Prevalence of Dementia in American Indians and Alaska Natives Compared to White, Black, and Hispanic Medicare Beneficiaries: Findings from the National Health and Aging Trends Study

Abstract

Objective

To estimate the prevalence of dementia among American Indian and Alaska Native (AIAN) Medicare beneficiaries and compare the prevalence of dementia in AIANs and other racial and ethnic minority groups—non-Hispanic Blacks (NHBs) and Hispanics—with non-Hispanic Whites (NHWs) Medicare beneficiaries.

Methods

We used survey data from Round 5 of the National Health and Aging Trends Study (NHATS, 2015) (N = 7,449), a nationally representative study of Medicare beneficiaries ages 65 years and older. We used logistic regression to estimate the age- and sex-adjusted prevalence of dementia with 95% confidence intervals by race/ethnicity as well as prevalence differences between groups.

Results

The majority of participants were between 65 and 74 years old. Approximately half of them were female. NHWs had a prevalence of 5% for dementia, and AIAIs had a prevalence of 9%, four percentage points higher than NHWs (95%CI 1%, 11%). We estimated a similar difference in the prevalence of dementia in AIAN Medicare beneficiaries compared to NHBs.

Conclusion

While previous research has reported that AIANs share a similar or lower prevalence of dementia than NHWs, our findings suggest a potentially higher prevalence of dementia among AIAN Medicare beneficiaries. A relatively small number of AIAN resulted in wide confidence intervals for many of our associations. Future research should focus on recruiting a larger sample and on dementia prevalence and unique risk factors within and among AIAN tribes.

Introduction

Alzheimer’s disease (AD) is the most common cause of dementia and the fifth leading cause of death for older adults in the USA, with shortened life expectancy compared with non-AD older adults [1]. In general, AD is characterized by a progressive, significant deterioration in cognitive and physical abilities, leading to dependence on caregivers. The national healthcare and long-term care costs for AD and related dementias (ADRD) are projected to grow from $355 billion in 2021 to $1.1 trillion by 2050 [1]. An additional estimated annual cost of $256.7 billion is estimated when unpaid care by informal caregivers is considered.

While some studies observed a significant decline in dementia prevalence between 2000 and 2012 [1, 2], the estimated number of people with AD and other dementias in the USA is predicted to be more than double (12.7 million) by 2050 due to demographic factors such as population growth and aging [1, 3], although disease estimates are largely based on studies of non-Hispanic white (NHW) older adults. Despite years of focused attention, there is little information on AD and other dementia estimates among American Indians and Alaska Natives (AIANs). In 2010, there were approximately 235,000 AIANs aged 65 and older, but this population is expected to increase to more than 918,000 by 2050 due in part to improvements in socio-economic and environmental factors [4–8]. With this increase in life expectancy, the number of AIANs with ADRD is expected to increase from 23,850 in 2010 to over 100,000 by 2050 [2]. However, there is a dearth of studies on the AIAN population, with even simple ADRD prevalence estimates lacking in the literature [3, 9–12].

This lack of information on dementia in AIAN older adults has significantly hampered the development of effective, culturally responsive protocols for prevention, detection, identification, and intervention, leaving the AIAN population at a greater risk of poorer outcomes. Without a basic understanding of the prevalence of dementia in this group, we will remain limited in our ability to create appropriate intervention strategies for AIAN older adults. Existing studies of dementia in the AIAN population have reported mixed results regarding levels of prevalence of ADRD among these groups [13–17]. For example, a study of the Cherokee of Oklahoma found that as “the genetic degree of Cherokee Indian ancestry increased,” the risk of developing AD decreased [17]. However, a recent study in Kaiser Permanente in Northern California reported that the cumulative 25-year risk for dementia at age 65 years was highest for Black (38%), followed by AIAN (35%), Latino (32%), Hawaiian/Pacific Islander (25%), White (30%), and Asian American (28%) populations [10]. AIANs are at an elevated risk for many chronic health conditions, including diabetes, cardiovascular disease, and traumatic brain injury, which are established risk factors for dementia [16, 18]. Given the projected growth of the older population in AIANs, more research is needed on the prevalence of dementia and cognitive impairment using validated assessment instruments, robust methodology, and nationally representative data [11, 18].

To address this research gap, the current study used data from the National Health and Aging Trends Study (NHATS) to estimate the prevalence of dementia among AIAN Medicare beneficiaries. We compared the prevalence of dementia in AIANs and other racial and ethnic minority groups—non-Hispanic Blacks (NHBs) and Hispanics—with non-Hispanic Whites (NHWs). We also compared the prevalence of dementia in AIANs to NHBs and Hispanics.

Methods

Sample

We used survey data from Round 5 of the NHATS (2015). The NHATS is a nationally representative study of Medicare beneficiaries 65 years and older that collects data annually from beneficiaries on age-related changes in “physical and cognitive capacity, how activities of daily life are carried out, the social, physical, and technological environment, and participation in valued activities” [19]. NHB and older respondents were oversampled in the NHATS survey [20]. The NHATS dataset included 8,038 Medicare beneficiaries. For this study, we excluded 403 who were nursing home residents and 136 participants in other residential facilities who did not complete the survey in 2015. We included community-dwelling participants and non-nursing home residents (e.g., assisted living facilities) who completed the survey (N = 7,499). Details about the sampling design, sample weights, and design variables have been previously published [21]. The current study was approved by the University of Louisville Institutional Review Board (20.111).

Measures

NHATS assesses the cognitive status of participants using the following methods: proxy reports (proxy respondents are used if the participants cannot respond) or self-reports indicating a doctor has told the participant they have dementia or AD; a proxy-reported assessment of dementia using AD8 Dementia Screening Interview [22, 23]; and a battery of cognitive tests, administered to the participant, designed to evaluate memory (immediate and delayed 10-word recall), orientation (date, month, year, and day of the week; naming the president and vice president), and executive function (clock drawing test).

NHATS identifies three outcomes of these cognitive tests: probable dementia, possible dementia, and no dementia. Probable dementia is defined by the presence of any one of the following: (1) a self- or proxy report of a diagnosis of dementia, (2) a proxy-reported score on the AD8 of ≥ 2, or (3) a score of ≤ 1.5 standard deviations from the mean on assessments of at least two of the three cognitive functioning domains tested (orientation, memory, and executive functioning). Possible dementia was defined by a score of ≤ 1.5 standard deviations from the mean on at least one of the three cognitive functioning domains (memory, orientation, and executive function). According to NHATS documentation [19], the narrower criteria of probable dementia had a specificity of 87% and a sensitivity of 61.6%, and the combined criteria for both probable and possible dementia showed a specificity of 61.6% and a sensitivity of 85.7% when evaluated in the Aging, Demographics, and Memory Study. We used the narrower, more specific definition of probable dementia as our primary outcome but also included results using the broad definition of probable or possible dementia [19.24].

Independent Variables

Race/Ethnicity

NHATS respondents were asked to report all racial/ethnic identities that applied. “What race do you consider yourself to be: White, Black or African American, American Indian, Alaska Native, Asian, Native Hawaiian, Pacific Islander, or something else.” If respondents identified selected multiple races, they were also asked to state the race they consider their primary racial identity (“Do you consider yourself primarily_____ ?”). Participants were asked whether they consider themselves Hispanic or Latino. If the respondents were endorsing Hispanic or Latino ethnicity, they were classified as that to the exclusion of any other racial identity they might have endorsed. We categorized those who reported being only AIAN or primarily identifying as AIAN as “primary-race AIANs” (unweighted sample size = 68, weighted sample size = 393,141). Participants who identified AIAN as one of multiple racial/ethnic backgrounds but did not indicate AIAN as primary were categorized as a “non-primary-race AIAN” (n = 154). Our analysis focused on primary race AIANs but included data from the non-primary AIAN group for supplementary analyses.

Demographics

We considered six pre-defined age groups by NHATS (65–69, 70–74, 75–79, 80 – 84, 85–89, 90 + years) and sex (male, female). Additional information was available on marital status (married, not married), living arrangement (living alone, living with others), educational attainment (high school diploma or less, more than a high school diploma), and self-reported presence of specific health conditions (heart disease, heart attack, high blood pressure, diabetes, and stroke).

Analytic Strategy

We present the estimated percentages of demographics and chronic health conditions by race/ethnicity with 95% confidence intervals. Logistic regression models were fit to the NHATS data with dementia as the outcome and categorical race/ethnicity variable as the predictor of interest. Separate regressions were fit using the narrow and broad definitions of dementia. We fit unadjusted and adjusted logistic regression models. Adjusted models include the non-modifiable risk factors such as age and sex [24]. Regression results were used to estimate the predicted prevalence of dementia in each racial and ethnic group, and the absolute differences in prevalences between groups averaged over other variables in the model [25]. We present the difference in the predicted prevalence of AIANs, NHBs, and Hispanics compared to NHWs. We do not present the results of the remaining group, including Asians, Native Hawaiians, Pacific Islanders, or “other,” because the subgroups are heterogeneous. We do not report analytic results using null hypothesis significance testing. Rather, we focus on estimated prevalences or associations along with estimates of precision (95% confidence intervals). This is in keeping with good practice as advocated by leading epidemiology textbooks, the American Statistical Association, and a statement in Nature signed by over 800 scientists [26–28]. We used Stata version 15 together with the NHATS analytic sampling weights to incorporate the NHATS complex sampling strategies. We also present the predicted prevalence using data on non-primary race AIAN Medicare beneficiaries as supplementary analyses.

Results

Characteristics of the Study Sample

The majority of participants in each racial/ethnic group were between 65 and 74 years old. Slightly more than half were female (Table 1). AIANs were less likely to live alone than NHWs and NHBs, and our additional comparisons indicated that Hispanics were less likely to live alone than NHWs and NHBs. Approximately 66% of AIANs reported having a high school education or less, compared to 39% of NHWs, 57 of NHBs, and 74% of Hispanics. AIANs (55%) were approximately as likely to be married as NHWs (58%), but more than NHBs (35%). NHBs and Hispanics showed less educational attainment and were less likely to be married than NHWs.

Table 1.

Characteristics of Medicare beneficiaries by race/ethnicity (NHATS 2015, N = 7,499, weighted; 95% CI)^*

Variables	AIAN (n = 68)	NHW (n = 5,136)	NHB (n = 1,546)	Hispanic (n = 441)
Age (years)
65–69	44.9 (32.4–58.2)	37.8 (35.0–37.6)	39.5 (36.5–40.6)	40.6 (34.3–46.2))
70–74	21.1 (13.0–29.4)	23.2 (21.4–23.35)	24.8 (22.3–26.6)	25.2 (20.9–29.6)
75–79	13.4 (9.0–20.8)	17.3 (16–18)	17.7 (15.9–19.7)	15.8 (12.8–19.7)
80–84	9.9 (5.6–16.3)	12.0 (11–13)	10.0 (9.2–11.6)	11.1 (8.6–14.4)
85–89	10.6 (5.1–20.3)	7.1 (6.6–7.7)	5.2 (5.1–6.8)	4.9 (3.7–7.4)
90 +	0.9 (.2–3.1)	3.6 (3.4–4.1)	2.6 (2.6–3.9)	2.4 (1.6–3.9)
Female	54.7 (39.4–67.4)	55.2 (53.9–56.4)	59.7 (57.3–62.5)	56.3 (51.6–61.4)
Living alone	26.9 (15.6–42.2)	30.8 (29–32.6)	35.9 (33.3–38.7)	21.0 (17.8–24.7)
Married	54.8 (42.7–66.4)	58.3 (56.660)	35.2 (32.9–37.5)	50.3 (45.3–55.1)
Education	66.3 (51–79.7)	39.0 (36.2–41.5)	57.2 (52.8–62.2)	73.7 (69.–78.3)
≤High school diploma
Heart attack	32.2 (20.8–41)	12.6 (11.7–13.7)	12.9 (11–15.3)	12.1 (9.5–15.8)
Heart disease	21.7 (11.9–36.4)	16.1 (15–17.7)	14.3 (12.3–16.6)	12.1 (9.1–16.5)
High blood pressure	83.6 (72.16–91)	61.6 (60.3–63.4)	76.5 (73.7–79.2)	67.9 (63.3–72.2)
Diabetes	34.6 (23.3–48)	22.6 (21.6–24.2)	36.2 (33.4–39)	41.6 (34.6–49.1)
Stroke	14.1 (8.3–23)	5.8 (5.2–6.7)	7.9 (6.4–9.5)	5.3 (3.5–8)

^*mean or percent and 95% confidence interval(CI); characteristics of others are not present

A large proportion of AIANs reported having chronic conditions. Compared to NHWs, AIANs reported more heart attacks (32% vs. 13%), diabetes (35% vs. 23%), and higher blood pressure (84% vs. 62%). However, AIANs reported a lower prevalence of diabetes than NHBs (36%) and Hispanics (42%). NHBs were more likely to report more heart attacks than AIANs and NHWs. Hispanics were less likely to report heart attacks than both NHWs and NHBs. Also, Hispanics were more likely to report high blood pressure and stroke than NHWs. We note that the modest sample size among AIANs makes many of these prevalence estimates somewhat imprecise (see Table 1).

Dementia Prevalence

NHWs had a prevalence of 5% for dementia (95% CI: 4%, 6%), and AIAIs had a prevalence of 9% (95% CI: 1%, 16%), four percentage points greater than NHWs (95% CI: 1%, 11%) Compared to NHWs, in a model adjusting for age and sex, prevalence of probable dementia was 5 percentage points greater among NHBs (prevalence difference (PD) = 5%; 95% CI: 3%, 7%) and 8 percentage points greater among Hispanics (PD = 8%; 95% CI: 5%, 12%). After controlling for age and sex, the prevalence and prevalence differences remained largely unchanged in all groups.

When we examined the broader range encompassing both probable and possible dementia (Table 2), we found larger differences in dementia prevalence between NHWs and other racial/ethnic groups. Prevalence of probable and possible dementia was 23% among AIANs (11 percentage points higher than NHWs), 25% among NHBs (13 percentage points higher than NHWs), and 28% among Hispanics (15 percentage points more than NHWs). The supplementary tables present the predicted probabilities of both probable dementia and probable and possible dementia for the non-primary race AIAN category. While these other results were similar to findings from our main analyses, the difference between the prevalence of probable and possible dementia in the non-primary race AIANs and NHWs was far less pronounced (see Tables 3 and 4 for supplementary data on non-primary race AIAN Medicare beneficiaries).

Table 2.

Predicted Prevalence of ADRD, NHATS 2015 (N = 7,449, weighted, 95% CI)

	Probable dementia			Probable and possible dementia
	Model 1	Model 2*	Difference in predicted prevalences*	Model 1	Model 2*	Difference in predicted prevalences*
NHW	0.05 (0.04–0.06)	0.05 (0.04–0.06)	Ref	0.12 (0.11–0.14)	0.12 (0.11–0.13)	Ref
AIAN	0.09 (0.01–0.16)	0.09 (0.02–0.16)	0.04 (0.01–0.11)	0.23 (0.12–0.34)	0.23 (0.13–0.34)	0.11 (0.01–0.22)
NHB	0.09 (0.08–0.11)	0.10 (0.08–0.11)	0.05 (0.03–0.07)	0.24 (0.22–0.26)	0.25 (0.23–0.27)	0.13 (0.10–0.15)
Hispanic	0.12 (0.09–0.16)	0.13 (0.10–0.17)	0.08 (0.05–0.12)	0.27 (0.21–0.31)	0.28(0.23–0.32)	0.15 (0.11–0.20)

^*adjusted for age and gender (all significant); weighted sample

Discussion

The older AIAN population is one of the fastest-growing groups in the USA [2, 5]. As a result, the prevalence of dementia is expected to increase as the population ages. Our results suggest that AIANs may be at a higher risk, though there is considerable imprecision in our estimates. Relative to NHWs, for whom we estimate a dementia prevalence of 5% (95% CI: 4%, 6%), AIANs have a 9% prevalence (95% CI: 2%, 16%). This four percentage point increase (95% CI: 1%, 16%) was somewhat imprecise given the small sample size. These data are compatible with AIANs having a modest one percentage point increase in dementia prevalence as well as a substantial 11 percentage point increase in dementia prevalence, given no errors other than random chance. This was the case for both the narrow and broad definitions of dementia; however, the difference in prevalence estimates was more pronounced with the broad, less specific definition. We included not only those who had dementia but also persons with cognitive impairment indicating possible dementia, which may impact the noticeable prevalence estimates.

Our results are consistent with the findings from previous studies [10, 29] showing that AIAN Medicare beneficiaries may be at a higher risk of dementia than NHWs. Although a strength of the current study is that it is based on national data, disparities throughout the life course that disproportionately impact AIANs (e.g., pre-and post-natal complications, healthcare differences, malnutrition, exposure to environmental toxins and other stressors, risk of traumatic brain injury, hypertension, stroke, and comorbid conditions) [30] are likely the cause of the increased risk of dementia [18], although we did not include these factors. We also found higher proportions of health conditions associated with dementia risk, such as high blood pressure, in AIANs than in NHWs [31]. AIANs were more likely to have lower formal educational attainment than NHWs, a potential risk factor for dementia [31, 32]. Possibly, tobacco use, physical inactivity, and malnutrition (e.g., lower level of fruit and vegetable consumption) among AIANs might also contribute to increased risk of dementia [33]. Future researchers should investigate a unique multi-domain pathway to dementia risk in AIANs and focus on decreasing the prevalence of identified risk factors in AIAN populations.

Our study found that AIANs have a similar prevalence of dementia as non-Hispanic Black older adults. This finding may reflect the consequences of inequities affecting these groups, similarly, leading to SES disparities and chronic conditions that increase the risk of dementia onset. Our results are similar to other national studies showing a higher prevalence of dementia among older adults of Hispanic ethnicity than among those who are non-Hispanic Whites and non-Hispanic Blacks.³⁴ Our finding that AIANs have a lower prevalence of dementia than Hispanics adds to the literature specific to the AIAN population, although we did not conduct a statistical comparison and observed large differences in the sample sizes. Reasons for these differences may be due to the use of national data with representation from across the country, which includes different healthcare systems and different environmental stressors, as well as the limited sample size of the AIAN group. Future research should investigate potential variations by region and race/ethnicity.

There are several limitations to this study. First, the small sample size is an important limitation. The unweighted sample size of AIANs was 68 (weighted AIAN sample size = 393,141) in the current study, resulting in wide confidence intervals and modest precision for our estimates; future national studies should oversample the AIAN population to gain greater precision in population estimates of the dementia prevalence. Second, NHATS participants were recruited from Medicare beneficiaries. Although more than 90% of AIANs age 65 and older list Medicare as a source of health insurance coverage, those who do not receive Medicare (and are thus not included in the NHATS sample) may be different than those who do, resulting in selection bias [34]. Furthermore, the average life expectancy for AIANs is shorter than that of other racial/ethnic groups. For example, AIANs born in 2020 have a life expectancy of 73 years, 5.5 years shorter than that of the overall US population (78 years) [5]. Also, due to a high prevalence of well-known AD risk factors (e.g., limited physical activity, hypertension, type 2 diabetes, high cholesterol) among AIANs, they might have AD earlier than other racial and ethnic groups [1]. Additionally, we did not have any available on tribal affiliation, reservation dwelling, or degree of native ancestry, allowing the investigation of unique factors that might impact dementia risks. Finally, the NHATS dementia classification is not based on clinical adjudication of diagnosis and reflects the lower sensitivity of its narrow definition and the lower specificity of its broad definition.

Despite these limitations, including wide confidence intervals due to the small sample size of AIANs, our results contribute to the limited information on the prevalence of dementia among AIANs in the USA. Future research should focus on recruiting a larger sample, allowing for more precise estimates and investigating dementia prevalence and risks by different factors such as AIAN tribal differences in culture and traditions or urban compared to rural settings to provide culturally tailored interventions to reduce the risks in this population. In addition, the continued projected growth in the older AIAN population may also mean that older AIANs need culturally competent health services through Indian Health Service or other appropriate systems, including those specifically intended for dementia diagnosis and care.