Effect of Mentally Challenging Occupations on Incident Dementia Differs Between African Americans and Non-Hispanic Whites

Abstract

Background:

Engaging in mentally challenging activities may protect against dementia in late life. However, little is known whether the association between mentally challenging activities and dementia risk varies with race/ethnicity.

Objective:

The current study investigates whether having jobs with higher mental stimulation is differentially associated with a decreased risk of dementia between African Americans (AAs) and non-Hispanic Whites (nHWs).

Methods:

The sample consisted of 1,079 individuals (66% nHWs, 28% AAs; age = 78.6±5.3) from the longitudinal Einstein Aging Study. Occupation information of each participant was collected retrospectively at baseline and was linked to the substantive complexity of work score from the Dictionary of Occupational Titles. Cox proportional hazards models were used to evaluate the associations of occupational complexity with risk of dementia.

Results:

Individuals whose jobs had moderate-to-high levels of complexity, compared to those with the lowest complexity, were at modestly decreased risk for incident dementia. When stratified by race, moderate-to-high levels of occupational complexity were significantly associated with lower risk of developing dementia for AAs (HR = 0.35). When risk of dementia was evaluated based on the combinations of race×occupational complexity, AAs with lowest occupational complexity showed the highest risk of developing dementia, while other combinations exhibited lower risk of developing dementia (HRs = 0.36~0.43).

Conclusion:

Our results suggest that moderate-to-high levels of complexity at work are associated with a decreased risk of incident dementia in AAs. Understanding the differential effects of mentally challenging occupations across race/ethnicity may suggest important intervention strategies that could mitigate racial disparities in dementia rates.

Keywords

African Americans dementia occupations race workplace

INTRODUCTION

Up to 50% of dementia cases worldwide may be attributable to preventable causes such as low education, physical inactivity, smoking, and cardiovascular disease [1]. Work-related exposures may be another important source for the risk or protective effects, considering nearly 70% of the adult population in Organization for Economic Co-operation and Development countries spends a substantial amount of time in the workplace [2]. Although several empirical studies found that mentally stimulating occupations were associated with lower risk of developing dementia [3 –10], there has been little evidence regarding the associations between mentally challenging occupations and dementia risk in a racially diverse population. Moreover, little is known whether the role of mentally challenging occupations differs across racial/ethnic groups. The aim of this study is to examine whether having mentally challenging occupations is associated with lower rates of incident dementia among racially diverse older adults, and whether the association between mentally challenging occupations and dementia varies between African Americans (AAs) and non-Hispanic Whites (nHWs).

Prior studies reported higher dementia rates among AAs compared to nHWs [11]. Because being AA and not engaging in mentally challenging occupations were found to be associated with greater risk of developing dementia in late life, AAs who did not engage in mentally stimulating occupations may be at the greatest risk of developing dementia. Although little is known about the combined effects between race and occupational complexity on dementia, prior theories and empirical studies suggest that there may be differences in socioenvironmental effects on health outcomes across racial groups. There are two competing hypotheses on this issue. First, the minority poverty framework [12], originally proposed from the literature on poverty among AAs, proposes that minority individuals may face enormous threats to various health outcomes due to combined disadvantages from race and socioeconomic/environmental factors. In addition, impacts from contextual conditions earlier in the life course become more pronounced with age as initial advantages or disadvantages early in life accumulate across the lifespan [13]. Therefore, negative health outcomes would be greatest among the most disadvantaged group (e.g., AAs with the least challenging occupations), while there would be smaller gap in health outcomes for other combinations.

Second, the diminishing returns hypothesis suggests that AA individuals face overt and covert discrimination and do not experience the same returns as nHWs for high achievement [12, 14]. AAs with cognitively challenging occupations, for example, may face racial discrimination in their everyday life such as diminished returns for investments for their efforts, limited promotion opportunities due to glass ceilings, and so on [14, 15], which are great sources of frustration and stress that may lead to negative health outcomes [16, 17]. In addition, as they try to alter these situations by hard work, they may pay the high cost in terms of mental, physical, and cognitive health. Therefore, this hypothesis predicts that higher levels of mental complexity at work among AAs do not lead to similar positive health outcomes comparable to nHWs.

Although there have been a few empirical studies that examined the effects of race with education, income, and occupational hierarchy on various health outcomes such as cardiovascular risk factors [18, 19], psychological distress [20], and self-rated health [12], there has been no empirical study that examined the effects of race and occupational complexity together on dementia. The current study aims to address this gap in the literature. The Einstein Aging Study (EAS) provides the opportunity to investigate the associations of mentally challenging occupations and incident dementia in a community-based, racially diverse sample. We aimed to examine 1) whether substantive occupational complexity in the longest held job throughout life was associated with lower rates of incident dementia in diverse older adult sample, and 2) whether the effects of mentally challenging occupations differ between AAs and nHWs.

METHODS

Study sample

The EAS is a longitudinal study of cognitive aging and dementia started in 1993. The sample consists of community-dwelling, racially diverse adults systematically recruited from Bronx County, NY. Between 1993 and 2004, information from the Health Care Financing Administration/Centers for Medicaid and Medicare Services rosters for Medicare-eligible persons aged 70 or above was used to develop sampling frames of community-residing participants in Bronx County. Due to the changes in policy releasing this information, the New York City Board of Elections-registered voter lists for the Bronx were used since 2004. Eligible participants were at least 70 years of age, non-institutionalized, ambulatory, fluent in English, without visual/auditory impairments, and without active psychiatric symptomatology. Participants received comprehensive annual assessments including health and psychosocial status, clinical evaluations, a neuropsychological battery, and medical history. Written informed consent was obtained in the initial clinic visit. The Institutional Review Board of the Albert Einstein College of Medicine approved the study protocol. Study details are described elsewhere [21].

For current study, a subsample of participants (n = 1,079) from the EAS who were dementia-free at baseline, had at least one wave of follow-up, were working for salary in their primary job, and retired after the ages of 50 were included in the analyses (Fig. 1). The subsample of selected participants had more years of education (13.7 versus 12.6, p < 0.001) and lower (i.e., better) baseline Blessed Information Memory Concentration (BIMC) test scores (2.3 versus 4.0, p < 0.001) compared to unselected participants. There were no differences in the mean baseline age, sex, and racial/ethnic composition between selected and unselected participants. Among nHWs, subsample of selected participants had more years of education (13.2 versus 11.9, p < 0.001) and lower baseline BIMC scores (3.1 versus 5.3, p < 0.001). Among AAs, there were no significant differences in years of education and baseline BIMC scores between selected and unselected participants. There were no significant differences in baseline age and sex between selected and unselected participants in both nHWs and AAs. Among 1,079 participants, 120 were diagnosed with dementia at follow-up, with 24.2 new cases of dementia per 1,000 persons per year. Table 1 shows number of incident dementia cases between AAs and nHWs by occupational complexity. Occupational complexity was divided into two categories: lower occupational complexity (the bottom quintile of occupational complexity) and higher occupational complexity (the remaining quintiles; see Analytic plans for rationale).

Fig. 1

Sample selection.

Table 1

Number of incident dementia cases between AAs and nHWs by occupational complexity

	Incident dementia		p for the
	No	Yes	chi-square test
AA with higher occupation	200 (92%)	17 (8%)	<0.001
AA with lower occupation	67 (74%)	23 (26%)
nHW with higher occupation	516 (90%)	59 (10%)
nHW with lower occupation	118 (89%)	15 (11%)

Lower occupational complexity: the bottom quintile; higher occupational complexity: the remaining quintiles. AA, African Americans; nHWs, non-Hispanic Whites.

Table 2 summarizes the baseline characteristics of the study participants stratified by occupational complexity and race/ethnicity. Higher occupational complexity group had higher levels of education (14.3 versus 11.5 years), lower baseline BIMC scores (2.1 versus 3.0), had greater percentage of nHWs (67.9% versus 57.3%), and lower percentage of participants with income below poverty line (16.5% versus 38.5%). Participants in nHWs, compared to AAs, were older at study entry (79.3 for nHWs versus 77.3 years old for AAs), retired at later ages (65.8 versus 64.4 years old), had higher scores in occupational complexity (5.2 versus 4.6), more years of education (13.9 versus 13.2 years), lower baseline BIMC scores (1.9 versus 3.1), had less percentages of females (56.1% versus 73.9%) and with income below poverty line (15.3% versus 30.8%), and had fewer medical comorbidities (1.8 versus 2).

Table 2

Baseline characteristics by occupational complexity and race (Mean (Standard error) (Range))

Variable	All participants	By occupational complexity			By race/ethnicity
	(N = 1079)	Lower occupational complexity	Higher occupational complexity	p	nHWs	AAs	Others	p
		(N = 232)	(N = 847)		(N = 708)	(N = 307)	(N = 64)
Baseline age	78.6 (5.3) (67~100)	78.9 (5.1)	78.5 (5.4)	0.263	79.3 (5.4)	77.3 (4.9)	76.6 (5.2)	<0.001
Retirement age	65.3 (5.8) (50~90)	65.1 (5.2)	65.4 (6)	0.394	65.8 (6.1)	64.4 (5.1)	65.0 (5.5)	0.003
Mean follow-up (y)	4.6 (3.5) (0.3~19.4)	4.6 (3.7)	4.6 (3.4)	0.972	4.5 (3.4)	4.8 (3.8)	4.1 (3.0)	0.240
Education (y)	13.7 (3.4) (2~24)	11.5 (2.9)	14.3 (3.3)	<0.001	13.9 (3.3)	13.2 (3.5)	13.5 (4.0)	0.007
Occupational complexity (continuous)	5.0 (1.9) (0.3~10)	2.2 (0.8)	5.8 (1.3)	<0.001	5.2 (1.9)	4.6 (2.0)	5.2 (1.9)	<0.001
BIMC score	2.3 (2.2) (0~13)	3.0 (2.5)	2.1 (2.1)	<.001	1.9 (2.0)	3.1 (2.3)	3.5 (2.7)	<0.001
Number of comorbidities	1.8 (1.2) (0~7)	1.9 (1.2)	1.8 (1.2)	0.972	1.8 (1.2)	2 (1.2)	2.1 (1.4)	0.022
Race, %
nHWs	65.6%	57.3%	67.9%	<0.001	–	–	–	–
Aas	28.5%	38.8%	25.6%
Others	5.9%	3.9%	6.5%
Female, %	61.6%	64.7%	60.8%	0.285	56.1%	73.9%	64.1%	<0.001
Low income, %	21.1%	38.5%	16.5%	<.001	15.3%	30.8%	37.3%	<0.001

nHWs, non-Hispanic Whites; AAs, African Americans; Low income, having income below poverty line. t-tests were conducted to test differences for age, follow-up, education, occupational complexity, BIMC score, and number of comorbidities. Chi-square tests were conducted for race, female, and low income.

Occupational complexity

We utilized the factor score of substantive complexity of work derived from the Dictionary of Occupational Titles (DOT) to indicate occupational complexity in the current study. Details on this measure are described elsewhere [22, 23]. Briefly, these scores are based on the 4^th edition DOT from U.S. Department of Labor and data from the April 1971 Current Population Survey. Occupational analysts performed on-site ratings for 46 occupation-related descriptions (e.g., complexity of work with data, specific vocational preparation) for more than 12,000 different occupational titles coded with DOT occupational categories. Because all ratings on occupational descriptions for the given occupation were assigned at one time by a single analyst in the 4^th edition of the DOT, multiple-item factor scales that tap underlying dimensions of occupational characteristics were developed to improve reliability [23], and we opted to use one of the factor scores that reflect mentally challenging occupations rather than using individual items in occupational characteristics. Ratings of 46 occupational characteristics for the DOT occupation categories were aggregated for the 574 census occupational categories and were factor analyzed using data from the April 1971 Current Population Survey [24], which include information for more than 60,000 workers. Four factors (substantive complexity, motor skills, physical demands, and undesirable working conditions) emerged, and the substantive complexity factor was used to operationalize occupational complexity in the current study. Ten items out of 46 occupational characteristics were loaded strongly on substantive occupational complexity factor: 1–3) general educational development in reasoning/mathematics/language, 4) specific vocational preparation, 5) complexity of functioning with data, 6–8) intelligence/numerical/verbal aptitude, 9) interests in abstract and creative activities, and 10) temperament for non-repetitive processes. All items were standardized and summed to form substantive complexity scale, and for ease of interpretation, the scale was converted to a 0–10 range.

Participants’ occupations of longest duration in their lifetime were identified from self-reports at the initial visit of the EAS, and trained coders classified each participant’s job using a computer system which allowed them to identify three-digit 1970 U.S. Census occupational code using occupation keywords. Then each participant’s job was linked with the substantive complexity of work score from the DOT [23] using 1970 Census codes. As a result of data linkage in occupational complexity scores, 183 unique occupation categories were identified. Table 3 shows seven broad occupation categories created by this occupation coding to provide an idea of the types of job that individuals had in this study. nHWs were more likely to have occupations in the managers and administrators category compared to other races, and AAs were more likely to have occupations in the service workers category. Given the differences in the mean occupational complexity score across broad occupational categories, the different proportions of race in the broad occupational categories seem to indicate different patterns in occupational complexity score across race.

Table 3

Occupational categories of the study sample

Occupational categories (mean±SD for occupational complexity score)	All participants	By race/ethnicity
	(N = 1,079)	nHWs (N = 708)	AAs (N = 307)	Others (N = 64)
Professional, technical (6.9±0.99)	28.7%	28.4%	28.3%	34.4%
Managers and administrators, sales workers (5.98±0.71)	18.8%	20.3%	16.6%	12.5%
Clerical (4.20±1.13)	29.3%	31.2%	25.7%	25.0%
Craftsmen (4.77±0.57)	5.2%	5.8%	3.3%	7.8%
Operatives and transport equipment operatives (2.23±0.69)	6.3%	6.8%	5.5%	4.7%
Laborers &farmers (1.30±0.41)	0.9%	1.4%	0.0%	0.0%
Service workers, private household workers (2.61±1.37)	10.8%	6.1%	20.5%	15.6%

nHWs, non-Hispanic Whites; AAs, African Americans. Occupational categories were created based on the first digit of 1970 Census occupation codes.

Dementia diagnosis

Dementia diagnosis was based on standardized clinical criteria from the DSM-IV [25], which requires impairment in memory and at least one additional cognitive domain, accompanied by evidence of functional decline. Diagnoses were assigned at consensus case conferences, which involved comprehensive review of neurocognitive test results, neurologic symptoms, and functional impairment. To ensure uniformity of diagnosis over time, all individuals evaluated before the release of DSM-IV in 1994 were re-evaluated according to DSM-IV criteria.

Other study measures

Baseline age (centered at the sample mean), retirement age (centered at the sample mean), gender (reference: male), race/ethnicity (reference: nHWs), years of education (centered at the sample mean), and late-life income (above poverty level versus poverty level) were obtained at the time of the baseline survey. Race/ethnicity was categorized as nHWs, AAs, and other (i.e., Hispanic White, Hispanic Black, Asian, and other racial/ethnic groups). To calculate comorbidities, presence of arthritis, angina, diabetes, chronic obstructive pulmonary disease, high blood pressure, myocardial infarction, congestive heart failure, Parkinson’s disease, and stroke was identified through self-reports at baseline and during annual follow-up assessments. Depressive symptoms were assessed using the 15-item Geriatric Depression Scale (GDS) [26] and scores of six or above indicated moderate to severe depressive symptoms. Binary variable was created for each item to indicate whether participants ever experienced each symptom during entire assessments or not (0 = no, 1 = yes). The final comorbidities score was defined by the total sum of these ten items (range = 0 to 10). Global cognition was assessed using BIMC test [27], with higher scores indicating more severe impairment (range = 0 to 33).

Analytic plans

Cox proportional hazards regression models were used to model the risk of dementia in association with occupational complexity. Participant’s age was used as the reference point in the Cox model for the timing of dementia diagnosis. The possible bias due to left truncation was statistically handled in all Cox regression analyses [28]. The Wald chi-square statistics were used to test the null hypothesis. We used BRESLOW method to handle ties, but there were no ties in the survival times in the current dataset. In the preliminary analysis, we examined both linear and quadratic effects of occupational complexity on incident dementia in the univariate Cox model because prior studies suggested potential nonlinear associations between mentally challenging/demanding activities and health outcomes [3 , 30]. Linear effect of occupational complexity was not significant but quadratic effect was significant (p = 0.015), suggesting a potential threshold effect (i.e., exposure to occupational complexity above a certain level may be protective against dementia). To further examine the nature of quadratic effect of occupational complexity on incident dementia, we used quintiles of occupational complexity and examined the hazard ratios (HRs) for incident dementia for quintile of complexity contrasted with the lowest quintile (Fig. 2). The HRs for all other quintiles were less than one (HRs = 0.56 to 0.68, ps = 0.038 to 0.136) suggesting lower risk of dementia. As other quintiles were not significantly different from each other in risk, we grouped them together for the compact and clear presentation of results. Consequently, a binary indicator for the occupational complexity (lower occupation: lowest occupational complexity; higher occupation: moderate-to-high occupational complexity) was used in all subsequent analyses. Since we did not find the evidence of non-linear effects of education, continuous scores of education were used for analyses.

Fig. 2

Hazard ratio (HR) and 95% confidence intervals (CIs) for dementia by quintile of occupational complexity based on the result from univariate Cox proportional hazard model. The 1^st (lowest) quintile of occupational complexity is the reference group.

To examine the association between occupational complexity and risk of dementia, models were sequentially constructed. Due to the high correlation between occupational complexity and education (r = 0.56), we first ran models without education and then included education to check potential confounding effects of education in the associations between occupational complexity and dementia. Therefore, Model 1 included occupational complexity, retirement age, sex, and race/ethnicity and Model 2 added education. To rule out the possible confounding effects of income [31, 32], vascular and other comorbidities [33], Model 3 further controlled for these factors. To examine the effects of occupational complexity in each race/ethnicity group, the above models (excluding race/ethnicity) were separately fit in nHWs and AAs (“other” race was not examined due to the small sample size, n = 64). Finally, to investigate the combined effects of race and occupational complexity on incident dementia, we assigned participants to one of the four combinations of race and occupational complexity (i.e., AAs with lower occupation, AAs with higher occupation, nHWs with lower occupation, nHWs with higher occupation).

Before conducting main analyses, we examined whether the proportional hazards assumptions were supported in the above models. Visual inspection of Kaplan-Meier curves [34, 35] and statistical/graphical tests of the Schoenfeld residuals suggested that the proportional hazards assumptions were satisfied in all sets of models. All major statistical analyses were performed using SAS 9.4 software (SAS Institute, Inc., Cary, NC).

RESULTS

The effects of occupational complexity on incident dementia

A total of 1,079 individuals were included in this study (mean age at baseline = 78.6, mean follow-ups=4.6 years). Table 4 summarizes findings from Cox proportional hazards regression models assessing the association between occupational complexity and incident dementia in entire sample. The results indicated that higher occupational complexity, compared to lower occupational complexity, was associated with a significantly lower dementia risk in Model 1 (HR = 0.63, 95% CI = [0.43, 0.93]). The effect was slightly attenuated when education was added (HR = 0.65, 95% CI = [0.43, 0.99], Model 2) and lost significance when further adjusted for income and comorbidities (HR = 0.68, 95% CI = [0.43, 1.07], Model 3), though neither of these variables were themselves significant.

Table 4

Results from Cox proportional hazards regression examining the effects of occupational complexity on dementia risk for all sample

	Model 1			Model 2			Model 3
	HR	95% CI	p	HR	95% CI	p	HR	95% CI	p
Occupational complexity (Ref: lower)	0.63	0.43, 0.93	0.021	0.65	0.43, 0.99	0.046	0.68	0.43, 1.07	0.098
Years of education	–	–	–	0.99	0.93, 1.05	0.68	0.99	0.93, 1.05	0.711
Retirement age	1.02	0.99, 1.05	0.258	1.02	0.99, 1.05	0.257	1.01	0.98, 1.05	0.527
Female	0.93	0.63, 1.37	0.714	0.93	0.63, 1.37	0.704	0.89	0.59, 1.34	0.586
African-American (Ref: nHWs)	1.43	0.95, 2.14	0.087	1.43	0.95, 2.14	0.087	1.38	0.89, 2.13	0.146
Other race (Ref: nHWs)	1.39	0.60, 3.22	0.444	1.38	0.60, 3.21	0.449	1.64	0.70, 3.87	0.258
Low income	–	–	–	–	–	–	1.34	0.85, 2.11	0.212
Comorbidities	–	–	–	–	–	–	0.91	0.77, 1.09	0.312

nHWs, non-Hispanic Whites; AAs, African Americans; Low income, having income below poverty line.

The effects of occupational complexity on incident dementia in nHWs and AAs

To investigate the effects of occupational complexity in nHWs and AAs, we ran Cox proportional hazards regression in two separate ethnic/race groups. Among AAs, higher, compared to lower, occupational complexity was associated with a reduced likelihood of incident dementia (Table 5, Model 1, HR = 0.31, 95% CI = [0.16, 0.63]) after controlling for retirement age and sex. Further adjustment for education (in Model 2), income, and comorbidities (in Model 3) reduced the HRs of occupational complexity slightly, but individuals in the top four quintiles, compared with those in the bottom quintile of occupational complexity, still had significantly lower risks of developing dementia (HR = 0.32, 95% CI = [0.15, 0.70] in Model 2; HR = 0.35, 95% CI = [0.16, 0.76] in Model 3). The effects of education, income, and comorbidities were not significant. Among nHWs, the effects of occupational complexity were not significant in any models.

Table 5

Results from Cox proportional hazards regression examining the effects of occupational complexity on dementia risk in nHWs and AAs

	AAs
	Model 1			Model 2			Model 3
	HR	95% CI	p	HR	95% CI	p	HR	95% CI	p
Occupational complexity (Ref: lower)	0.31	0.16, 0.63	0.001	0.32	0.15, 0.7	0.004	0.35	0.16, 0.76	0.008
Years of education	–	–	–	0.99	0.88, 1.12	0.869	0.98	0.87, 1.12	0.799
Retirement age	1.03	0.96, 1.11	0.371	1.03	0.96, 1.11	0.389	1.02	0.95, 1.1	0.575
Female	0.77	0.34, 1.75	0.533	0.77	0.34, 1.75	0.539	0.83	0.36, 1.91	0.661
Low income	–	–	–	–	–	–	1.31	0.63, 2.71	0.466
Comorbidities	–	–	–	–	–	–	0.81	0.59, 1.13	0.214
	nHWs
	Model 1			Model 2			Model 3
	HR	95% CI	p	HR	95% CI	p	HR	95% CI	p
Occupational complexity (Ref: lower)	0.97	0.51, 1.87	0.937	1.06	0.53, 2.12	0.865	1.15	0.57, 2.32	0.708
Years of education	–	–	–	0.97	0.9, 1.05	0.469	0.98	0.91, 1.06	0.566
Retirement age	1.01	0.97, 1.05	0.676	1.01	0.97, 1.05	0.688	1.01	0.97, 1.05	0.7
Female	0.96	0.59, 1.57	0.882	0.95	0.58, 1.56	0.844	0.91	0.55, 1.49	0.699
Low income	–	–	–	–	–	–	1.48	0.78, 2.81	0.235
Comorbidities	–	–	–	–	–	–	0.96	0.76, 1.21	0.733

nHWs, non-Hispanic Whites; AAs, African Americans; Low income, having income below poverty line.

The combined effects of race and occupational complexity on incident dementia

Next, we assigned participants to one of the four combinations of race and occupational complexity (i.e., AAs with lower occupation, AAs with higher occupation, nHWs with lower occupation, nHWs with higher occupation). Risk of dementia was evaluated based on this group membership in a Cox analysis. In the fully adjusted model, dementia risk was significantly lower in AAs with higher occupational complexity (Table 6, Model 3, HR = 0.39, 95% CI = [0.20, 0.78]), nHWs with higher occupational complexity (HR = 0.43, 95% CI = [0.25, 0.76]), and nHWs with lower occupational complexity (HR = 0.36, 95% CI = [0.17, 0.76]) compared to AAs with lower occupational complexity. There were no significant differences in the risk of dementia among AAs with higher occupation, nHWs with higher occupation, and nHWs with lower occupational complexity. The covariates-adjusted survival curve comparing the probability of dementia-free survival by occupational complexity and race/ethnicity is shown in Fig. 3.

Table 6

Results from Cox proportional hazards regression examining the effects of race×occupational complexity on dementia risk

	Model 1			Model 2			Model 3
	HR	95% CI	p	HR	95% CI	p	HR	95% CI	p
Race×OCC group*
AAs with higher OCC	0.35	0.18, 0.68	0.002	0.38	0.19, 0.75	0.005	0.39	0.20, 0.78	0.008
nHWs with higher OCC	0.37	0.22, 0.62	0.000	0.40	0.23, 0.69	0.001	0.43	0.25, 0.76	0.003
nHWs with lower OCC	0.36	0.17, 0.76	0.007	0.36	0.17, 0.75	0.007	0.36	0.17, 0.76	0.007
Education (Ref: lower)	–	–	–	0.97	0.91, 1.03	0.351	0.98	0.92, 1.04	0.429
Retirement age	1.02	0.98, 1.05	0.361	1.02	0.98, 1.05	0.384	1.01	0.98, 1.05	0.453
Female	0.93	0.61, 1.41	0.727	0.92	0.61, 1.41	0.710	0.89	0.58, 1.35	0.577
Low income	–	–	–	–	–	–	1.38	0.86, 2.23	0.181
Comorbidities	–	–	–	–	–	–	0.90	0.75, 1.08	0.267

nHWs, non-Hispanic Whites; AAs, African Americans; OCC, occupational complexity; Low income, having income below poverty line. *Reference: AAs with lower occupational complexity.

Fig. 3

Covariates-adjusted dementia-free survival curves by occupational complexity and race/ethnicity. nHW, non-Hispanic Whites; AA, African Americans; OCC, occupational complexity.

DISCUSSION

We examined whether occupational complexity of longest held job throughout life was associated with incident dementia in a racially diverse, systematically recruited sample of older adults, and specifically tested whether the effects of occupational complexity differed between nHWs and AAs. We observed significant associations between a decreased dementia risk and moderate-to-high levels of occupational complexity; the association became nonsignificant as we further adjusted for income levels and comorbidities although these variables were not significant. The significant effects of occupational complexity on incident dementia were mainly driven by the AAs in the lowest quintile of occupational complexity. Among AAs, moderate-to-high levels of occupational complexity were significantly associated with a decreased dementia risk even after controlling for well-known risk factors such as education, income, cardiovascular or other comorbidities.

Race, occupational complexity, and dementia

Previous studies have found that working in mentally challenging occupations was associated with lower risk of developing dementia [3 –9]. However, race/ethnicity factors were not examined at all or were included only as covariates in analyses without evaluating the effects of racial groups. The present findings expand on previous studies by providing evidence of the differential effect of occupational complexity on dementia across race/ethnicity.

We observed significant associations between occupational complexity and dementia among the AA population. The association between occupational complexity and dementia was not linear and a threshold or breaking point existed beyond which occupational complexity may be protective. Among AAs, moderate-to-high levels of occupational complexity were associated with lower risk of developing dementia compared to the lowest levels of occupational complexity. When we compared individuals from different combinations of race/ethnicity and occupational complexity, AAs with lowest occupational complexity exhibited highest risk of developing dementia among all combinations of race/ethnicity and occupational complexity. In line with the minority poverty hypothesis [12], it is likely that AA individuals with lowest occupational complexity face combined disadvantages due to race and little mental stimulation throughout their adulthood. The combined effects may become more pronounced in later life as disadvantages accumulate through midlife to old age [13]. Therefore, AAs who had the lowest complex occupations may show worse health outcomes than AAs who had higher levels of mentally demanding occupations or nHWs who had the lowest occupational complexity. Interestingly, AAs with higher occupational complexity enjoyed the similar cognitive benefits to nHW individuals. That is, AAs may show special benefit from having jobs with cognitive stimulation, which can help to remedy disparities in dementia rates.

Many researchers argue that racial taxonomies are socially constructed [36, 37]. Even within the same race/ethnicity group, there may be large within-race variations and heterogeneities as well as between-race heterogeneities for socio-environmental factors such as occupation, levels of education, income, and access to healthcare [37, 38]. The present finding suggests one of the important socio-environmental factors (i.e., occupation) that may account for the heterogeneities in AAs. If there exist social factors influencing cognitive health outcomes within the given race/ethnicity group, failure to account for those factors may result in inconsistent findings or misinterpretation of results. Instead of using race as a proxy for the social/environmental or even genetic factors, race may need to be further decomposed into specific factors that account for heterogeneity within the race group in dementia/health/intervention studies [36, 37].

We did not find the significant effect of occupational complexity among entire sample in the fully adjusted model, and significant effects of occupational complexity were found only among AAs. The mean levels of occupational complexity were higher among nHWs compared to AAs; moreover, nHWs had a lower proportion of the lowest occupational complexity compared to AAs, which might have contributed to this divergent result. Potential mechanisms underlying the differential effects of occupation between AAs and nHWs will be discussed in the following section. However, it should be noted that the EAS cohort is a systematically recruited sample to represent the community from which it was derived, and it matches closely the US Census demographic data for the Bronx at the time of recruitment [21]. Thus, selection bias that may arise from clinics or volunteers could be minimized. Most of the prior studies [4–6 , 8] that examined the associations between occupational demands/complexity and dementia did not report racial/ethnic composition in their samples. Moreover, current study differs from prior studies in terms of sample characteristics, measures of occupational complexity, occupation characteristics from different geographic locations (Bronx, NY, the US; other areas in the US; European countries), and sampling methods (e.g., volunteers or clinic samples, living in the community or institutions), all of which make it hard to directly compare the current study results to other studies. Future disparity studies from community-based sample in different geographical locations will be necessary to replicate the current finding.

Potential mechanisms

Because AD neuropathology progressively and slowly develops across the span of several decades, different underlying mechanisms between AAs and nHWs may be more clearly revealed in midlife or even earlier periods. There may be several potential mechanisms that account for the differential effects of occupational complexity across race/ethnicity.

First, differences in quantity and quality of education between AAs and nHWs may influence employment options, resulting in different dementia rates between AAs and nHWs by occupational complexity. AA older adults in the US experienced discrimination in the quality and quantity of schooling, as large portion of them attended de jure segregated schools [39]. Then poorer early-life education among AAs compared to nHWs may limit occupational choices among AAs, which may also limit opportunities for cognitively stimulating activities associated with occupation. In addition, systematic discrimination in obtaining mentally complex occupations among AAs may contribute to the current findings. Even if AAs have similar educational levels to nHWs, AAs with equal educational qualifications are less likely to obtain jobs commensurate with their education in comparison with HWs [40], which may deprive the opportunities among AAs to accumulate cognitive skills and resources across adulthood. Then, the influence from social and contextual conditions earlier in the life course becomes more pronounced with age as initial advantages or disadvantages accumulate across the lifespan [13], resulting in the worst cognitive health outcomes among AA older adults who had occupations with the lowest complexity.

Second, perceived discrimination at work among AAs may account for the different associations between occupation and dementia risk across race. AAs with lower occupational complexity are more likely to experience unfair treatment, discrimination, and stigma [41]. Then, enhanced neuroendocrine responses from repeated and chronic stress ‘wear and tear’ the body and influence neural mechanisms underlying cognitive function [16, 17]. For AAs with higher occupational complexity, however, cognitive skills and resources built through mentally engaging occupations may alleviate the negative effects of chronic stress [42].

Third, lower occupational complexity may be associated with psychological or behavior risk factors (e.g., smoking, stress) of vascular disease, which in turn may increase the risk of dementia. Vascular dementia accounts for a large proportion of dementia cases in AAs [43]. In the current study, late-life vascular factors did not appear to substantially contribute to the risk of dementia in AAs. However, previous studies found that late-life dementia is more associated with midlife vascular risk factors, such as obesity, hypertension, and high cholesterol, than late-life vascular factors [43, 44]. We do not have measures of midlife vascular factors, which should be examined in future studies.

Limitations and future directions

A few limitations of our study have to be discussed. First, the measure of occupational complexity was drawn from the DOT, which assessed occupational characteristics based on workers function and traits that each occupation requires. This raises the possibility of reversed causation (i.e., individuals with specific personality or cognitive abilities are likely to select themselves into particular occupations), and caution is required in interpreting the current results. It was beyond the scope of this study to evaluate whether participant’s mentally challenging experiences in the workplace or preexisting traits (e.g., cognitive abilities, interests, or personality) affected the choice of occupation, which is predictive of late-life cognitive health. Future study needs to investigate whether there are mediating or confounding effects of occupational complexity in the association between earlier factors (i.e., premorbid intelligence) and dementia by race/ethnicity. Second, the present study did not test or control for the effects of lifespan factors (e.g., childhood SES) due to the unavailability of those measures. Third, we could not test another potential mediating or confounding factors, such as health behaviors, access to healthcare, and midlife vascular factors, which may underlie the associations reported here.

In the prior study with the EAS participants and the same occupational complexity measure, we found that occupational complexity was significantly associated with higher levels of cognition and more rapid cognitive decline among those who did not develop dementia [22]. Although we did not examine the role of race/ethnicity in levels and rates of change in cognition, other studies found that racial differences were significant in levels of cognition but were not significant or small in rates of cognitive decline [45 –47]. In addition, given that there may be no direct relations between rates of cognitive decline and dementia when levels of cognition were accounted for [48], the association among race/ethnicity, occupational complexity, and cognition may differ across normative/pathologic cognitive aging processes (i.e., levels of cognition, rates of cognitive decline, and dementia), which requires further investigation.

Conclusion

Identifying modifiable risk and protective factors of dementia in diverse race/ethnicity can inform interventions and policies. Although prior evidence suggests potential protective effects of mentally challenging occupations against dementia, this is the first to evaluate differential and combined effects of occupational factors with race on dementia. Using a racially diverse sample from the EAS, we found evidence of reduced likelihood of incident dementia related to moderate-to-high levels of work complexity among AAs. The effect of occupational complexity was not accounted for by education, income, cardiovascular and other comorbidities. Given that large variations within the same race/ethnicity exist, future dementia research on health disparities is necessary to take both race/ethnicity and socio-environmental (e.g., occupational) factors influencing cognitive outcomes into account. Future intervention studies or policies on racial disparities should investigate whether targeting AAs in the least complex occupations would be helpful to reduce disparities and prevent dementia among AA individuals.

Footnotes

ACKNOWLEDGMENTS

This research was supported by National Institutes on Aging (NIA) grants P01 AG003949, NIA grant T32 AG049676 to the Pennsylvania State University, the Sylvia and Leonard Marx Foundation, and the Czap Foundation.

Authors’ disclosures available online ().

References

Barnes

, Yaffe

(2011) The Projected impact of risk factor reduction on Alzheimer’s disease prevalence, Alzheimers Dement 7, S511.

Employment rate (OECD), https://data.oecd.org/emp/employment-rate.htm

Karp

, Andel

, Parker

, Wang

H-X

, Winblad

, Fratiglioni

(2009) Mentally stimulating activities at work during midlife and dementia risk after age 75: Follow-up study from the Kungsholmen Project, Am J Geriatr Psychiatry 17, 227–236.

Smyth

, Fritsch

, Cook

, McClendon

, Santillan

, Friedland

(2004) Worker functions and traits associated with occupations and the development of AD, Neurology 63, 498–503.

Potter

, Helms

, Burke

, Steffens

, Plassman

(2007) Job demands and dementia risk among male twin pairs, Alzheimers Dement 3, 192–199.

Kröger

, Andel

, Lindsay

, Benounissa

, Verreault

, Laurin

(2008) Is complexity of work associated with risk of dementia? The Canadian Study of Health and Aging, Am J Epidemiol 167, 820–830.

Andel

, Crowe

, Pedersen

, Mortimer

, Crimmins

, Johansson

, Gatz

(2005) Complexity of work and risk of Alzheimer’s disease: A population-based study of Swedish twins, J Gerontol B Psychol Sci Soc Sci 60, 251–258.

Stern

, Gurland

, Tatemichi

, Tang

, Wilder

, Mayeux

(1994) Influence of education and occupation on the incidence of Alzheimer’s disease, JAMA 271, 1004–1010.

Then

, Luck

, Heser

, Ernst

, Posselt

, Wiese

, Mamone

, Brettschneider

, König

H-H

, Weyerer

, Werle

, Mösch

, Bickel

, Fuchs

, Pentzek

, Maier

, Scherer

, Wagner

, Riedel-Heller

(2017) Which types of mental work demands may be associated with reduced risk of dementia?, Alzheimers Dement 13, 431–440.

10.

Then

, Luppa

, Schroeter

, König

H-H

, Angermeyer

, Riedel-Heller

(2013) Enriched environment at work and the incidence of dementia: Results of the Leipzig longitudinal study of the aged (LEILA 75+), PloS One 8, e70906.

11.

Chin

, Negash

, Hamilton

(2011) Diversity and disparity in dementia: The impact of ethnoracial differences in Alzheimer’s disease, Alzheimer Dis Assoc Disord 25, 187–195.

12.

Farmer

, Ferraro

(2005) Are racial disparities in health conditional on socioeconomic status?, Soc Sci Med 60, 191–204.

13.

Dannefer

(1987) Aging as intracohort differentiation: Accentuation, the Matthew effect, and the life course, Sociol Forum 2, 211–236.

14.

Cole

, Omari

(2003) Race, class and the dilemmas of upward mobility for African Americans, J Soc Issues 59, 785–802.

15.

Clark

, Anderson

, Clark

, Williams

(1999) Racism as a stressor for African Americans: A biopsychosocial model, Am Psychol 54, 805–816.

16.

Lupien

, McEwen

, Gunnar

, Heim

(2009) Effects of stress throughout the lifespan on the brain, behaviour and cognition, Nat Rev Neurosci 10, 434–445.

17.

McEwen

(1998) Stress, adaptation, and disease: Allostasis and allostatic load, Ann N Y Acad Sci 840, 33–44.

18.

Waitzman

, Smith

(1994) The effects of occupational class transitions on hypertension: Racial disparities among working-age men, Am J Public Health 84, 945–950.

19.

Frierson

, Howard

, DeFina

, Powell-Wiley

, Willis

(2013) Effect of race and socioeconomic status on cardiovascular risk factor burden: The Cooper Center Longitudinal Study, Ethn Dis 23, 35–42.

20.

Kessler

, Neighbors

(1986) A new perspective on the relationships among race, social class, and psychological distress, J Health Soc Behav 27, 107–115.

21.

Katz

, Lipton

, Hall

, Zimmerman

, Sanders

, Verghese

, Dickson

, Derby

(2012) Age and sex specific prevalence and incidence of mild cognitive impairment, dementia and Alzheimer’s dementia in blacks and whites: A report from the Einstein Aging Study, Alzheimer Dis Assoc Disord 26, 335–343.

22.

Hyun

, Katz

, Lipton

, Sliwinski

(2019) Mentally challenging occupations are associatedwithmore rapid cognitive decline at later stages of cognitive aging. J Gerontol B Psychol Sci Soc Sci. doi: 10.1093/geronb/gbz122.

23.

Roos

, Treiman

(1980) DOT scales for the 1970 Census classification. In Work, jobs, and occupations: A critical review of occupational titles, Miller AR, Treiman DJ, Cain PS, Roos PA, eds. National Academy Press, Washington, DC, US, pp. 336-389.

24.

NationalAcademy of Sciences.Committee on Occupational Classification and Analysis (1984) Dictionary of Occupational Titles (DOT): Part I – Current Population Survey, April 1971, Augmented with DOT Characteristics and Dictionary of Occupational Titles (DOT): Part II – Fourth Edition Dictionary of DOT Scores for 1970 Census Categories: Version 2.

25.

American Psychiatric Association (1994) Diagnostic and statistical manual of mental disorders. 4th ed., American Psychiatric Association, Washington, DC.

26.

Sheikh

, Yesavage

(1986) Geriatric Depression Scale (GDS): Recent evidence and development of a shorter version, Clin Gerontol 5, 165–173.

27.

Blessed

, Tomlinson

, Roth

(1968) The association between quantitative measures of dementia and of senile change in the cerebral grey matter of elderly subjects, Br J Psychiatry 114, 797–811.

28.

SAS (2015) SAS/STAT® 14.1. User’s guide, SAS Institute Inc., Cary, NC.

29.

Lyketsos

, Chen

L-S

, Anthony

(1999) Cognitive decline in adulthood: An 11.5-year follow-up of the Baltimore Epidemiologic Catchment Area study, Am J Psychiatry 156, 58–65.

30.

De Jonge

, Reuvers

, Houtman

, Bongers

, Kompier

(2000) Linear and nonlinear relations between psychosocial job characteristics, subjective outcomes, and sickness absence: Baseline results from SMASH, J Occup Health Psychol 5, 256–268.

31.

Evans

, Hebert

, Beckett

, Scherr

, Albert

, Chown

, Pilgrim

, Taylor

(1997) Education and other measures of socioeconomic status and risk of incident Alzheimer disease in a defined population of older persons, Arch Neurol 54, 1399–1405.

32.

Scazufca

, Almeida

, Menezes

(2010) The role of literacy, occupation and income in dementia prevention: The São Paulo Ageing & Health Study (SPAH), Int Psychogeriatr 22, 1209–1215.

33.

Vassilaki

, Aakre

, Cha

, Kremers

, St. Sauver

, Mielke

, Geda

, Machulda

, Knopman

, Peterson

, Roberts

(2015) Multimorbidity and risk of mild cognitive impairment, J Am Geriatr Soc 63, 1783–1790.

34.

Foreman

, Lai

, Miller

(2008) Surviving left truncation using PROCPHREG. In Western Users of SAS Software Meeting, University City, CA Citeseer.

35.

Lamarca

, Alonso

, Gomez

, Muñoz

(1998) Left-truncated data with age as time scale: An alternative for survival analysis in the elderly population, J Gerontol A Biol Sci Med Sci 53, M337–M343.

36.

Williams

(1997) Race and health: Basic questions, emerging directions, Ann Epidemiol 7, 322–333.

37.

Ighodaro

, Nelson

, Kukull

, Schmitt

, Abner

, Caban-Holt

, Bardach

, Hord

, Glover

, Jicha

(2017) Challenges and considerations related to studying dementia in blacks/African Americans, J Alzheimers Dis 60, 1–10.

38.

Bryc

, Durand

, Macpherson

, Reich

, Mountain

(2015) The genetic ancestry of African Americans, Latinos, and European Americans across the United States, Am J Hum Genet 96, 37–53.

39.

Glymour

, Manly

(2008) Lifecourse social conditions and racial and ethnic patterns of cognitive aging, Neuropsychol Rev 18, 223–254.

40.

Hyllegard

, Lavin

(1992) Higher education and challenging work: Open admissions and ethnic and gender differences in job complexity. In Sociological Forum. Springer, pp. 239-260.

41.

Zahodne

, Manly

, Smith

, Seeman

, Lachman

(2017) Socioeconomic, health, and psychosocial mediators of racial disparities in cognition in early, middle, and late adulthood, Psychol Aging 32, 118–130.

42.

Hyun

, Sliwinski

, Almeida

, Smyth

, Scott

(2018) The moderating effects of aging and cognitive abilities on the association between work stress and negative affect, Aging Ment Health 22, 611–618.

43.

Babulal

, Quiroz

, Albensi

, Arenaza-Urquijo

, Astell

, Babiloni

, Bahar-Fuchs

, Bell

, Bowman

, Brickman

(2019) Perspectives on ethnic and racial disparities in Alzheimer’s disease and related dementias: Update and areas of immediate need, Alzheimers Dement 15, 292–312.

44.

Alzheimer’s Association (2018) 2018 Alzheimer’s disease facts and figures, Alzheimers Dement 14, 367–429.

45.

Gross

, Mungas

, Crane

, Gibbons

, MacKay-Brandt

, Manly

, Mukherjee

, Romero

, Sachs

, Thomas

(2015) Effects of education and race on cognitive decline: An integrative study of generalizability versus study-specific results, Psychol Aging 30, 863–880.

46.

Sachs-Ericsson

, Blazer

(2005) Racial differences in cognitive decline in a sample of community-dwelling older adults: The mediating role of education and literacy, Am J Geriatr Psychiatry 13, 968–975.

47.

Weuve

, Barnes

, de Leon

, Carlos

, Rajan

, Beck

, Aggarwal

, Hebert

, Bennett

, Wilson

(2018) Cognitive aging in Black and White Americans, Epidemiology 29, 151–159.

48.

Korten

, Henderson

, Christensen

, Jorm

, Rodgers

, Jacomb

, Mackinnon

(1997) A prospective study of cognitive function in the elderly, Psychol Med 27, 919–930.