This longitudinal, observational study assesses the physical, cognitive, and social health of older African American participants annually. Recruitment is continuous, with the first participants enrolled in 2000. Here we describe the sample and findings up to 2019. The protocol is approved by the OHSU Institutional Review Board (IRB #00001480) and all participants provided written informed consent prior to engaging in study procedures. Data is available from our local repository (#6845), and is submitted to the National Alzheimer's Coordinating Center Repository, which adds regional diversity to the national sample (https://naccdata.org/requesting-data/nacc-data).

Participants (n = 177) were recruited from the Portland, Oregon metropolitan area. To facilitate recruitment, the AADAPt research team collaborates with local community organizations, such as PreSERVE Coalition (https://www.preserve-coalition.org), a group of community members and academicians that promote brain health in the African American community. Churches, sororities, and health care systems also promote recruitment opportunities by providing space for information sessions, study presentations, and networking opportunities.

To be included in the study, participants need to self-identify as Black/African American, age 55 years or more. They cannot have a dementia diagnosis at study entry, and they need to able to participate in cognitive and functional testing. They must have a study partner to provide collateral historical data. Participants are encouraged to continue with the study for as long as they are able.

To encourage participation and reduce barriers to enrollment, transportation to research visits is provided free of charge. Participants are remunerated $30 per visit for their time and effort. Participants are thanked and updated on study findings at an annual brunch hosted by the research team in the community's neighborhood.

Enrolled participants attend clinic-based visits annually. The health history of participants, including family history of dementia, is documented. They are offered the option of assenting to brain donation after death for genetic and other analyses.

During the annual visits, a broad range of physical, cognitive, and social data is collected on each participant (see Table 1). The results presented here are from data collected at the baseline visit on 177 participants. These AADAPt participants had the option of receiving a brain MRI for volumetric analyses (Mueller et al., 1998), and blood samples for apolipoprotein E evaluation (Farrer et al., 1997).

Baseline descriptive findings of health-related risk factors (e.g., hypertension, diabetes), cognition, and social functioning are presented for the overall cohort. Additionally, comparisons were made between those classified as cognitively and functioning intact at baseline [i.e., Global Clinical Dementia Rating (CDR) = 0, Morris, 1993] and those with at least questionable cognition and/or daily functioning (i.e., CDR > 0) with a series of three MANCOVAs that controlled for group differences on age and education: (1) Daily functioning [i.e., basic and instrumental activities of daily living (Lawton and Brody, 1969; Fillenbaum, 1985)], (2) cognition [i.e., demographically-corrected z-scores on WRAT-R Reading, MMSE, Digit Span Forward and Backward, Category Fluency, CERAD Word List Acquisition and Delay Recall, Logical Memory I and II (Shirk et al., 2011; Weintraub et al., 2018, 2009)], and (3) physical functioning [i.e., CIRS (Parmelee et al., 1995), Tinetti Gait and Balance (Tinetti, 1986)]. For these MANCOVAs, follow-up ANCOVAs examined which measures within the domain were significantly different between the two groups. Finally, two ANCOVAs, which also controlled for age and education, compared the two groups on (1) depression [i.e., GDS/CESD (Radloff, 1977; Sheikh and Yesavage, 1986)] and (2) social activity [i.e., a composite measure of social activities (Morgan et al., 1985)]. Since data on this cohort was collected over a long period of time, this was considered as another possible covariate. However, there were not group differences (i.e., Global CDR = 0 vs. >0) on time of data collection (p = 0.08). Comparisons were considered statistically significantly different if p < 0.05. All analyses were performed using SPSS 29.0 software.

The sample included 177 participants who met the inclusion criteria and completed a baseline visit (Table 2). Most were cognitively intact at enrollment (76%, CDR = 0), and, based on responses from a collateral source, the participants were quite intact for basic and instrumental activities of daily living. Nearly half of the participants lived alone at the time of the baseline evaluation (45%), 37% lived with a spouse or partner, and 17% lived with a family member (e.g., children, grandchildren, and sibling). Most participants were born in the southern United States (70%) and moved to the Pacific Northwest in their youth. Over a third of the cohort (37%) was born in Texas or Louisiana.

The majority of the sample (76%) had a CDR global rating of 0.0, with 24% having a rating of 0.5 (and < 1% had a CDR of 1.0, but had a MMSE score that allowed inclusion). For the entire sample, 7% rated their memory as “excellent,” 59% as “good,” 29% as “fair,” and 5% as “poor” or “very poor.” Global cognition, as measured by the MMSE (Folstein et al., 1975), was consistent with subjective data, mean = 27.6 (SD 2.4), which falls above the conventional cutoff for dementia. Premorbid intellect was average (i.e., WRAT Reading), as was simple attention (i.e., Digit Span Forward), semantic fluency (i.e., Category Fluency), and verbal learning and memory (i.e., CERAD List Acquisition and Recall, Logical Memory I and II). Complex attention (i.e., Digit Span Backward) was low average.

Participants were optimistic about their physical health, with 71% rating their overall physical health as “good” or “excellent,” 26% viewed their health as “fair,” and 3% endorsed “poor” or “very poor.” Despite this optimistic view of their health, a sizable proportion reported current or past stroke or transient ischemic attack (26%), diabetes (33%), hyperlipidemia (54%) and hypertension (83%), and 50% reported ever smoking. Over half (58%) reported a family history of dementia. Even with these multiple vascular risk factors and dementia family history, baseline medical burden was comparable to another cohort of community dwelling older adults [Kaye et al., 2011; CIRS: mean = 21.4, SD = 3.9 (55th percentile of Kaye et al., 2011)]. Their balance and gait were also comparable to community dwelling older adults [Tinetti Gait: median = 0 (50th percentile of Kaye et al., 2011), Tinetti Balance median = 1.5 (~50th percentile of Kaye et al., 2011)].

In this cohort, 53% reported past or current depression. However, on screening measures of depressive symptoms (Table 1), minimal such symptoms were endorsed. Similarly, in a subset of this cohort, their collaterals endorsed minimal neuropsychiatric symptoms in them on the Neuropsychiatric Inventory total score = 0.23 (0.83) (Cummings et al., 1994).

Participants reported active social engagement. For example, the modal number of current close friends that participants reported was five. Most reported engaging in the following activities daily or weekly: attending church (76%), going out to eat (57%), and having visitors (54%). They followed the news, with 57% reading the newspaper or 92% listening to news on the radio/television on a daily basis. Over one-third of the cohort reported daily use of a computer. Social activity was also related to other outcomes. For example, using a composite of social activities greater social engagement was associated with better global cognition (MMSE: r = 0.24, p = 0.005), lower depression symptom scores (GDS/CESD: r = −0.21, p = 0.01), and lower medical comorbidity (CIRS: r = −0.21, p = 0.008).

In a subset of 88 participants who completed an MRI of the brain, the average hippocampal volume was 1.3 cc (SD = 0.2), and total brain volume = 908 cc (SD = 103), WMH 5.0 cc (SD = 6.3; Mueller et al., 1998). Apolipoprotein E was assessed in 121 of the AADAPt participants. Of these, 67.5% of the sample had no ϵ4 alleles, 30% of the sample had one ϵ4 allele, and 2.6% of the sample had two ϵ4 alleles (Farrer et al., 1997; Logue et al., 2023).

When the individuals with intact CDR scores (CDR = 0) were compared to those with questionable or more severe dementia (CDR > 0), the intact participants were younger (p < 0.001) and had more years of education (p < 0.001). The groups did not differ by gender (p = 0.48; Table 2). As such, age and education were used as covariates in these group comparisons. In an initial MANCOVA on daily functioning, the two groups differed on collateral ratings of their abilities to perform activities of daily living, with the CDR = 0 group performing significantly better than the CDR > 0 group [F_{(2, 145)} = 7.1, p = 0.001]. In follow-up ANCOVAs, this difference was present on both scales of basic (p < 0.001) and instrumental (p = 0.006) activities.

In a MANCOVA examining cognition, there were overall differences between these two groups [F_{(9, 120)} = 4.0, p < 0.001], with the CDR = 0 group performing better than the CDR > 0 group. In follow-up ANCOVAs, there were no group differences on premorbid intellect or simple and complex attention. Conversely, those rated as intact had significantly better global cognition (p < 0.001), semantic fluency (p = 0.047), list learning (p = 0.004) and recall (p < 0.001), and story learning (p = 0.01) and recall (p < 0.001) than those with impaired CDR scores. Such scores differences suggest that the impaired group presented with an amnestic cognitive profile that could indicate preclinical Alzheimer's disease.

In a MANCOVA on physical functioning, there were significant group differences [F_{(3, 144)} = 4.1, p = 0.008]. In follow-up ANCOVAs, those with CDR = 0 had fewer medical comorbidities (CIRS: p = 0.04), better balance (Tinnetti balance: p = 0.01), and better gait (Tinnetti gait: p = 0.003) compared to those with CDR > 0. In an ANCOVA examining responses on self-reported depression screenings, those with CDR=0 endorsed significantly fewer such symptoms compared to those with CDR > 0 [F_{(1, 170)} = 28.5, p = 0.003]. Finally, in an ANCOVA using a composite of social activities, individuals rated as CDR > 0 engaged in significantly fewer social activities than those rated as CDR = 0 [F_{(1, 150)} = 39.3, p = 0.03].

AADAPt participants were recruited from the metropolitan Portland area, which limits the geographic generalizability of these findings. Participants were also self-selected, which could create some bias in this sample. Our sample has limited representation from those with lower educational levels.