Religiosity, Religious Beliefs, and Cognitive Impairment among Black and White Men with Modest Incomes

Abstract

Objective:

The objective of this study was to examine the association between religious practices, beliefs, and cognitive impairment among Black and White men with modest incomes in the 2016 Health and Retirement Study.

Methods:

Data were drawn from Black and White men who reported annual incomes at or below $50,000 (n=926). The primary outcome was any cognitive impairment, a dichotomous variable derived from a modified version of the Telephone Interview for Cognitive Status. The religious variables were religious service attendance, private prayer frequency, and religious beliefs.

Results:

Results from regression models indicated that religious service attendance was inversely related with cognitive impairment among White men (PR=0.64, CI:0.48-0.87). Private prayer (PR=0.67, CI:0.47-0.97) and religious beliefs (PR=0.91, CI:0.84-1.00) were inversely related to cognitive impairment among Black men.

Discussion:

Our results suggest that religious practices and beliefs may contribute to cognitive preservation among Black and White men, but longitudinal studies are needed to examine these associations further.

Introduction

Racial disparities in Alzheimer’s disease and related dementias (ADRD) have been widely documented (Alzheimer’s Association., 2019, 2020). Cognitive impairment is a major risk factor for ADRD and older Black and other adults from marginalized racial groups have lower levels of cognitive functioning than their non-Hispanic White peers (Castora-Binkley et al., 2015; Weuve et al., 2018). Persistent patterns of racial differences in cognitive impairment have been generally attributed to racial disparities in medical comorbidities including hypertension and diabetes. Social factors may additionally contribute, with socioeconomic status being one of the few social factors evaluated in studies examining health disparities in cognitive impairment and ADRD. Results from this small body of work indicate that educational attainment and related factors (e.g., literacy) attenuate a small portion of the racial gap in cognitive impairment and ADRD outcomes (Chen et al., 2021; Forrester et al., 2018). The effort to explain disparities in cognitive impairment is critical and results are promising; however biological and social factors only account for a fraction of the racial gap in ADRD risk factors and related outcomes (Barnes & Bennett, 2014; Chen et al., 2021). Racial disparities may be a function of contextual rather than individual differences. Attention is therefore shifting to environmental, sociocultural, and behavioral factors (Hill et al., 2015) that may impact disparities in later life cognitive outcomes.

Highlighting distinctions between social and economic environments of Black and White adults can aid in understanding their influence on cognitive impairment. A robust body of research provides evidence indicating that Black adults are disproportionately more likely than their White peers to live in or near areas characterized by insufficient employment opportunities, substandard educational resources, inadequate housing, family disruption, general disorder, pollution, and danger (Caunca et al., 2020; Diez Roux & Mair, 2010; LaViest & Wallace, 2000; Massey, 2004; Ross & Mirowsky, 2001; Usher et al., 2018; Wilson, 1987). Black adults also have frequent encounters with individuals and institutions perceived to be racially motivated (Chatters et al., 2021; Feagin, 1991; Feagin & Sikes, 1994; Jones, 2000, 2002; Krieger, 2003; LaFave et al., 2022). Research has suggested that distressing environments, particularly ones characterized by low income and disinvestment in the community, and persistent hostile encounters can contribute to disparities in onset and progression of chronic diseases and related conditions like cognitive impairment (Alzheimer’s Association., 2019; Barnes & Bennett, 2014; Bruce et al., 2009; Bruce et al., 2015b; Umberson et al., 2014). It has been suggested that these unfavorable social and economic circumstances associated with structural racism, institutional discrimination and persistent unfair treatment over a lifetime contribute to the poor health profile of Black men and their truncated life spans relative to other race and gender populations (Brown & Hargrove, 2018; Bruce et al., 2009; Bruce et al., 2015a; Thorpe Jr & Whitfield, 2018b; Thorpe, Kennedy-Hendricks, et al., 2015; Thorpe et al., 2017). Black men with limited economic resources or opportunities experience oppression, commodification, and criminalization like no other group of men in the United States (Bruce et al., 2015a; Bruce & Thorpe Jr., 2019; Griffith et al., 2019; Thorpe Jr & Whitfield, 2018a; Williams, 2003, 2015). Bruce, Griffith, and Thorpe (2015b) introduced a biopsychosocial heuristic model of chronic disease integrating elements from the life course approach (Szanton et al., 2010; Thorpe & Kelly-Moore, 2013), cumulative disadvantage theory (Dannefer, 2003; Ferraro, 2006; Ferraro et al., 1997; Ferraro & Kelley-Moore, 2003; O’Rand, 1996), and the weathering hypothesis (Geronimus, 1992; Geronimus et al., 2001; Geronimus et al., 1996; Geronimus et al., 2006) to specify how social and economic factors contribute to racial health disparities. Few studies have examined how social and economic factors can have direct and indirect implications for cognitive functioning in Black men (Sims et al., 2015) and there is a need to examine the full range of factors accelerating or delaying cognitive impairment among this population.

Faith is thought to have health benefits and the science examining the relationship between religiosity and/or spirituality and health has grown considerably in the past two decades (Bruce et al., 2017; Koenig et al., 2012). Faith and related religious institutions have been associated with survival in African Americans (Giger et al., 2008; Lincoln, 1974) and Black men in particular (Bruce & Thorpe Jr., 2019; Robinson et al., 2018; Robinson & LaBalle, 2018). Religious institutions have been regarded as safe places where males are welcomed and given affirmative spiritual messages to help them cope with hostile and stressful social environments (Frazier & Lincoln, 1974; Lincoln, 1974). Recent studies have also indicated that faith and related institutions can have health benefits for Black men. Attending religious services has been associated with reductions in the likelihood of mobility limitation or mortality among Black men (Bruce et al., 2022; Bruce et al., 2021) and positive religious coping was associated with resilience and improvements in quality of life (Bruce et al., 2020; Ellis et al., 2015).

Few studies have examined the association between religiosity and cognitive status among Black adults and of these, even fewer have focused exclusively on men (Bruce & Thorpe Jr., 2019; Henderson et al., 2021). This investigation is guided by the heuristic model introduced by Bruce, Griffith, and Thorpe (2015b) and its purpose is to examine the cross-sectional association between religiosity, religious beliefs, and cognitive impairment in Black and White men with modest incomes in a sample of adults 50 years of age and older in the United States. We hypothesized that greater frequency of religious practices and higher levels of religious beliefs would be associated with a lower likelihood of cognitive impairment. We also hypothesized that these associations would vary by race, being stronger in Black than White men.

Methods

Data for this study were drawn from the Health and Retirement Study (HRS), a biennial nationally representative survey of non-institutionalized adults in the United States who are 50 years of age or older. HRS was launched in 1992 and data collection was based on a birth cohort-based study design integrating new cohorts as they reached 50-60 years of age. Participants are included in HRS via a multistage area population sampling design that oversampled Black and Hispanic adults (Heeringa & Connor, 1995). In-person interviews and self-administered surveys were used to collect data on health and related biologic, economic, psychological, and social factors to track aging and retirement process for individuals during middle- and late-life (Crosswell et al., 2020; Sonnega et al., 2014). HRS data are publicly available and de-identified and a detailed description of the HRS is available elsewhere (Heeringa & Connor, 1995; Juster & Suzman, 1995). Analyses using these data are not classified as human subjects research and not subject to IRB review.

The present analyses included individuals in 2016 who identified themselves as male, Black or White, and reported annual incomes at or below $50,000, an amount below the median income for men in 2016 ($51,640) (Semega et al., 2017). This study also included participants who completed the core questionnaire as well as the Psychosocial and Lifestyle Questionnaire, often referred to as the “Leave Behind Questionnaire” (LBQ). The LBQ included psychological, social, and lifestyle measures to complement core HRS data and was administered to participants for completion after the in-person portion of the HRS interview (Crosswell et al., 2020). Participants with missing data for any of the study variables were excluded. The study sample for these complete case analyses included 926 Black and White men and the algorithm used to define this cohort of participants is depicted in Figure 1.

Figure 1.

Algorithm used to define cohort of participants

Measures

The cognitive impairment measure used in this study was derived from the Telephone Interview for Cognitive Status (TICS) (Brandt et al., 1988; Whitfield et al., 2009). The TICS is a cognitive screening test (Amano et al., 2022) that was modeled after the Mini-Mental State Exam (Folstein et al., 1975) to be administered over the phone. This screening test has been used in large population studies and includes measures of episodic memory, serial of 7 subtraction, counting backwards, object naming, and recall of the date, current president, and current vice-present (Crimmins et al., 2011). Langa and colleagues (2010) developed an approach to defining cognitive impairment and dementia and established cutpoints from the composite scores on the TICS (range 0 to 27) to classify participants into three categories: probable dementia (0-6 points), cognitive impairment without dementia (7-11 points) and normal cognitive function (12-27 points) (Langa et al., 2010; Maharani et al., 2020). We collapsed the cognitive impairment without dementia (CIND) and dementia categories into a single “any cognitive impairment” category (Hale et al., 2020) to ensure a robust analysis given the small number of Black and White men classified as experiencing dementia (n=49). Individuals with TICS scores less than 12 were coded as 1 and all others were coded as 0.

Religious service attendance, private prayer frequency, and religious beliefs were the primary independent variables of interest. Religious service attendance was derived from a core HRS item asking, “About how often have you attended religious services during the past year?”. The original response categories were “more than once per week” (coded 1), “once per week” (coded 2), “two or three times per month” (coded 3), “one or more times per year” (coded 4) and “not at all” (coded 5). For these analyses, religious service response categories were truncated to align with the current literature (Bruce et al., 2017; Chen et al., 2020) and reverse coded where greater frequency of attendance was represented by a higher score: “not at all” (coded 0), “less than weekly” (coded 1), and “at least weekly” (coded 2). Private prayer frequency was derived from responses to an LBQ item asking respondents to report how often they “pray privately in places other than a church or synagogue?”. Response categories ranged from “daily” (coded 1) to “never” (coded 7). For these analyses, the response categories were truncated and reverse coded in manner where greater frequency of prayer was represented with a higher score: “never” (coded 0), “less than monthly” (coded 1), “at least once a month” (coded 2), “at least once a week” (coded 3), and “daily” (coded 4). The religious beliefs variable was drawn from the Brief Multidimensional Measure of Religiousness/Spirituality (Fetzer Institute., 2003) included in the LBQ. This instrument asked respondents to indicate the degree of agreement or disagreement with the following statements, “I believe in a God who watches over me”; “The events in my life unfold according to a divine of greater plan”; “I try hard to carry my religious beliefs over into all my other dealings in life”; and “I find strength and comfort in my religion”. The response categories ranged from “strongly disagree” (coded 0) to “strongly agree” (coded 5). The sum of responses was used as an index where a higher score represented a higher level of religious beliefs (possible range = 0-20).

Covariates that have been found to be related to either religiosity and/or cognitive impairment were included (Barnes & Bennett, 2014; Bruce & Thorpe Jr., 2019; Henderson et al., 2021; Hosseini et al., 2017). Demographic covariates included race, age, and marital status. Race was a dichotomous variable indicating whether respondents identified themselves as Black or White. Age was continuous variable derived from an item asking respondents to report their age in years. Marital status was represented by a dichotomous variable indicating whether a respondent reported being married or not. The socioeconomic covariate included education, categorized as three indicator variables denoting less than 12 years of education, high school graduate, or completed more than 12 years of formal education. The health-related covariates in this study were derived from items asking respondents to rate their health status, report the number of chronic conditions they ever had, recent depressive symptoms, and indicate social engagement activities. Self-rated health was represented by a dichotomous variable indicating whether respondents rated their health as fair or poor (coded 1) vs. good or excellent (coded 0). The number of chronic conditions variable reflected the total number of affirmative responses to items asking if they had ever been told by a clinician that they had a chronic disease (e.g., hypertension, diabetes, stroke, heart failure, arthritis, lung disease; possible range = 0-7). The depressive symptoms variable was drawn from the Center for Epidemiologic Studies Depression Scale (CES-D) (Radloff, 1977). The CES-D was developed to measure the frequency of depressive symptoms and has been used in other studies considering the association between religion and cognition-related outcomes (Corsentino et al., 2009; Reyes-Ortiz et al., 2008). The HRS included an eight-item version of the CES-D (Steffick & HRS Health Working Group., 2000) and participants had two response options for each item, no (coded 0) or yes (coded 1). The sum of responses was used as an index where a higher score represented a greater number of depressive symptoms (possible range = 0-8). The social engagement variable was derived from responses to LBQ items asking respondents about the frequency of participation in public-facing engagement activities including active involvement with younger family members, volunteer events with local children or adolescents, charity work, enrollment in educational or training courses, attendance at social, sport, or other clubs, membership in non-religious political or community groups, and participation in a local arts group. The response categories were “never” (coded 0), “not in the last month” (coded 1), “at least once a month” (coded 2), “several times a month” (coded 3), “once a week” (coded 4), “several times a week” (coded 5) and “daily” (coded 6). The sum of responses was used as an index where a higher score represented a higher level of social engagement (possible range = 0-42).

Statistical Analyses

Results were generated using unweighted data from participants with complete data on all study variables to minimize bias from sample weights and imputations (Fletcher & Dixon, 2012; Korn & Graubard, 1995; Lines et al., 2022). Descriptive statistics were generated to describe characteristics of the total sample and by race. Group differences between Black and White men were assessed with t-tests for continuous variables or chi-square tests for categorical variables. The prevalence of any cognitive impairment in the sample exceeded 10%; therefore, we estimated Modified Poisson Regression models with robust standard errors to generate prevalence ratios and corresponding 95% confidence intervals (McNutt et al., 2003; Thorpe et al., 2017; Zou, 2004). To examine the association between each respective religious measure and cognitive impairment, separate models for religious service attendance, prayer frequency, and religious beliefs were estimated, each adjusting for demographic, socioeconomic, and health covariates. Previous work has established that the patterns of association between religious variables and outcomes can vary by race (Bruce et al., 2020; Bruce et al., 2021) and race-stratified Modified Poisson Regression models with robust standard errors were estimated to determine the degree to which religiosity and religious beliefs varied by race. For all analyses, statistically significant p-values are less than or equal to 0.05. The descriptive and regression analyses were completed using STATA (Version 17).

Results

Descriptive statistics for the total sample and by race are presented in Table 1. Statistically significant racial disparities were observed. Black men comprised 23.3% of the analytic sample, were on average 7 years younger than their White peers (64.9 ± 9.7 vs 72.0 ± 11.3) and were less likely to report being married (39.5% vs 61.7%). Black men had more depressive symptoms (1.9 ± 2.0) than their White peers (1.5 ± 2.0) on average. Black men also had higher levels of social engagement (6.8 ± 6.7) than White men (5.8 ± 5.3) in the study. Nearly half of Black men (48.1%) in the study had TICS scores classifying them as having any cognitive impairment while the corresponding percentage of White men was 27.6%. The percentage of Black men attending religious services (74.5%) was significantly larger than the corresponding proportion of White men (67.5%) and Black men had higher average religious belief index scores (4.2 ± 1.3) than their White peers (3.9 ± 1.4). There were no statistically significant race differences in education, self-rated health, the average number of chronic conditions, or prayer frequency.

Table 1.

Characteristics of Black and White Men Reporting Income Less than $50,000 Dollars in the 2016 Health and Retirement Study

Variables	Total (n=926)	White Men (n=710)	Black Men (n=216)	p-value
Age, mean (std)	70.3 (11.4)	72.0 (11.3)	64.9 (9.7)	<0.001
Married, (%)	56.5	61.7	39.5	<0.001
Education, %				0.250
Less than 12 years	25.3	24.1	29.2
12 years	38.1	38.2	38.0
More than 12 years	36.1	37.8	32.9
Self-rated poor/fair health, (%)	38.0	36.2	44.0	0.088
Chronic Conditions, mean (std)	2.5 (1.5)	2.6 (1.5)	2.4 (1.6)	0.205
Depressive Symptoms, mean (std)	1.6 (2.0)	1.5 (2.0)	1.9 (2.0)	0.002
Social Engagement, mean (std)	6.0 (5.6)	5.8 (5.3)	6.8 (6.7)	0.017
Religious Service Attendance, (%)				0.022
None	31.4	33.5	24.5
Less than weekly	31.0	29.2	37.0
Weekly or more	37.6	37.3	38.4
Personal Prayer Frequency, (%)				0.769
Never	22.0	22.3	21.3
Less than monthly	6.6	6.8	6.0
At least once per month	13.3	14.0	11.1
At least once per week	16.3	15.9	17.6
Daily	41.8	41.1	44.0
Religious Belief Index, mean (std)	4.0 (1.4)	3.9 (1.4)	4.2 (1.3)	0.001
Cognitive Function, (%)				<0.001
Normal	67.6	72.4	51.9
CIND	27.1	22.8	41.2
Demented	5.9	4.8	6.9
Any Cognitive Impairment, (%)	32.4	27.6	48.1	<0.001

The proportions presented are the rounded, so they may not add up to 100.

Any cognitive impairment includes individuals with Telephone Interview for Cognitive Status (TICS) scores classifying respondents into cognitive impairment without dementia (CIND) or probable dementia categories.

Model results, including prevalence ratios (PR) and confidence intervals (CI) for the association between religious service attendance and any cognitive impairment are presented in Table 2. In the total sample, men who attended religious services less than weekly had a lower prevalence of any cognitive impairment than their peers who did not attend religious services (PR=0.78, CI:0.62-0.98) after adjusting for demographic, socioeconomic, and health factors. Examining the association within race, the prevalence ratio for White men who attended religious services less than weekly was statistically significant (PR=0.64, CI:0.48-0.87). White men who attended religious services less than weekly had a lower prevalence of any cognitive impairment than their White peers who did not attend at all, whereas for Black men, religious service attendance was not associated with any cognitive impairment.

Table 2.

Association between Any Cognitive Impairment and Religious Service Attendance among Black and White Men Reporting Income Less $50,000 Dollars in the 2016 Health and Retirement Study

	Total Sample			White Men			Black Men
Religious Service Attendance	Model 1	Model 2	Model 3	Model 1	Model 2	Model 3	Model 1	Model 2	Model 3
Never	Ref
Less than weekly	0.80 0.64 – 1.02	0.80 0.63 – 1.00	0.78 0.62 – 0.98	0.67 0.49 – 0.91	0.65 0.48 – 0.87	0.64 0.48 – 0.87	1.16 0.78 – 1.71	1.16 0.80 – 1.69	1.11 0.76 – 1.62
At least weekly	0.92 0.74 – 1.14	0.93 0.75 – 1.15	0.95 0.76 – 1.18	0.80 0.62 – 1.03	0.79 0.61 – 1.02	0.82 0.63 – 1.07	1.27 0.87 – 1.85	1.36 0.93 – 1.94	1.40 0.95 – 2.08
Black	2.17 1.80 – 2.61	2.03 1.68 – 2.46	1.98 1.64 – 2.40
Age	1.03 1.02 – 1.04	1.03 1.02 – 1.04	1.03 1.02 – 1.04	1.04 1.02 – 1.05	1.04 1.03 – 1.05	1.04 1.03 – 1.05	1.01 1.00 – 1.02	1.01 1.00 – 1.02	1.02 1.01 – 1.03
Married		0.99 0.82 – 1.19	0.97 0.80 – 1.16		1.10 0.86 – 1.41	1.07 0.84 – 1.37		0.83 0.62 – 1.11	0.82 0.62 – 1.10
Education
Less than 12 years
12 years		0.76 0.63 – 0.92	0.82 0.68 – 0.99		0.63 0.50 – 0.80	0.68 0.53 – 0.87		1.06 0.79 – 1.42	1.13 0.83 – 1.54
More than 12 years		0.35 0.27 – 0.46	0.38 0.29 – 0.50		0.26 0.19 – 0.37	0.28 0.20 – 0.40		0.60 0.40 – 0.90	0.65 0.43 – 0.98
Fair or poor health			1.48 1.22 – 1.79			1.40 1.09 – 1.78			1.54 1.15 – 2.05
Chronic Conditions			0.93 0.87 – 0.99			0.93 0.86 – 1.01			0.92 0.83 – 1.02
Depressive Symptoms			1.03 0.98 – 1.08			1.03 0.97 – 1.09			1.04 0.96 – 1.12
Social Engagement			1.02 1.00 – 1.03			1.00 0.98 – 1.03			1.02 1.00 – 1.04

Note: Table entries are unstandardized coefficients.

Model 1 adjusts for demographic covariates only.

Model 2 additionally adjusts for the socioeconomic covariate.

Model 3 is the fully adjusted model that includes demographic, socioeconomic, and health covariates.

Prevalence ratios for the association between private prayer frequency and any cognitive impairment are presented in Table 3. In the total sample, personal prayer frequency was unrelated to the prevalence of any cognitive impairment, but the prevalence ratios for the race variable were significant across all of the total sample models. The prevalence ratio for race (PR=1.95, CI:1.61-2.37) indicated that the prevalence of any cognitive impairment for Black men was nearly twice the corresponding prevalence for their White peers, even after adjustment for religious variables, demographic, socioeconomic, and health factors. The race-stratified models in Table 3 suggested that the association between private prayer and any cognitive impairment varied by race. As shown in Table 3, the prevalence ratio in the fully adjusted model for Black men who reported daily private prayer had a lower prevalence of any cognitive impairment than their Black peers who did not pray (PR=0.67, CI:0.47-0.97). Private prayer frequency was not statistically associated with cognitive impairment in White men. Table 4 presents prevalence ratios for the association between religious beliefs and any cognitive impairment. The patterns in the models for religious beliefs were like those for private prayer. In the total sample and model for White males, the prevalence ratio for the religious belief index was not statistically significant (Table 4). The prevalence ratio for the race variable (PR=1.96, CI:1.61-2.38) was significant, even after adjusting for demographic, socioeconomic, and health factors. In the race-stratified models, the prevalence ratio for religious beliefs was statistically significant for Black men (PR=0.91, CI:0.84-1.00) but not White men (PR=1.02, CI:0.94-1.11), indicating that the level of religious beliefs among Black men was inversely related to any cognitive impairment in the fully adjusted model.

Table 3.

Association between Cognitive Status and Personal Prayer Frequency among Black and White Men Reporting Income Less than $50,000 Dollars in the 2016 Health and Retirement Study

	Total Sample			White Men			Black Men
Personal Prayer Frequency	Model 1	Model 2	Model 3	Model 1	Model 2	Model 3	Model 1	Model 2	Model 3
Never	Ref.	Ref	Ref	Ref	Ref	Ref	Ref	Ref	Ref
Less than Monthly	0.69 0.45 – 1.07	0.75 0.48 – 1.15	0.73 0.48 – 1.11	0.65 0.35 – 1.19	0.76 0.42 – 1.36	0.79 0.44 – 1.41	0.79 0.43 – 1.46	0.80 0.43 – 1.42	0.66 0.37 – 1.15
At least once per month	0.73 0.53 – 1.01	0.76 0.56 – 1.04	0.76 0.56 – 1.04	0.70 0.47 – 1.05	0.78 0.53 – 1.16	0.80 0.55 – 1.18	0.78 0.48 – 1.28	0.76 0.46 – 1.23	0.73 0.45 – 1.19
At least once per week	0.86 0.65 – 1.13	0.91 0.70 – 1.19	0.90 0.68 – 1.18	0.81 0.56 – 1.17	0.91 0.64 – 1.29	0.97 0.68 – 1.39	0.91 0.61 – 1.34	0.88 0.59 – 1.30	0.81 0.54 – 1.21
Daily	0.74 0.60 – 0.92	0.80 0.64 – 0.99	0.81 0.64 – 1.02	0.77 0.59 – 1.02	0.83 0.63 – 1.08	0.89 0.67 – 1.19	0.69 0.49 – 0.96	0.74 0.53 – 1.04	0.67 0.47 – 0.97
Black	2.12 1.76 – 2.55	2.00 1.65 – 2.42	1.95 1.61 – 2.37
Age	1.03 1.02 – 1.04	1.03 1.02 – 1.04	1.03 1.02 – 1.04	1.03 1.02 – 1.05	1.04 1.02 – 1.05	1.04 1.03 – 1.05	1.01 1.00 – 1.03	1.01 1.00 – 1.02	1.02 1.00 – 1.03
Married		1.00 0.84 – 1.20	0.98 0.82 – 1.18		1.07 0.84 – 1.36	1.04 0.82 – 1.32		0.89 0.67 – 1.19	0.88 0.66 – 1.16
Education
Less than 12 years		Ref	Ref		Ref	Ref		Ref	Ref
12 years		0.77 0.63 – 0.93	0.82 0.67 – 0.99		0.63 0.50 – 0.80	0.67 0.53 – 0.86		1.04 0.77 – 1.40	1.07 0.79 – 1.45
More than 12 years		0.36 0.28 – 0.47	0.39 0.29 – 0.50		0.27 0.19 – 0.37	0.29 0.20 – 0.41		0.61 0.41 – 0.91	0.61 0.41 – 0.91
Fair or poor health			1.45 1.20 – 1.75			1.39 1.09 – 1.78			1.44 1.08 – 1.92
Chronic Conditions			0.94 0.88 – 1.00			0.93 0.86 – 1.01			0.94 0.85 – 1.04
Depressive Symptoms			1.02 0.98 – 1.07			1.03 0.97 – 1.09			1.02 0.94 – 1.09
Social Engagement			1.02 1.00 – 1.04			1.00 0.98 – 1.03			1.03 1.01 – 1.05

Note: Table entries are unstandardized coefficients.

Model 1 adjusts for demographic covariates only.

Model 2 additionally adjusts for the socioeconomic covariate.

Model 3 is the fully adjusted model that includes demographic, socioeconomic, and health covariates.

Table 4.

Association between Cognitive Status and Religious Beliefs among Black and White Men Earning Less than $50,000 Dollars in the 2016 Health and Retirement Study

	Total Sample			White Males			Black Males
	Model 1	Model 2	Model 3	Model 1	Model 2	Model 3	Model 1	Model 2	Model 3
Religious Beliefs Index	1.00 0.94 – 1.07	0.98 0.92 – 1.05	0.99 0.93 – 1.05	1.05 0.96 – 1.15	1.01 0.93 – 1.10	1.02 0.94 – 1.11	0.91 0.83 – 0.99	0.91 0.84 – 0.99	0.91 0.84 – 1.00
Black	2.13 1.76 – 2.58	2.01 1.66 – 2.44	1.96 1.61 – 2.38
Age	1.03 1.02 – 1.04	1.03 1.02 – 1.04	1.03 1.02 – 1.04	1.04 1.02 – 1.05	1.04 1.03 – 1.05	1.04 1.03 – 1.05	1.01 1.00 – 1.02	1.01 1.00 – 1.03	1.02 1.01 – 1.03
Married		0.97 0.82 – 1.18	0.97 0.81 – 1.16		1.03 0.81 – 1.30	1.01 0.80 – 1.27		0.93 0.70 – 1.25	0.92 0.69 – 1.23
Education
Less than 12 years	Ref	Ref	Ref		Ref	Ref		Ref	Ref
12 years		0.75 0.63 – 0.91	0.81 0.67 – 0.98		0.61 0.49 – 0.77	0.67 0.52 – 0.85		1.07 0.80 – 1.43	1.12 0.83 – 1.52
More than 12 years		0.35 0.27 – 0.46	0.38 0.29 – 0.50		0.26 0.19 – 0.37	0.29 0.20 – 0.41		0.61 0.41 – 0.91	0.62 0.41 – 0.94
Fair or poor health			1.48 1.22 – 1.78			1.41 1.10 – 1.80			1.48 1.11 – 1.97
Chronic Conditions			0.94 0.88 – 1.00			0.93 0.86 – 1.01			0.93 0.84 – 1.03
Depressive Symptoms			1.02 0.98 – 1.07			1.03 0.98 – 1.03			1.02 0.94 – 1.10
Social Engagement			1.01 1.00 – 1.03			1.00 0.98 – 1.03			1.03 1.00 – 1.05

Note: Table entries are unstandardized coefficients.

Model 1 adjusts for demographic covariates only.

Model 2 additionally adjusts for the socioeconomic covariate.

Model 3 is the fully adjusted model that includes demographic, socioeconomic, and health covariates.

Discussion

Cognitive impairment is a major risk factor for Alzheimer’s disease, the seventh leading cause of death among all adults in the United States (Center for Disease Control and Prevention., 2024). Like other chronic conditions, the burden of cognitive impairment and Alzheimer’s disease is not equally distributed across populations (Adkins-Jackson et al., 2023). Black adults have lower levels of cognitive functioning and a higher prevalence of Alzheimer’s disease and related dementias than their non-Hispanic White peers (Alzheimer’s Association., 2019, 2020; Castora-Binkley et al., 2015; Weuve et al., 2018). Racial disparities in cognitive impairment have been attributed to multiple factors; however, each only slightly attenuates the gap in outcomes between Black and non-Hispanic Whites. Research is critically needed to identify factors that can potentially delay cognitive impairment among high-risk populations. This study examined the extent to which religiosity and religious beliefs are associated with any cognitive impairment among Black and White men with modest incomes. Our results suggest that religious practices and beliefs are inversely related to any cognitive impairment, but with the patterns of associations varying by race. Occasional religious service attendance was significant for White men in the study while personal private prayer and religious beliefs were associated with lower prevalence of any cognitive impairment among their Black peers.

Our findings are consistent with recent work reporting religiosity to be inversely related with cognitive impairment in mixed samples of men and women (Henderson et al., 2021; Hosseini et al., 2017). Our study extends this line of research by highlighting the salience of gender, social class, and race. Gender is often overlooked in health studies focusing on religion and related factors. Men have lower levels of worship service attendance than women (Pew Research Center., 2018); however, religious institutions, symbols, and practices tend to be male-centered (Douglas, 1994; Hopkins, 2014). Our findings indicate that religiosity and religious beliefs may protect men from cognitive impairment and emphasize the importance of examining gender how differences at institutional and individual levels impact cognitive outcomes.

Social class is a major driver of health and is typically represented with variables such as income and education. However, these indicators are inextricably linked with race at individual and community levels (Bruce et al., 1998; Thorpe, Bell, et al., 2015; Thorpe et al., 2020) and can complicate studies examining racial disparities in outcomes like cognitive impairment. Racial disparities in age-related outcomes can be a function of the stark socioeconomic differences between Black and White men (Bruce et al., 2021; Thorpe et al., 2020). We attempted to control for some of these factors by restricting our sample to Black and White men reporting incomes less than $50,000. Descriptive analyses from this sample of men with modest incomes indicated that racial differences in education, self-rated health, chronic conditions, and personal prayer frequency were not statistically significant. This suggests the racial differences in the associations between religiosity and religious beliefs among men are less influenced by income disparities and other covariates; however, the manner through which race and social class collectively influence cognitive impairment is less clear. Race is a sociopolitical construct initially established to legally justify unequal access to economic and social resources essential to health and well-being (Thorpe Jr. et al., 2019). Pernicious forms of racial and economic inequality are embedded in societal structures and its institutions (Beech et al., 2021), and additional studies are needed to specify how these factors influence cognitive functioning and cognitive decline over the life course – as well as factors that may be protective against cognitive impairment.

Religion is an important social resource that can influence behavior and decision making at multiple levels. The pathways through which religious practices and beliefs influence cognitive function have not been specified; however, results from our study indicate that mechanisms linking religious factors and cognitive status among men vary by race. Religion and religious institutions have been considered a important source of coping resources or strategies, and numerous studies have examined the impact of religious coping on variety of health outcomes over the past three decades (Ano & Vasconcelles, 2005; Bruce et al., 2022; Bruce et al., 2020; Bruce & Thorpe Jr., 2019; Hill & Pargament, 2008; Pargament et al., 1998). A qualitative study published by Ellis and colleagues (2015) found that religious practices were a key component of stress coping strategies employed by Black men. Remarkably few studies have examined the degree to which religious resources have implications for chronic disease management and outcomes (Bowie et al., 2017; Dickey & Ogunsanya, 2018). Future studies are needed to determine the direction and degree to which religious coping is associated with delayed disease onset or progression among men in general and Black men in particular.

Our study contributes to the current body of minority aging and health disparities research, but also has limitations worth noting. First, the outcome measure is derived from the TICS, a screening cognitive test. This measure is a limited indicator of cognitive impairment when compared to a more comprehensive neuropsychological battery; however, it is noteworthy that the TICS is used for large population studies rather than clinical encounters (Castanho et al., 2014). Second, religious service attendance and private prayer were used as proxies for religiosity and reflect only two forms of religious practice. Religiosity can be expressed in multiple ways that can vary across and within faith traditions (Bruce et al., 2022). Third, our analysis does not include traditional social support measures; however, recent studies using nationally representative data found that the presence of social support variables did not significantly influence the association between religiosity and health outcomes like mortality (Bruce et al., 2022; Bruce et al., 2017). Nonetheless, we cannot exclude confounding from other aspects of social support not captured with traditional measures. Fourth, the health measures in this study do not sufficiently capture disease states or physiological dysregulation. We additionally cannot exclude residual confounding from factors such as disease severity or duration. Finally, data for this study were drawn from one wave of the HRS – an observational study; therefore, causality cannot be inferred from our results. Moreover, our analysis used survey responses which may be subject to limitations such as recall bias and social desirability.

Our results underscore the potential benefits of religious practices and beliefs for cognitive functioning for middle age and older men. Additional studies are needed to specify the protective mechanisms associated with religiosity, spirituality, and related beliefs. Theoretical and methodological advances in the next generation of studies can pave the way for transdisciplinary approaches integrating commensurable biological, psychological, social, and theological tenets. Results from these efforts can advance our understanding of how religiosity, related beliefs, and other forms of spirituality can preserve cognitive function, slow cognitive decline, and promote longevity among Black men, while simultaneously informing future interventions seeking to reduce racial disparities in aging-related health outcomes.

Data availability statement.

The Health and Retirement Study data are publicly available and can be downloaded at https://hrsdata.isr.umich.edu/data-products .