testing interactions between covariates and log(time). Age violated the proportional hazards assumption; therefore, a time-varying coefficient was included by modeling its interaction with log(time), while all other covariates satisfied the assumption. Statistical significance was inferred by a two-sided p-value<0.05. Analyses were performed using SAS 9.4 (SAS Institute Inc., Cary, NC).

Results

Table 1 shows the distribution of characteristics of the analytic sample overall and by John Henryism level. The mean age was 52.2 years (± 11.8) and 64.3% were female. Most participants had high education, with 78.4% reporting some college or trade school or more. About 40% reported affluent income (40.1%) and most reported that their mother’s education was less than a high school diploma (54.2%). Mean disadvantage scores varied by socioeconomic indicator, with higher average disadvantage scores observed for mother’s education (mean 8.0 ± 4.3) compared with participant education (mean 2.9 ± 3.1) and household income (mean 0.95 ± 0.95). John Henryism scores were approximately normally distributed (mean 29.8 ± 4.2), with 35.7% of participants classified as having low John Henryism, 33.7% moderate, and 30.6% high. With respect to health characteristics, the majority of participants had ideal smoking status (89.0%), while non-ideal physical activity (76.4%) and non-ideal BMI (87.6%) were common. Nearly half of participants had hypertension (49.6%), and 17.7% had diabetes. The mean depressive symptom score was 9.4 (± 5.9).

Table 1:

Characteristics of Jackson Heart Study participants stratified by levels of John Henryism (N=1704)

	Overall	Low John Henryism (n=608)	Moderate John Henryism (n=575)	High John Henryism (n=521)	P-value
Age in years (mean ± SD)	52.2 ± 11.8	51.0 ± 11.6	51.8 ± 11.9	54.1 ± 12.1	<0.001
Sex, n (%)					0.772
Male	609 (35.7)	224 (36.8)	203 (35.3)	182 (34.9)
Female	1095 (64.3)	384 (63.2)	372 (64.7)	339 (65.1)
Socioeconomic Disadvantage
Education, n (%)					<0.001
Less than High School	121 (7.1)	22 (3.6)	40 (7.0)	59 (11.4)
High school graduate/GED	247 (14.5)	79 (13.0)	78 (13.6)	90 (17.3)
Some college/trade school or more	1336 (78.4)	507 (83.4)	457 (79.5)	372 (71.4)
Disadvantage score, mean ± SD	2.9 ± 3.1	2.5 ± 2.7	2.9 ± 3.0	3.4 ± 3.4
Household Income, n (%)					0.521
Poor	137 (8.1)	43 (7.1)	50 (8.7)	44 (8.4)
Lower-Middle	319 (18.7)	108 (17.8)	101 (17.6)	110 (21.1)
Upper-Middle	564 (33.1)	214 (35.3)	185 (32.2)	165 (31.7)
Affluent	682 (40.1)	242 (39.9)	238 (41.5)	202 (38.8)
Disadvantage score, mean ± SD	0.95 ± 0.95	0.93 ± 0.93	0.94 ± 0.97	1.0 ± 0.96
Mother’s Education, n (%)					0.013
Less than High School	923 (54.2)	317 (52.1)	299 (52.0)	307 (58.9)
High School graduate/GED	359 (21.0)	127 (20.9)	118 (20.5)	114 (21.9)
Some college/trade school	231 (13.6)	81 (13.3)	93 (16.2)	57 (10.9)
College degree or more	191 (11.2)	83 (13.7)	65 (11.3)	43 (8.3)
Disadvantage score, mean ± SD	8.0 ± 4.3	7.8 ± 4.3	7.8 ± 4.3	8.5 ± 4.1
John Henryism Score, mean ± SD	29.8± 4.2	25.2 ± 2.7	30.6 ± 1.1	34.2 ± 1.1	<0.001
Smoking Status, n (%)					0.248
Non-ideal/Current Smoker	187 (11.0)	68 (11.2)	71 (12.4)	48 (9.2)
Ideal/Non-smoker or Former-smoker	1517 (89.0)	540 (88.8)	504 (87.7)	473 (90.8)
Physical Activity, n (%)
Non-ideal	1301 (76.4)	478 (78.6)	438 (76.2)	385 (73.9)	0.176
Ideal	403 (23.6)	130 (21.4)	137 (23.8)	136 (26.1)
Body mass index (BMI), n (%)					0.583
Non-ideal/BMI >25	1492 (87.6)	535 (88.0)	506 (88.0)	451 (86.6)
Ideal/BMI ≤25	212 (12.4)	73 (12.0)	69 (12.0)	70 (13.4)
Hypertension Status, n (%)					0.081
No	859 (50.4)	322 (53.0)	295 (51.3)	242 (46.4)
Yes	845 (49.6)	286 (47.0)	280 (48.7)	279 (53.6)
Diabetes Status, n (%)					0.484
No	1402 (82.3)	508 (83.6)	473 (82.3)	421 (80.8)
Yes	302 (17.7)	100 (16.5)	102 (17.7)	100 (19.2)
Depressive symptoms (mean ± SD)	9.4 ± 5.9	9.9 ± 7.3	9.2 ± 6.8	8.9 ± 6.7	0.230

Note: Abbreviations: GED, general education development, SD, standard deviation.

Socioeconomic disadvantage scores were reverse coded in order to capture disadvantage, where greater scores represented greater disadvantage. The range of years of education was 0–19 for both participant and parent. Income disadvantage score ranged from 0–3, and was calculated by reversed coding assigned values 0, 1, 2, 3 to the income categories. John Henryism was captured from the John Henryism Active Coping Scale, which includes 12 items (0–36). John Henryism categories (low, moderate, high) were based on a tertile distribution of the scores from participants. Smoking was categorized as ideal if the participant reported they were not a current smoker or if they had never smoked and were categorized as non-ideal if they were a current smoker or had quit less than 1 year prior to the baseline examination. Physical activity was categorized as non-ideal if they reported 0 minutes of moderate/vigorous activity, < 150 minutes of moderate activity, < 75 minutes of vigorous activity or <150 minutes of combined moderate and vigorous activity over a week. Participants had ideal physical activity if they reported ≥150 minutes of moderate activity, ≥ 75 minutes of vigorous activity or ≥ 150 minutes of combined moderate/vigorous activity. Body mass index (BMI) was considered non-ideal if values were > 25, BMI was categorized as ideal if values were 25 or less. Depressive symptoms scores from the CES-D ranged from 0–60, where scores ≥ 16 indicate a high risk for clinical depression. P-values were calculated using chi-square and Kruskal-Wallis tests as appropriate of differences within John Henryism level, and were considered statistically significant if p-value <0.05

The association between sample characteristics and levels of John Henryism are also shown in Table 1. Aside from household income (p=0.521), respondents who scored higher (i.e., more disadvantaged) on SED also tended to report significantly higher John Henryism scores (p<0.05). Sex, health behaviors, cardiometabolic risk factors, depressive symptoms, and incident HF cases were similar across John Henryism strata, with no statistically significant differences observed.

During follow-up, 100 participants (5.9%) experienced incident HF. After verifying the unadjusted direct associations between SED scores and HF [Education: HR 1.18 (95% CI 1.13, 1.25); Income: HR 1.41 (95% CI 1.16, 1.70); Mother’s Education: 1.11 (95% CI 1.06, 1.17)], we assessed whether John Henryism levels modified these associations. The interactions between John Henryism levels and education disadvantage score (p=0.84) and childhood disadvantage score (p=0.79) were not statistically significant, but the interaction between income disadvantage score and levels of John Henryism was statistically significant (p=0.04). Thus, evidence of effect measure modification by John Henryism was observed when John Henryism was categorized into tertiles; however, the interaction was not statistically significant when John Henryism was modeled continuously, suggesting that the modifying effect may be non-linear or driven by differences at the extremes of the John Henryism distribution rather than a linear gradient.

Table 2 shows the association between income disadvantage score and incident HF stratified by John Henryism category across sequentially adjusted models. Consistent with the John Henryism hypothesis, greater income disadvantage was associated with a substantially higher hazard of HF among individuals with high John Henryism, and this association remained robust across all levels of covariate adjustment. In fully adjusted models accounting for age, sex, health behaviors, cardiometabolic risk factors, and depressive symptoms (Model 4), a 1-unit increase in income disadvantage score was associated with a two-fold higher hazard of HF among individuals with high John Henryism (HR 2.00; 95% CI 1.39, 2.86). Among individuals with low John Henryism, income disadvantage was also associated with an elevated hazard of HF in the unadjusted model (Model 1 HR 1.40; 95% CI 1.04, 1.90). However, this association was attenuated and no longer statistically significant after adjustment for age and sex (Model 2 HR 1.19; 95% CI 0.87, 1.63) and other covariates. There was no statistically significant association between income disadvantage and HF among individuals with moderate John Henryism across models. Table S1 shows the estimates for covariates in the fully adjusted model.

Table 2.

The association of income disadvantage score with incident heart failure by John Henryism category

	Hazards Ratio (HR) 95% Confidence Interval
	Low John Henryism (n=608)	Moderate John Henryism (n=575)	High John Henrysim (n=521)
Incident heart failure events	n=42	n=25	n=33
Model 1	1.40 (1.04, 1.90)	0.96 (0.63, 1.45)	1.90 (1.36, 2.66)
Model 2	1.19 (0.87, 1.63)	0.98 (0.64, 1.51)	1.92 (1.37, 2.69)
Model 3	1.10 (0.78, 1.53)	0.94 (0.60, 1.50)	1.90 (1.34, 2.70)
Model 4	1.12 (0.79, 1.58)	0.94 (0.59, 1.49)	2.00 (1.39, 2.86)

The p-value for interaction for income disadvantage score and John Henryism on incident heart failure was p=0.0415 in the unadjusted model.

Income disadvantage score ranged from 0–3, and was calculated by reversed coding assigned values 0, 1, 2, 3 to the income categories, where higher scores indicated greater disadvantage or lower income. Effect estimates for income disadvantage score represent a 1-unit increase change in income disadvantage associated with heart failure.

Model 1 – Unadjusted

Model 2 – adjusted for interaction term for age and sex

Model 3 – Model 2 + smoking status, body mass index, hypertension, diabetes

Model 4-Model 3+ depressive symptoms

As a sensitivity analysis, we evaluated effect measure modification in Table S2, which summarizes results from a single Cox proportional hazards model including the fully adjusted interaction term between income disadvantage score and John Henryism, with stratum-specific hazard ratios. The results from the stratum-specific estimates were not appreciably different from the stratified models in Table 2. Table S2 also shows results for education disadvantage and childhood disadvantage scores. However, together, these findings indicate that income disadvantage is associated with increased HF risk across John Henryism strata, with the strongest and most consistent associations observed among individuals with high John Henryism.

Discussion

We hypothesized that a strong disposition to repeatedly engage in high-effort coping with socioeconomic adversity (i.e., John Henryism) would help explain the direct association observed between SED and incident HF in Black adults enrolled in the JHS. The significance of the interaction term between SED scores and John Henryism was not consistent, likely reflecting differences in the socioeconomic constructs as well as temporal alignment with John Henryism. Consistent with other studies, incident HF in the current study was elevated among increasing income disadvantage; but, as hypothesized, this elevated risk was most strongly found among individuals who scored high on the John Henryism Scale for Active Coping. This suggests that the physiological wear-and-tear resulting, theoretically,⁵¹ from repeated high-effort coping with SED, specifically income disadvantage, may play a role in the elevated risk for HF.

The increased likelihood of HF being associated with a disposition to repeatedly engage in high-effort coping in a context of socioeconomic adversity is consistent with the allostatic load theory.^51,52 Chronic exposure to difficult life stressors can cause recurring surges in stress hormones and blood glucose, damaging blood vessel endothelia and accelerating the risk for hypertension and related life-shortening cardiometabolic diseases including, but not limited to HF. Our findings suggest that the elevated risk of HF in Black Americans, especially those most economically disadvantaged might be partially explained by the physiological wear-and-tear,^35,51 from repeated high-effort coping with SED. Of note, incident HF among highly disadvantaged individuals who scored in the moderate range of John Henryism had lower effect sizes, though non-significant. While our findings need replication, it nevertheless raises the possibility that a more “moderate” or “strategic” engagement with difficult recurring life stressors that severely disadvantaged people face could slow the physiological wear and tear without undermining their upward mobility aspirations.

Very few studies have examined whether John Henryism moderates the association between SED and the incidence of CVD. The first longitudinal test of the John Henryism hypothesis for CVD examined acute fatal and nonfatal myocardial infarction among Finnish men.⁵³ This study found an interaction between occupation and John Henryism, such that men who worked in blue-collar positions and scored high on John Henryism experienced the greatest hazard for acute myocardial infarction. Another study by Subramanyam et al. found that John Henryism moderated the association between low income and prevalent hypertension among men in the JHS. Similar to our study, however, these authors did not find significant interactions between John Henryism and education.⁵⁴

Our findings should be interpreted within the study limitations. First, the study data are observational which limits our ability to make strong causal inferences. Additionally, generalizability is limited to the Jackson, MS metropolitan area. Second, the results were generated from a complete-case analysis, excluding those who had missing data or those who were lost-to-follow-up, which decreased power. Additionally, SED measures and John Henrysim were captured at one timepoint and were self-reported. Thus, we cannot account for changes in SED or John Henryism at different stages of the life course, nor assume that the measurements are without error. Third, while our analysis considered multiple covariates that have important roles in the association of income and HF, we cannot guarantee that all confounders were considered. Considering these limitations, there are also notable strengths. This study was conducted in a relatively large sample of Black adults in the Southern United States, where CVD cases are common. Additionally, socioeconomic indicators known to be strongly associated with CVD were used to capture disadvantage. Also, the analyses considered the confounding effects of behavioral and lifestyle factors, and the HF events were adjudicated and included hospitalization records.

Conclusions

This study found support for the John Henryism hypothesis. The risk of HF was greatest among those who had increasing income disadvantage and high John Henryism. Though this study was observational, our findings underscore the importance of intersectoral, community-level interventions to improve human capital (e.g., educational and technical skills) of individuals and families living in disadvantage. In addition, primary health care providers should consider screening patients, especially lower-income adults for both chronic stress exposure and high-effort coping behaviors, and when appropriate, connect them to culturally competent health and social services or resources.

Finally, to advance scientific understanding of the molecular basis of the ongoing HF epidemic, adequately powered longitudinal studies should investigate this association and the biomarkers (e.g., proteomic, epigenomic) that illuminate mechanistic pathways between SED, repeated high-effort coping with adversity and HF. Identification of such biomarkers, and the socio-economic contexts giving rise to them could provide a more complete understanding of how HF develops in Black Americans and how the health and financial burden it imposes on them can be decreased.

Supplementary Material

Supplemental Materials: Tables S1–2

What is known?

• Heart failure manifests earlier and has a faster progression in US Black adults compared to other ethnic groups.

• Cardiovascular morbidities such as hypertension, and limited access to socioeconomic resources are major drivers of US racial disparities in heart failure.

• Previous studies have not sufficiently captured the reasons for the faster progression of HF in Black Americans which limits the ability to design timely and effective interventions.

What the study adds?

• Repeated high-effort coping with adversity (i.e., high John Henryism) within a context of limited problem-solving resources (e.g., inadequate family income) could be an important pathway that, along with other factors, contribute to the excess risk of heart failure in US Black adults.

• The findings from this study emphasize the need for studies that investigate real world mechanisms and the need for structural and community efforts to reduce heart failure disparities.

Non-standard Abbreviations and Acronyms

ARIC

Atherosclerosis Risk in Communities Study

CES-D

Centers for Epidemiologic Studies-Depression

CVD

Cardiovascular Disease

FPG

Fasting Plasma Glucose

GED

General Education Development

HF

Heart Failure

HPA

Hypothalamic-Pituitary-Adrenal

JHS

Jackson Heart Study

MS

Mississippi

SD

Standard Deviation

SED

Socioeconomic disadvantage