A systematic review/meta‐analysis of prevalence and incidence rates illustrates systemic underrepresentation of individuals racialized as Asian and/or Asian‐American in ADRD research

We investigate Alzheimer's disease and related dementia (ADRD) prevalence, incidence rate, and risk factors in individuals racialized as Asian and/or Asian‐American and assess sample representation. Prevalence, incidence rate, risk factors, and heterogeneity of samples were assessed. Random‐effects meta‐analysis was conducted, generating pooled estimates. Of 920 records across 14 databases, 45 studies were included. Individuals racialized as Asian and/or Asian‐American were mainly from Eastern and Southern Asia, had higher education, and constituted a smaller sample relative to non‐Hispanic white cohorts. The average prevalence was 10.9%, ranging from 0.4% to 46%. The average incidence rate was 20.03 (12.01‐33.8) per 1000 person‐years with a range of 75.19–13.59 (12.89‐14.33). Risk factors included physiological, genetic, psychological, behavioral, and social factors. This review underscores the systemic underrepresentation of individuals racialized as Asian and/or Asian‐American in ADRD research and the need for inclusive approaches accounting for culture, language, and immigration status.


INTRODUCTION
Dementia prevalence will experience exponential growth worldwide over the next three decades. 1Alzheimer's disease (AD) is the most common dementia etiology, followed by vascular, Lewy body, frontotemporal, and Parkinson's disease.In 2023, 6.7 million Americans ≥ 65 years of age (an older adult) were living with AD in the United States (US), equating to roughly 1 in 9 people (10.7%) in this age band. 2 AD is expected to rise to 12.7 million by 2050, almost doubling from the 2023 estimate. 2Racial and ethnic disparities in dementia incidence are pervasive and embedded in structural and social determinants of health (S/SDOH). 3,45][6] Institutional racism extends to scientific practices that aggregate distinct national and cultural groups into racialized categories that reflect external perceptions of commonalities.Existing dementia studies often aggregate data for individuals racialized as Asian, Asian-American, and Native 8][9] The aggregation of diverse groups of people into an "Asian" category results in findings that suggest such a group has a decreased risk for dementia than other groups, which may lead to biases that have significant ramifications for certain groups and individuals. 7aracterizing an individual as "Asian" and/or "Asian-American", is a part of racialization, a process that assigns a category and value to individuals based on perceived similarities. 10This label is a social construct that is highly contextual and subject to change based on time and space."Asian" refers to a nebulous racialized category used to denote the identity of diverse peoples that currently constitute 60% of the global population and represent over 2,300 distinct languages.In the United States, "Asian" is used as a racialized category by the Office of Management and Budget (OMB) and Census to distinguish individuals of perceived or self-reported Asian ancestry, including people from the Far East, Southeast Asia, and the Indian subcontinents while excluding persons from the Middle East, West Asia, and the Caucasus region. 11This arbitrary categorization diminishes the complexity of nationality, history, and cultural backgrounds.Interestingly, this racialization is assigned to groups irrespective of country of origin as future generations of "Asian" descendants are racialized similarly to someone born in an "Asian" country.On occasion, these individuals may be referred to as "Asian-American" to acknowledge their affiliation with the United States and to demarcate their connection to the Asian continent.Despite individuals racialized as Asian and/or Asian-American in the United States being considered a smaller portion of the population, they represent one of the fastest-growing racialized groups in the United States, and older adults in this group are projected to grow 145% by 2030. 12Globally, the number of individuals racialized as Asian (based on OMB definition) is projected to grow to 5.25 billion by 2055. 13However, individuals racialized as Asians in the United States often experience exclusion by being grouped into a monolith in scientific research obfuscating an understanding of which specific risk and protective factors impact the incidence and prevalence of Alzheimer's disease and related dementia (ADRD).
5][16] When integrated with individuals racialized as White, individuals racialized as Asian and/or Asian-American experience a myriad of interpersonal discrimination, including microaggressions, some leading to more rapid cognitive decline. 17Most notably, the perceived socioeconomic and integration successes of a few have fueled the construction of the "model minority myth" that erroneously posits that individuals racialized as Asian and/or Asian-American have achieved socioeconomic and social success, yielding benefits to health and well-being in the United States.Therefore, these individuals are not underrepresented, do not experience health disparities, and should be excluded from social and healthcare policy reforms. 18This myth has led to a lack of standards of care for individuals racialized as Asian and/or Asian-American, resulting in systemic underestimation of ADRD prevalence.Of the limited studies that include individuals racialized as Asian and/or Asian-American, estimates of AD or ADRD prevalence are often inconsistent due to variations in data collection of race and ethnicity, cognitive and clinical assessments used, and statistical methods reported.For example, factors such as the inclusion of resident location (community vs. institution), 19 choices of clinical assessments or neurological batteries used for dementia diagnosis (many of which are not accessible for people who speak an "Asian" language), 7 or specifying an Asian nationality under consideration 19 all contribute to the uncertainty of prevalence rates among individuals racialized as Asian and/or Asian-American.There is an underestimation of the various etiologies ensconced under the umbrella of ADRD, which are not mutually exclusive, as mixed dementia tends to be more prevalent in minoritized groups compared to older adults racialized as non-Hispanic White (nHW) individuals. 20,21ven these limitations in understanding the prevalence and potential risk factors for ADRD among older adults racialized as Asian and/or Asian-American, these populations face multiple risks, including underdetection, 22 delayed diagnosis, 7,19 and suboptimal management of cognitive impairment.For example, a recent study of racial and ethnic differences in ADRD diagnosis rates found that Medicare beneficiaries racialized as Asian and/or Asian-American received fewer diagnostic evaluations for dementia compared to beneficiaries racialized as nHW individuals. 23A growing body of literature is finding that individuals racialized as Asian and/or Asian-American face significant barriers to healthcare access including language and communication issues, inequitable cultural competency, and less/poorer insurance coverage. 24,25Furthermore, the spotlight is shifting to the role of immigration as a risk factor for ADRD.Given that immigrants from Asia make up the largest immigrant group in the United States, 26  (3) conduct an analysis on the risk factors associated with dementia for adults racialized as Asian and/or Asian-American identified in the existing literature.

Literature Search Strategy
The published literature was systematically searched on October

Data Extraction and Synthesis
Search results were uploaded to Covidence 27 for further screening.To obtain prevalence and incidence rates for the meta-analysis, the number of ADRD cases, incidences, and person-years for both groups racialized as Asian and/or Asian-American and comparison groups were extracted from the datasets used in the studies.In the event that multiple studies utilized the same dataset/cohort study, the prevalence or incidence rate was extracted from the articles specifically reporting on either rate or both for that study.
To analyze the risk factors associated with ADRD among adults racialized as Asian and/or Asian-American, the identified risk factors (explanatory or dependent variables) that showed significant associations with ADRD in each study were extracted.These factors were then categorized into physiological, genetic, psychological, behavioral, and social categories.

Statistical Analysis
Following data abstraction, the heterogeneity between studies was assessed using the Cochran Q statistic, and its magnitude was quantified using the I 2 statistic.An I 2 value greater than 50% indicates significant heterogeneity. 23In the presence of heterogeneity, random effects meta-analysis was conducted to obtain pooled estimates of the prevalence and incidence rate of ADRD.Forest plots were utilized to visually represent the heterogeneity between different studies and display the pooled estimates.The pooled estimates and forest plots were performed using R version 4.3 using the Metafor package. 29

RESULTS
There were 45 publications that examined data from fourteen unique datasets (Figure 1).Among the 45 studies, 26 studies used ADRD incidence rate as an outcome, which described the development of ADRD over a specific period, and 17 studies used ADRD prevalence, which reported the occurrence of ADRD at one-time point.One study reported incidence diagnosis and one study reported both prevalence and incidence rate.A summary of the demographic characteristics of samples racialized as Asian and/or Asian-American included in these datasets is provided in Table 1.Among these studies, 15 utilized datasets incorporating ADRD data from multiple racial and ethnic groups 7,22,23,[30][31][32][33][34][35][36][37][38][39][40] nationwide or in certain regions.8][39] Three small-scale studies [40][41][42] recruited participants from multiple racial and ethnic groups from institutions or communities.

Individuals racialized as Asian and/or Asian-American
Operationalization of race and ethnicity varied across studies.
Of 15 studies that focused on racial comparisons 7,8,22,23,30-42 four studies 7,22,30,31 23 The remaining studies relied on community partners to assist with participants' racial and ethnic identity; however, most studies did not utilize nationality for their definition.Only Graves et al. (1996)  43 specifically delineated Japanese Americans in the KAME project as a person with at least 50% of their ancestors identifying as being from Japan.Since different studies used varying terminology to describe race and ethnicity, the terminology reported in this section reflects how the original studies operationalized Asian and/or Asian-American as a racialized group.Veterans Health Administration dataset. 36presentation was relatively higher in regional large-scale datasets like Kaiser Permanente Northern California and Multiethnic

TA B L E 1 Key demographics of Asian or
Cohort Study since they are based in California, where populations racialized as Asian and/or Asian-American represented 17% of the population in 2022. 44 participants who immigrated from China, Japan, and the Philippines, accounted for over 50% of Asian and/or Asian-American, while less than 5% of participants identified as immigrants or descendants of ancestors from the rest of Asian countries combined. 8,33,37Three studies in New York City and one in San Francisco centered on the experiences of Chinese Americans.Of these studies, two recruited participants from nursing homes, 35,40 and the other two studies recruited patients from medical schools. 41Finally, the Honolulu-Asia Aging Study cohort located in Honolulu, Hawaii, and The Kame Project located in Seattle, Washington State, recruited Japanese American patients. 12,43

ADRD Assessment
The operationalization of ADRD varied among the 45 studies (Table 2).
Among studies which collected primary data, Fan et al. (2019) 41  (2011) 45,46 after conducting a cognitive assessment battery adopted for the National Alzheimer's Coordinating Center Uniform Data Set (UDS). 47  cross-sectional datasets (Table 3).Six of the studies  41 cognitive tests, 42 and healthcare service utilization tend to have a higher proportion 32 of participants diagnosed with ADRD.
Twenty-two studies used the incidence rate from longitudinal studies of ADRD as the outcome variable (Table 4).Among these, 12 studies were based on the Honolulu-Asia Aging Study cohort.2000) 63 for the meta-analysis.Six studies 7,33,34,[37][38][39]   reported a higher hazard for Asians compared to nHW individuals; two studies 7,8 reported a lower hazard rate for Asians for ADRD, and a fourth study 30 did not find a statistically significant difference in hazard between Asians and nHW individuals.

Risk factors associated with ADRD
5][56][57][58][59][60][61][62][63][64][65][66][67][68][69][70][71][72][73] The significant risk factors are organized in Studies have shown that immigration status and language may also be important risk factors for ADRD but are significantly understudied. 4ong the studies identified in this review, most studies were not able to collect information on the immigration status or primary language used by the participants.Hayes-Larson et al. (2022) 33 found that ADRD incidence rate appeared higher among the foreign-born Filipinos and Japanese, but not among Chinese and individuals racialized as nHW individuals.Another investigation 59 found that there was not a significant association between written Japanese proficiency and ADRD.

DISCUSSION
To our knowledge, this systematic review and meta-analysis is the first to examine the evidence from empirical studies, quantify heterogeneities of prevalence and incidence rate, and elucidate the challenges pertaining to the sample representation, diagnosis criteria, and risk factors across the studies involving older adults racialized as Asian and/or Asian-American in the United States.Lim et al. (2020) 78 conducted the first scoping review of ADRD in the AAPI population to map the extant literature.This systematic review and meta-analysis further evaluates empirical evidence and found that the pooled effects estimated by random effects models were 12.90% (95% CI: 4.4%−25%) for prevalence and 20.03% (95% CI 12.01-33.38)for incidence rates.There was significant heterogeneity (I 2 ) due to the different inclusion criteria of Asian groups and measures of ADRD.Most research has concentrated on physiological and genetic factors as the principal risk factors.In contrast, psychological, behavioral, and social factors have not been explored as extensively.Additionally, pathology data in ADRD are often lacking for minoritized groups, especially for Asian Americans. 79The disparity between clinical and pathological diagnosis of ADRD was highlighted in seven studies included in the present review, but all the data used in these studies were generated by the Honolulu-Asia Aging Study cohort.
There is a profound gap in the ADRD literature on prevalence and incidence rates among United States adults racialized as Asian and/or Asian-American. 8 Asian and/or as American make up less than 3% of the cohorts, and most community studies are only able to work with a small sample of participants from one given Asian group.The overall lack of diversity in the available data, coupled with a selection bias in each dataset, constitutes an important limitation in our capacity to effectively measure ADRD prevalence and incidence among adults racialized as Asian and/or Asian-American. 78,80r findings suggest that aggregating diverse and heterogeneous populations into a monolithic "Asian" category masks significant differences between groups and among individuals. 81In regional datasets such as Multiethnic Cohort Study and Kaiser Permanente Northern California datasets, while the percentage of participants racialized as Asian and/or Asian-American was much larger, the specific groups included were predominantly East Asian such as Chinese, Filipino, and Japanese.Participants from other distinct countries and regions were often excluded or lumped into the "other Asian" group, which may also mask important socio-economic differences among populations in addition to cultural and political ones.In studies included in this systematic review, the average income and education level of study participants racialized as Asians were often significantly higher than those of participants racialized as nHW individuals, suggesting significant sample selection bias in existing datasets.[84][85] The Census continues to use the OMB classification, standardized in 1980, which consists of six (addition of 'Other') racialized categories defined by the dominant (European and/or European-American) culture. 86Before 1980, the Social Security Administration only used three categories (White/Black/Other). 86 As a result, racial identifiers in Medicare datasets and many other large datasets are generated using these broad, externally informed categories.However, subsequent studies have found that people identified as Hispanic and Asian are often miscategorized. 86,87e overall practice of simplifying the data collection process was also reflected in the fact that few studies collected data on language and immigration status, including refugee status, which deeply influences screening, assessing, and diagnosing ADRD.Given that immigrants from Asia make up the largest immigrant group 26  proficiency is associated with a delayed diagnosis of ADRD, 88 as well as lower income, education, and access to health insurance. 89Research on immigrants racialized as Hispanic observed a higher prevalence of ADRD compared to native-born individuals racialized as Hispanic. 78A recent systematic review in Australia found that refugees and asylum seekers had higher rates of depression, social isolation, physical inactivity, and diabetes (modifiable ADRD risk factors) compared to their Australian-born counterparts.These studies shed light on the potential risk factors that might be observed in individuals racialized as Asian and/or Asian-American, given similar exposures. 90e lack of a concerted effort to engage diverse groups in ADRD research leads to an underestimation of the needs of individuals racialized as Asian and/or Asian-American, many at risk of ADRD and delayed or missed diagnosis.This issue is further exacerbated by the lack of data disaggregation in research among groups categorized as Asian and/or Asian-American.For example, among Cambodian and Bangladeshi individuals in the United States, poverty rates are approximately 20%, compared to 12.6% in individuals racialized as Asian-American.Similarly, while 75% of individuals racialized as Asian-American hold a bachelor's degree or higher, only 18% of Burmese and Hmong individuals attained a bachelor's degree. 9nce poorer access to educational attainment and lower income are significantly associated with the faster progression of dementia, 34,39 the lack of representation of individuals racialized as Asian and/or Asian-American with lower education and/or income in ADRD studies introduces significant sample bias.
ADRD, as a concept and formal medical diagnosis with different etiologies, intersects with language, social and familial roles, community beliefs, experiences, and perceptions, which can result in individuals being stigmatized by their community. 3,91However, those risk factors have not been thoroughly examined in the studies included in this sys-tematic review and meta-analysis.Related to language, dementia does not exist in the lexicon of many languages.For example, among Hmong Americans, dementia is more commonly related to chronic confusion or "tem toob," which is a term that normalizes forgetting things as one ages. 92As a result, a formal medical diagnosis of dementia, associated screening/assessment, and treatment may not translate or have the same impact without the appropriate cultural context.More often, the term dementia can have a negative and stigmatizing connotation.
The elders are often held in high regard in the family/community, and a diagnosis like dementia may be interpreted as labeling the person with a mental illness. 93There is a need for the family to protect/insulate their elders from perceived stigmatizing conditions that may undermine their value and respect, cultivated over decades. 93This requires care providers and researchers to be sensitized to the cultural values of a community, understand the heterogeneity among adults racialized as Asian and/or Asian-American, and use skilled and culturally responsive approaches to care and communication.
There were several important limitations to this review.We excluded studies conducted on MCI among samples racialized as Asian and/or Asian-American, which both limit the generalizability of study results and reduce the number of studies we reviewed on ADRD prevalence and incidence rate.Additionally, given that two-thirds of the studies originated from three cohort studies, two on the West Coast and one not in the mainland United States, geographical and sociopolitical factors may shape access to healthcare and time to a diagnosis differently.Generalizability may be limited when examining the same ethnoracial group in different cities.While not a limitation, it is important to note that we focused exclusively on studies that examined the experience of individuals racialized as Asian and/or Asian-American living in the United States.Also, the English language restriction and inclusion of literature published since 1990 may have excluded published work with additional incidence and prevalence rates.
This review reveals critical areas for future work.More studies are needed to comprehensively understand the unique needs of participants racialized as Asian/Asian-American. Recruiting participants who are racialized as Asian/Asian-American could be challenging due to the stigma associated with ADRD, lack of sufficient funding, and access to specific communities.Engaging older adults racialized as Asian/Asian-American and advocating for representation in research requires tailored approaches.These may include conducting community outreach programs with healthcare experts from these diverse communities, providing language assistance (e.g., interpreters), developing validated and linguistically tailored assessments, offering free, culturally-sensitive cognitive screening, and implementing brain health education to destigmatize ADRD.
When we collect data, disaggregated data are needed on the nationality, generational status, language preference, and immigration/refugee status of participants racialized as Asian and Asian-American. 94These data will provide a more robust understanding of how these factors impact ADRD risk in the United States.
Our study suggests that language plays a critical role in the screening, assessment, and diagnosis of ADRD.Among the studies in this review, the wide variation in prevalence and incidence rate of ADRD among participants racialized as Asian and/or Asian-American suggests that a lack of validation of diagnostic instruments across different languages may limit research on ADRD.Additional research is necessary to improve the accuracy and validity of diagnostic tests among individuals racialized as Asian and/or Asian-American in the United States.In addition, moving forward, studies should report on English-language proficiency in the study sample, whether English or another language was used in the assessment (including translated instruments), and any obstacles present during data collection that may impact the validity of the results.Finally, the lack of data disaggregation among groups racialized as Asian and Asian-American masks critical differences between groups.In addition to disaggregation, data collection efforts should include diverse categories for nationality and ethnicity, structural and social determinants of health (racism, capitalism, sexism), multidimensionality poverty, and recruitment location of the participants.

CONCLUSION
In conclusion, this systematic review and meta-analysis sheds light Out of 45 studies, 22 (48.9%)used the Honolulu-Asia Aging Study cohort.Six studies used data from Kaiser Permanente Northern California healthcare members who participated in the California Men's Health Survey and Kaiser Permanente Research Program on Genes, F I G U R E 1 PRISMA flow diagram of studies screened and included in this systematic review.Environment, and Health.Two studies used data from The KAME project (not an acronym) in Seattle, Washington, and the remaining studies used other cohorts nationwide or community samples with ADRD data.
used the OMB six-category self-reported ethnoracial groups.Kornblith et al. (2022) 36 only used five categories, which did not distinguish between Hispanic and non-Hispanic regarding groups racialized as Black and White.Multiethnic Cohort Study 8 and Kaiser Permanente Northern California 7,33,34,37-39 collected country of origin to further classify participants racialized as Asian and/or Asian-American.The study by Tsoy et al. (2021) used the Research Triangle Institute's algorithm to identify race/ethnicity based on the surnames of participants.
Only 45 studies were identified based on 13 unique datasets across the United States.Additionally, two thirds (30/45) of the studies were based on three projects: Honolulu-Asian Aging Study, KAME, and Kaiser Permanente Northern California.While these three cohorts are well-characterized, the findings cannot be generalized to individuals of the same ethnoracial group in different cities or individuals racialized as Asian in the overall United States.In retrospective studies using national datasets such as the National Alzheimer's Coordinating Center or Medicare claims data, participants racialized as on the gaps and challenges in ADRD research concerning individuals racialized as Asian and/or Asian-American.The study emphasizes the need for more comprehensive and nuanced approaches to studying ADRD within this population.This review brings attention to limitations in sample representation, diagnostic criteria, and cultural considerations among older adults racialized as Asian and/or Asian-American.It underscores the scarcity of comprehensive datasets, especially for groups racialized as non-East Asian, and the impact of language, culture, and immigration on ADRD progression, diagnosis, and care.This review proposes that future research should collect data on citizenship, generational status, and linguistic proficiency and incorporate analyses that examine country of origin, immigration status, and structural and social determinants of health.Studies should also enhance reporting of language usage during assessments to foster more inclusive and nuanced ADRD research with individuals racialized as Asian and/or Asian-American.
Asian-American participants included across the 45 studies.
a Based on the Research Triangle Institute (RTI; based on an algorithm using beneficiaries' surnames to identify race/ethnicity).b Based on the weighted sample.c The demographic information in based on Mayeda et al. (2016) and (2017).d Kaiser Permanente Northern California health plan members who completed one of two harmonized health surveys: California Men's Health Study or Kaiser Permanente Research Program on Genes, Environment, and Health Survey.
49aves et al. (1996)12and White et al. (1996)48employed the Diagnostic and Statistical Manual of Mental Disorders (DSM) IV49and Prevalence of Alzheimer's disease and related dementia among Asian participants.
III-Revised 50 in both KAME project and Honolulu-Asia Aging Study cohort after administering the cognitive functioning screen, behavioral, and neurological examination.The Honolulu-Asia Aging Study also adopted Cummings and Benson (1992) 51 criteria.White et al. (1996) 48 further incorporated CT scans and blood tests (complete blood cell count, chemistry profile, vitamin B 12 level, folate level, rapid plasma reagin, and thyroid function tests).Additionally, White et al. (1996) 48 also used the autopsy criteria for comparison in the Honolulu-Asia Aging Study cohort.Silverman et al. (1992) 40 was the only study that utilized the Alzheimer's Disease Risk Questionnaire 52 to obtain information on ADRD.Sixteen studies reported the prevalence of ADRD among different groups racialized as Asian and/or Asian-American based on TA B L E 3 Incidence rate of Alzheimer's disease and related dementia among Asian participants.

Table 5 .
Most of the studies (n = 19) investigating specific risk factors of ADRD among individuals racialized as Asian and/or Asian-American were based in the Honolulu-Asia Aging Study cohort.Physiological factors and genetic factors were the most frequently studied factors, with 12 studies focused on them.Conversely, psychological, social, and behavioral factors were not as extensively explored, having been the focus of only eight studies.
Putative Alzheimer's disease and related dementia risk factors.