Volume 19, Issue 8 p. 3426-3434
RESEARCH ARTICLE
Open Access

The impact of attending historically Black colleges and universities on cognitive decline in Black adults: A longitudinal analysis in the KHANDLE and STAR cohorts

Marilyn D. Thomas

Corresponding Author

Marilyn D. Thomas

School of Medicine, Department of Psychiatry and Behavioral Sciences, Weill Institute for Neurosciences, University of California, San Francisco, San Francisco, California, USA

School of Medicine, Department of Epidemiology and Biostatistics, University of California, San Francisco, San Francisco, California, USA

Correspondence

Marilyn D. Thomas, Department of Psychiatry and Behavioral Sciences, UCSF School of Medicine, 1001 Potrero Ave, Box 0852, San Francisco, CA 94110, USA.

Email: [email protected]

Search for more papers by this author
Camilla Calmasini

Camilla Calmasini

School of Medicine, Department of Epidemiology and Biostatistics, University of California, San Francisco, San Francisco, California, USA

Search for more papers by this author
Harmon Khela

Harmon Khela

Departments of Biology and of Public Health Studies, Johns Hopkins University, Baltimore, Maryland, USA

Search for more papers by this author
Taylor M. Mobley

Taylor M. Mobley

Department of Epidemiology, UCLA Fielding School of Public Health, University of California, Los Angeles, Los Angeles, California, USA

Search for more papers by this author
Elizabeth Rose Mayeda

Elizabeth Rose Mayeda

Department of Epidemiology, UCLA Fielding School of Public Health, University of California, Los Angeles, Los Angeles, California, USA

Search for more papers by this author
Christina Mangurian

Christina Mangurian

School of Medicine, Department of Psychiatry and Behavioral Sciences, Weill Institute for Neurosciences, University of California, San Francisco, San Francisco, California, USA

School of Medicine, Department of Epidemiology and Biostatistics, University of California, San Francisco, San Francisco, California, USA

Search for more papers by this author
Lisa L. Barnes

Lisa L. Barnes

Rush Alzheimer's Disease Center, Rush University Medical Center, Chicago, Illinois, USA

Search for more papers by this author
Paola Gilsanz

Paola Gilsanz

School of Medicine, Department of Epidemiology and Biostatistics, University of California, San Francisco, San Francisco, California, USA

Kaiser Permanente Division of Research, Oakland, California, USA

Search for more papers by this author
Rachel A. Whitmer

Rachel A. Whitmer

Kaiser Permanente Division of Research, Oakland, California, USA

Department of Public Health Sciences, University of California Davis School of Medicine, Davis, California, USA

Alzheimer's Disease Research Center, University of California Davis Health, Sacramento, California, USA

Search for more papers by this author
M. Maria Glymour

M. Maria Glymour

School of Medicine, Department of Epidemiology and Biostatistics, University of California, San Francisco, San Francisco, California, USA

Search for more papers by this author
First published: 17 February 2023

Abstract

INTRODUCTION

Black students attending predominantly White institutions (PWIs) versus historically Black colleges and universities (HBCUs) report more harmful discrimination and develop worse mental health outcomes, potentially offsetting the established benefits of college for lowering dementia incidence.

METHODS

Black participants in two cohorts (the Kaiser Healthy Aging and Diverse Life Experiences [KHANDLE] and the Study of Healthy Aging in African Americans [STAR]) who had attended college (N = 716) self-reported the college name (classified as HBCU vs. PWI) and completed three waves of executive function (EF) and verbal episodic memory (VEM) assessments. HBCU effects on cognitive level and decline were estimated using adjusted linear mixed-effects models.

RESULTS

HBCU (vs. PWI) attendees averaged better EF (β = 0.05 [−0.22, 0.32]) and VEM (β = 0.21 [−0.06, 0.46]) at age 70 though neither association was statistically significant. HBCU attendance was associated with slightly faster VEM decline (β = −0.03 [−0.05, 0.00]).

DISCUSSION

Harmonized analyses with larger studies are needed to estimate important effects of HBCU attendance.

Highlights

  • Higher education is robustly linked to lower dementia risk, yet Black–White inequities persist among college-educated adults.
  • Black students attending predominantly White institutions (PWIs) versus historically Black colleges and universities (HBCUs) report more harmful discrimination and develop worse mental health outcomes, which may offset the established benefits of college for lowering dementia incidence.
  • HBCU (vs. non-HBCU) attendees averaged better executive function and verbal episodic memory (VEM) at average age 70, though confidence intervals were wide and associations were not statistically significant, and averaged slightly faster decline in VEM.
  • Harmonized analyses using larger nationally representative studies are likely needed to avoid underestimating the health effects of HBCU attendance.

1 BACKGROUND

Prior studies consistently show racial and socioeconomic patterning of dementia;1-4 however, disentangling each independent effect is difficult given how socioeconomic opportunities and position are inequitably distributed across racial groups.5 The forces of structural racism create racial stratification in nearly every life domain and the drivers of dementia in Black adults may thus diverge from the predominant risk factors for dementia in White adults. Understanding drivers of dementia among Black Americans requires centering on Black experiences that may be sources of resilience against the harmful effects of structural racism.

Higher education is a promising social determinant of health and robustly linked to lower dementia risk1, 2, 6 yet Black–White inequities in dementia persist at the highest levels of education.7 It is widely hypothesized that higher education delays the onset of cognitive decline by increasing cognitive reserve,8-10 described as “capacity of the mature adult brain to sustain the effects of disease or injury (i.e., reduced brain integrity) sufficient to cause clinical dementia.”10 However, a recent study showed that higher years of education contributed to cognitive reserve providing protection against the adverse impact of reduced brain integrity for White adults, but not for Hispanic or Black adults.11 This evidence suggests that the average college experience may not offer the same level of protection against cognitive decline for Black compared to White adults. No studies have evaluated dementia risk among Black adults with distinctively Black educational experiences compared to those with experiences at predominantly White institutions (PWIs), warranting investigation.

Attending an historically Black college or university (HBCU) versus a PWI may offer Black adults more benefits of education for long-term dementia risk. HBCUs were founded after the Civil War when legal racial segregation was established and Black Americans were often barred from higher education.12, 13 HBCUs are largely located in 19 states, mostly in the South with the highest Black populations, and in the District of Columbia.12 After the 1954 Brown v. Board of Education ruling segregation in public schools unconstitutional, Black student enrollment increased for all college institutions in the United States.12 The passing of the 1964 Civil Rights Act expanded financial resources to HBCUs and low-income students, increasing Black student enrollment even further12, 13 with nearly 18% all US Black college students attending an HBCU.12, 14 Over time, HBCUs have contributed to greater social, economic, and political empowerment for Black Americans and Black communities.15, 16

High Black student enrollment was fundamental to HBCUs initially, and these institutions remain predominantly Black (75%).17 Among the myriad reasons Black students choose to attend HBCUs, common influences include the desire to be in an environment that celebrates Black culture and has historical and ongoing institutional pride, high academic quality, and minimal race-related social stressors.18-20 Compared to Black students attending PWIs, Black students who attend HBCUs develop better psychosocial wellness and academic achievement, earn higher incomes, and have better occupational status.21-30 Thus, it is reasonable to postulate that attendance at an HBCU may mitigate Black–White inequities in dementia risk via reduced exposure to economic and racism-related stressors that contribute to neurodegenerative or cerebrovascular disease.5

Evidence of HBCU impacts on aging-related Black health is extremely sparse. This is due in part to the fact that despite wide availability of large data sources comprised of demographic and health variables, few datasets include the specific names or school codes of colleges attended and subsequent cognitive outcomes. A recent study among Medicaid recipients in Northern Manhattan showed that attending an HBCU versus a PWI provided mid-life Black adults some protection against developing metabolic syndrome,31 a condition linked to increased progression of dementia,32, 33 suggesting a pathway through which HBCU attendance may protect against cognitive decline for Black adults.

In the current study, we examined the impact of attending an HBCU compared to a PWI on cognitive decline among Black adults residing in Northern California who attended college during the Civil Rights era. Given the protective role of both cultural capital on education outcomes for Black adults and education on cognitive impairment in general, we hypothesized that Black adults who attended an HBCU would have better cognitive health at study baseline and a slower average rate of cognitive decline than Black adults who attended a PWI.

2 METHODS

2.1 Study participants and data collection

We used pooled data from Black participants of the Kaiser Healthy Aging and Diverse Life Experiences (KHANDLE) cohort and the Study of Healthy Aging in African Americans (STAR). The KHANDLE and STAR studies were approved by the human subjects review boards at Kaiser Permanente Northern California and at University of California San Francisco; all participants provided informed consent.

The KHANDLE cohort includes community-dwelling older adults residing in the San Francisco Bay and Sacramento areas of California and aims to evaluate how race/ethnicity, life course health, and sociocultural factors influence late-life brain health and cognitive decline. Individuals eligible for KHANDLE were long-term members of Kaiser Permanente Northern California (KPNC), an integrated health-care delivery system; were age 65 years or older on January 1, 2017; spoke English or Spanish; and had previously participated in at least one Kaiser Permanente multiphasic health checkup (MHC) between 1964 and 1985. Stratified random sampling by race/ethnicity and educational attainment was used with the goal of recruiting approximately equal proportions of Asian, Black, Latino, and White participants and achieving diversity in educational attainment. Exclusion criteria included: electronic medical record diagnosis of dementia or other neurodegenerative disease (frontotemporal dementia, Lewy body disease, Pick's disease, Parkinson's disease with dementia, Huntington's disease), and presence of health conditions that would impede participation in study interviews, defined by hospice activity in the past 12 months, history of severe chronic obstructive pulmonary disease in the past 6 months, congestive heart failure hospitalizations in the past 6 months, and history of end stage renal disease or dialysis in the past 12 months.

RESEARCH IN CONTEXT

  1. Systematic review: The authors reviewed publications on academic and socioeconomic impacts of attending an historically black college or university (HBCU) for Black Americans using common methods (e.g., PubMed, scholarly books). Most health research about HBCU attendance focused on sexual and mental health and no prior studies evaluated cognitive aging. We have appropriately cited the relevant literature.

  2. Interpretation: Black HBCU attendees (9%) averaged non-significantly better cognitive function at average age 70 compared to Black attendees at predominantly white colleges (91%), though confidence intervals were wide. HBCU attendees averaged slightly faster declines in verbal episodic memory. Effect estimates were imprecise and important potential benefits or harms cannot be ruled out.

  3. Future directions: Harmonized analyses using larger national studies are needed to evaluate the effects of HBCU attendance on later life cognitive outcomes. Cognitive aging research on life experiences that differentially affect Black adults should be prioritized to address disparities.

The STAR cohort includes community-dwelling African American adults residing in the San Francisco Bay area of California, primarily the cities of Oakland and Richmond. STAR aims to evaluate how life course vascular and sociocultural factors influence the trajectory of cognitive aging and burden of cognitive impairment among Black Americans. As with KHANDLE, individuals eligible for STAR were long-term members of KPNC and had previously participated in MHC exams between 1964 and 1985. Additional eligibility requirements were that members identified as Black and were age 50 years or older on January 1, 2018. Stratified random sampling by age and educational attainment was used with the goal of recruiting approximately equal proportions of participants ages 50 to 64 and 65 and older (range 53–90 years). The same exclusion criteria as in KHANDLE were applied to the STAR cohort.

At baseline, 1712 individuals were enrolled in KHANDLE and 764 were enrolled in STAR. We restricted the sample to 1197 participants (70%) who identified as Black or African American (N = 443 KHANDLE; N = 754 STAR), of whom 62% attended college (N = 737). Three waves of cognitive data were available for each cohort and follow-ups were conducted on average every 16 to 18 months in KHANDLE and 13 to 14 months in STAR. We excluded participants who were missing data on baseline cognitive scores, college attended, and our covariates of interest (3%), leaving 716 participants for analysis.

2.2 Study measures

2.2.1 College attendance

Participants self-reported the name of each college they ever attended at baseline. The college name of each participant was cross-referenced with the National Center for Education Statistics list of 101 HBCUs across the United States.17 A dichotomous variable was generated indicating if a participant ever attended an HBCU (non-HBCU/PWI = 0; HBCU = 1).

2.2.2 Cognitive outcomes

Outcomes included two cognitive domains assessed across all three study waves: executive functioning and verbal episodic memory (VEM). They were derived from the Spanish and English Neuropsychological Assessment Scales (SENAS), which was given to all participants at each study wave in their preferred language (English or Spanish). The SENAS battery of cognitive tests has previously undergone extensive development for valid comparisons of cognitive change across racial/ethnic and linguistically diverse groups. Executive function (EF) composite scores were obtained using component tasks of category fluency, phonemic (letter) fluency, and working memory (digit-span backward, visual-span backward, list sorting). VEM composite scores were derived from a multi-trial word-list-learning test. Details of the administration procedures, development, and psychometric characteristics have been extensively described in previous publications.34, 35 Each domain was z-standardized using the analytic sample baseline mean and standard deviation.

2.2.3 Covariates

We used a directed acyclic graph (DAG) to represent our hypothesized causal structure and select covariates expected to confound the association between the type of college one attends and late-life cognition (Figure S1 in supporting information).19, 21 Participants self-reported early-life experiences via interview survey at baseline. Childhood socioeconomic status (cSES) was a continuous seven-category composite measure defined as the sum of scores between responses from a three-category measure of family finance status (pretty well off financially = 1, about average or it varied = 2, poor = 3) and a five-category measure of hunger frequency from lack of money for food (never = 1, rarely = 2, sometimes = 4, often = 5) with higher scores reflecting lower cSES (range 2–8). Paternal and maternal education were assessed dichotomously (high school or less/missing = 0, any college or more = 1). Childhood academic support was a continuous nine-category composite measure defined as the sum of scores between responses from a five-category measure of the amount of time participants were encouraged to succeed in school (none = 1, little = 2, some = 3, most = 4, all = 5) and a five-category measure of received help with homework (none = 1, little = 2, some = 3, most = 4, all = 5) with higher scores reflecting greater support (range 2–10). Childhood academic discrimination was a dichotomous measure asking participants if they were ever unfairly discouraged by a teacher or advisor from continuing their education (no = 0, yes = 1). The remaining covariates included age, sex (female = 0, male = 1), study cohort (KHANDLE = 0, STAR = 1), and an indicator variable to account for whether cognitive assessment was implemented by phone (no = 0, yes = 1).

2.3 Statistical analysis

In addition to descriptive statistics for the study sample characteristics, we conducted tests of heterogeneity (e.g., chi-square) for differences between STAR and KHANDLE. We fit linear mixed-effects models using random intercepts and current age as the time scale, adjusted for age at cognitive assessment (mean-centered at age 70), sex, maternal and paternal education, childhood academic support, childhood academic discrimination, the study cohort, and interview method. Final models included an interaction between HBCU attendance and current age to evaluate differences in rate of cognitive decline and time-scale interactions with all covariates. Because cognitive test performance can improve when the same test is taken repeatedly,36 we accounted for practice effects by including an offset term in our models. Practice effects were estimated by fitting linear mixed-effects models using balanced pre-pandemic data (two waves) with random effects only, cognition at assessment as the outcome, current age as the time scale, and adjustment for sex, race, and education.37, 38 We conducted sensitivity analyses using a variety of time-scale specifications (i.e., baseline age, years since baseline, three-category study wave) to assess the robustness of findings. Finally, because STAR participants were younger at baseline than KHANDLE participants (aged 50+ vs. 65+) and thereby may have different predictors of cognitive aging, we included sensitivity analyses restricting to STAR participants age 65 or older.

3 RESULTS

3.1 Descriptive statistics

Table 1 shows the distribution of participant characteristics in the overall sample and by the type of college attended (HBCU or PWI). Among the overall sample of 716 Black participants, mean age at baseline and standard deviation was 70 years ± 8.4. Most participants attended a PWI (91%). Participants who attended PWIs were on average older, more likely to be STAR participants, more likely to report academic discrimination in childhood, but also more likely to report childhood academic support.

TABLE 1. Distribution of Northern California Kaiser Permanente member study sample characteristics by type of college attended.
Total (N = 716) HBCU (n = 63) PWI (n = 653)
Variable Level n (%) n (%) n (%)
Cohort KHANDLE 234 (32.7) 28 (44.4) 206 (31.6)
STAR 482 (67.3) 35 (55.6) 447 (68.5)
Sex Female 509 (71.1) 41 (65.1) 468 (71.7)
Male 207 (28.9) 22 (34.9) 185 (28.3)
Maternal education ≤HS/Missing 496 (69.3) 45 (71.4) 451 (69.1)
>HS 220 (30.7) 18 (28.6) 202 (30.9)
Paternal education ≤HS/Missing 564 (78.8) 48 (76.2) 516 (79.0)
>HS 152 (21.2) 15 (23.8) 137 (21.0)
Childhood academic No 457 (63.8) 50 (79.4) 407 (62.3)
Discrimination Yes 259 (36.2) 13 (20.6) 246 (37.7)
Childhood academic mean (sd) mean (sd) mean (sd)
Support Range 1–10 8.01 (2.3) 7.94 (2.3) 8.60 (1.6)
Childhood SES Range 1–8 3.39 (0.9) 3.38 (0.9) 3.46 (0.9)
Age at baseline Range 53–90 69.8 (8.4) 69.2 (8.2) 75.8 (8.2)
  • Note: Columns may not total 100% due to rounding.
  • Abbreviations: HBCU, historically black college and university; HS, high school diploma; KHANDLE, Kaiser Healthy Aging and Diverse Life Experience; PWI, predominantly White institution; sd, standard deviation; SES, socioeconomic status; STAR, Study of Healthy Aging in African Americans.

The distribution of participant characteristics in the overall sample by study cohort are reported in Table S1 in supporting information. On average, KHANDLE participants were older, more likely to attend an HBCU, and had lower cSES compared to STAR participants.

3.2 Linear mixed-effects models

Figure 1 shows the plotted standardized beta coefficient estimates for the association between HBCU attendance and cognition predicted at mean age 70, adjusted for covariates. HBCU attendees averaged better EF (βEF = 0.05 [−0.22, 0.32]) and VEM (βVEM = 0.21 [−0.06, 0.46]) compared to PWI attendees at baseline, but confidence intervals were wide and neither association was statistically significant.

Details are in the caption following the image
Mean differences in cognitive test scores. Adjusted linear mixed-effect regression estimates of cognitive outcomes on college attended among Black participants in Kaiser Healthy Aging and Diverse Life Experiences (KHANDLE) and Study of Healthy Aging in African Americans (STAR) (N = 716). Indicated along the x-axis, the standardized coefficients estimate the effect of historically black colleges and universities (HBCU) attendance on executive function (EF; purple) and verbal episodic memory (VEM; magenta). Each model includes covariate interactions with age, including an age x college interaction term testing the null hypothesis that the rate of change in cognition is the same for Black participants attending an HBCU and Black participants attending a predominantly White institution (PWI). The predictor indicated along the y-axis is the college effect at mean age 70.

Figure 2 shows the adjusted annual rate of change from mean age 70. HBCU attendees did not notably differ from PWI attendees with respect to annual rate of change in EF (βEF = −0.01 [−0.03, 0.02]), but averaged slightly faster decline in VEM (βVEM = −0.03 [−0.05, 0.00]). Model estimates are reported in Table S2 in supporting information. Similar results emerged in our sensitivity analyses using various time-scale interactions (Table S3 in supporting information); however, when restricting to STAR participants aged 65 or older at baseline, there were no rate differences in cognitive decline (Table S4 in supporting information).

Details are in the caption following the image
Rate of change in cognitive outcomes. Age by college attended interaction plots showing standardized adjusted rate differences in cognitive decline among Black participants in Kaiser Healthy Aging and Diverse Life Experiences (KHANDLE) and Study of Healthy Aging in African Americans (STAR) who attended an historically black college and university (HBCU; magenta) versus a predominantly White institution (PWI; purple) (N = 716). Mean age in years is indicated along the x-axis. Changes in executive function and verbal episodic memory are depicted on the left and right plots, respectively.

4 DISCUSSION

4.1 Summary of findings

In a study of 716 middle- to older-aged Black adults who attended college and currently live in Northern California, nearly 10% had attended an HBCU instead of a PWI. HBCU attendance was more common among younger individuals and among those with lower childhood academic support. We did not find evidence that HBCU attendance was associated with EF, VEM, or annual rate of change in EF, but confidence intervals were wide. We found a small, unexpected difference in annual rate of change in VEM such that HBCU attendees averaged slightly faster decline than PWI attendees. However, there was no difference in annual change in VEM when restricting the analyses to participants aged 65 or over at study baseline. Moreover, when using a three-category time-scale interaction for study wave instead of current age to relax a linear assumption, the accelerated rate was only observed between waves 1 and 2 and there was no difference between waves 1 and 3. It is plausible that the faster rate of decline is highly influenced by a few HBCU attendees with extreme values in the full cohort, suggesting that the significant association we detected was due to chance. Given this, study findings should be interpreted with careful consideration.

4.2 Limited HBCU and aging-related health literature

Despite estimate imprecision, our study is an important contribution to the literature. Substantial evidence shows that higher education reduces dementia risk, yet there has been little to no examination of whether, and to what extent, HBCU attendance impacts late-life cognitive aging in Black adults. Few studies have been able to tackle this question because of limited Black enrollment and lack of information on HBCU attendance. HBCUs offer Black individuals and Black communities increased self-, social, cultural, and economic empowerment,15, 16 likely contributing to Black HBCU attendees having better psychosocial health, perceived self-image, academic success, wages, and occupations than Black individuals that attend PWIs.21-23, 25, 27-29 Racial prejudice and discrimination are commonly reported by Black attendees of PWIs,39, 40 so evaluating HBCU attendance as a distinctive experience for older Black Americans—an experience that may offer the benefits of educational attainment without increasing exposure to interpersonal discrimination—is an important research priority. Our study is one of the first to quantitatively evaluate long-term health effects of HBCU attendance. Previous work has shown that attending an HBCU versus a PWI predicted lower risk of metabolic syndrome in mid-life Black adults, and that this effect was strongest for those who attended predominantly Black high schools and lived in predominantly Black communities,31 but no previous study has evaluated later life cognitive outcomes. Our study findings underscore the need for more nuanced data collection on race-based educational experiences to better understand Black–White inequities in cognitive aging across all levels of educational attainment.

4.3 Interpretation of findings

Given that structural racism is implicated as a fundamental cause of cognitive health,41 HBCU attendance may be a surrogate for other unmeasured racism-related confounders that we were not able to account for in the current study, contributing to imprecise estimates. For example, racial discrimination in the workplace is among the most salient forms of racism reported by mid-life Black women with higher socioeconomic status,42, 43 which has been linked to lower subjective cognitive functioning in Black women.44 Studies also show that Black HBCU alums can begin employment with higher wages21, 25 and better occupational status than Black PWI alums27 but then receive fewer wage increases and promotions over time,26, 27, 45, 46 with even fewer returns on HBCU benefits for Black women.22 For instance, one study showed no difference in initial wage or wage increases between Black women attending an HBCU versus a PWI.22 Other studies have found that higher performing Black high school students are more attracted to prestigious PWIs than HBCUs,21, 46 which was especially true after the passing of the 1964 Civil Rights Act, the era in which our study participants attended college.46 Also, during this era, racial segregation was omnipresent in the areas surrounding HBCUs, contributing to increased exposure to broader social and economic inequalities that disadvantaged Black students in multiple life domains (e.g., health, school quality, community resources).41 To elucidate causal mechanisms by which HBCU versus PWI attendance contribute to cognitive decline in Black adults over the life course, future research should include other aspects of HBCUs and PWIs (e.g., college prestige, location, quality) and life course exposures and experiences of their Black alums (e.g., cognitive ability, socioeconomic status, wealth and debt, perceived stress and coping).

Low prevalence (<10%) in our primary exposure of interest—Black adults attending an HBCU—also likely contributed to our non-significant findings by widening confidence intervals. Precision of an effect estimate is proportional to the variance of the exposure, thus fewer observations in the exposure group can decrease the statistical power required to detect an effect. Similarly, the annual rate of change estimates should be interpreted with caution given the limited statistical power. A cautious interpretation is further supported by our sensitivity analyses: non-significant age-restricted models further excluded 195 participants (27%) from our analytic sample. Because studies that collect population-level measures of dementia are comprised of mostly White adults, harmonized analyses using larger nationally representative studies are likely needed to avoid underestimating the health effects of HBCU attendance.

4.4 Strengths and limitations

The current study had other methodological considerations. Our participants were sampled from Northern California Kaiser Permanente members, restricting generalizability. Furthermore, as there are no established HBCUs in California, the exposure to risk factors for cognitive decline related to the type of college attended among our Northern California Black study participants plausibly varied in frequency, duration, and severity over time.47, 48 Lack of exposure randomization also limits causal inference; however, our longitudinal study design establishes temporality, which allows for the estimation of cognitive decline. Moreover, the follow-up assessment periods may not have been long enough to capture the full spectrum of age-related changes in cognition. Internal validity was strengthened by adjusting models for study factors and early-life socioeconomic and academic confounders that contribute to cognitive aging in Black Americans, although unmeasured confounding remains a concern. Given the current limitations of data sources, and the complex mechanisms by which structural racism operates as a social determinant of health over the life course, overcoming the issue of confounding to target key drivers of cognitive aging and related racial inequities is fraught with challenges. This project could not evaluate potential benefits of HBCUs on community opportunities, such as early-life exposure to Black cultural capital and Black excellence, but HBCU exposure is an important topic for research using national data sources. Despite this, our study is the first to evaluate HBCU attendance and cognitive outcomes among Black older adults, centering Black experiences and exploring sources of Black resilience. Like prior research applying causal frameworks to evaluate HBCU impacts on Black academic achievement, wages, and psychosocial health,22, 23, 25, 45, 49, 50 evaluating how HBCUs affect dementia risk requires implementing rigorous methods for assessing causal effects in the absence of formal randomization.

4.5 Conclusion

We found little evidence that HBCU attendance was associated with better cognition or slowed decline in EF or VEM among a sample of middle-aged and older Black adults. Effect estimates were imprecise, so substantial harm or benefits cannot be ruled out. Harmonized analyses with larger studies including other aspects of HBCUs (e.g., prestige, alumni socioeconomic status) are needed. Research on late-life health outcomes should incorporate race-based experiences that may differentially offer protective advantages to Black adults.

ACKNOWLEDGMENTS

Outside of this work, Dr. Mangurian is supported by several grants from the National Institutes of Health (NIMH, NIMHD, NIAID, NIDA), Department of Defense, Health Resources & Services Administration, the Doris Duke Charitable Foundation, the California Health Care Foundation, Genentech, and United Health Group. This work was supported by the National Institute on Aging (K99AG076973, NIGMS UL1GM118985 [PI: Thomas]; RF1AG050782 [PI: Whitmer]; R01AG066132 [PI: Gilsanz] 5R01AG052132 [PIs: Whitmer, Glymour, Gilsanz, Mayeda]). The funding sources did not have any role in study design; in data collection, analysis, and interpretation; in manuscript writing; or in the decision to submit the article for publication. The authors have no conflicts of interest to report.

    CONFLICT OF INTEREST STATEMENT

    The authors declare no conflicts of interest. Author disclosures are available in the supporting information.