Design of MOSAAIC (Multi-Ethnic Observational Study in American Asian and Pacific Islander Communities)

Abstract

Background

Asian American (AsA) and Native Hawaiian and Pacific Islander (NHPI) populations are underrepresented in U.S. health studies.

Objectives

The MOSAAIC (Multi-ethnic Observational Study in American Asian and Pacific Islander Communities) is a cohort study designed to improve understanding of disparities in cardiovascular disease and other health conditions affecting AsA and NHPI groups.

Methods

Through broad eligibility criteria and outreach tailored to each study community, MOSAAIC will include 11,500 adults aged 18+ years living in 5 field center regions: New York, Philadelphia, Chicago, San Francisco Bay Area, and Honolulu. The protocol specifies sample-size targets to allow valid comparisons among persons having personal or family origins in 4 geographically defined regions, including East Asia (the largest groups being Chinese and Korean), South Asia (Indian), Southeast Asia (Vietnamese and Filipino) and Oceania (NHPI). Measurements to be obtained at an in-person examination include clinical laboratory tests, medical history, medication use, spirometry, cognitive and physical function, anthropometry, and electrocardiogram. Surveys in multiple languages to assess medical, behavioral, lifestyle, social and environmental influences on health were developed by a multistep process to ensure equivalency between translations and cultural appropriateness. Biospecimens are stored for future studies. Long-term follow-up will be conducted to ascertain and adjudicate major health events including myocardial infarction, stroke, heart failure, and mortality.

Conclusions

The National Institutes of Health established MOSAAIC as a resource for wide-ranging epidemiologic investigation of factors related to cardiometabolic, pulmonary, or mental health in persons of diverse AsA and NHPI background.

Central Illustration

Asian American (AsA) and Native Hawaiian and Pacific Islander (NHPI) persons in the United States comprise 7.7% (25.5 million) of the overall population but remain underrepresented in national health studies. In 1948, the National Heart, Lung, and Blood Institute (NHLBI) of the National Institutes of Health (NIH) (then known as the National Heart Institute) established the Framingham Heart Study, which was followed by a series of similar cohort studies to characterize the epidemiology of cardiovascular disease among U.S. residents. As of 2021, only 2% of prior NHLBI cohort study participants have been of AsA or NHPI backgrounds.¹ None of the prior NHLBI cohort studies on disease incidence and risk factors have been focused on AsA and NHPI populations, nor have they been designed to compare these populations by background, birthplace, language preference, duration of residence in the United States, age groups, and other characteristics known to affect health. As a result, differences in health among AsA and NHPI groups often go unrecognized because of data that are sparse, aggregated, or absent completely.

To address these gaps, NHLBI and other NIH institutes funded the MOSAAIC (Multi-ethnic Observational Study in American Asian and Pacific Islander Communities) project in 2023. Using a community-based, longitudinal cohort study design, MOSAAIC will improve the understanding of variation in measures of health and disease among several AsA and NHPI groups living in multiple U.S. regions. This report summarizes the key research goals and design features of MOSAAIC.

Genesis of MOSAAIC: Proposal Evaluation and Award

In 2021, an NIH-wide workshop identified knowledge gaps concerning the health of AsA and NHPI populations, particularly with regard to cardiometabolic, pulmonary, and mental health conditions.¹ The workshop concluded that little was known about the nature of and reasons for variation in coronary heart disease, stroke, and chronic lung disorders across AsA and NHPI population groups. In August 2022, the NIH solicited applications to establish clinical and community field centers (CCFCs) and a coordinating center (CC) for the MOSAAIC project through RFA-HL-23-015 (CCFCs)² and RFA-HL-23-016 (CC).³ NHLBI serves as the lead institute for MOSAAIC and additional support is provided by multiple institutes, including the National Human Genome Research Institute, National Institute for Mental Health, National Institute for Minority Health and Health Disparities, and National Institute for Neurological Disorders and Stroke. Applications were peer reviewed for their ability to recruit and follow a combined cohort of at least 10,000 adults (2,000 at each of the 5 CCFCs) from AsA and NHPI populations into a prospective cohort study, and for their solid community partnerships and experience to ensure participants’ retention over long-term follow-up. The study investigators integrated the specific scientific and operational approaches described in each application into a single, harmonized protocol with guidance and approval from the NIH program team. The project is organized into a 2-year UG3 pilot phase (2023-2025), during which time each CCFC would enroll at least 100 individuals; and a 5-year UH3 phase, which includes full-scale participant recruitment and follow-up (2025-2030). The initial award included a cohort of adults between the ages of 18 and 64 years. In 2025, additional funding from the National Institute on Aging helped expand the cohort to include individuals aged 65 years and older, adding measures of physical function and increasing the total sample size to 11,500 (NOT-HL-24-029).⁴

Study Objectives

In the initial (baseline) phase of MOSAAIC, over a 7-year funding period (Figure 1), the objectives are as follows:

• 1.To establish an epidemiologic cohort of 11,500 consenting adults living in 5 CCFC regions within the United States, who self-identify as having East Asian, South Asian, Southeast Asian, Native Hawaiian, and/or Oceania or Pacific Islander background.
• 2.To characterize study participants with respect to factors that are known or suspected to be related to cardiometabolic, pulmonary, or mental health; to characterize the epidemiology of age-related cognitive and physical function decline; and to estimate the prevalence, mean values, and distribution of key health and risk factors, with sufficient sample size to stratify by country of origin or background, larger geographic region, Indigenous ancestry (Native Hawaiian), and other demographic characteristics.
• 3.To evaluate associations among baseline disease risk factors and to evaluate relationships of risk factors with health outcomes.
• 4.To serve as a platform for future ancillary studies, using the stored biospecimens and collecting new data to address study questions beyond those proposed in the core study protocol.

Figure 1.

The MOSAAIC Study Timeline

Initiated in 2023, the UG3 funding phase of MOSAAIC (Multiethnic Observational Study in American Asian and Pacific Islander Communities) (August 2023 to July 2025) involved protocol development and a pilot enrollment phase during October 2024 through July 2025. Full-scale recruitment will continue through July 2028 during the UH3 funding phase, which also involves follow-up and outcomes data collection.

Study Locations and Organization

The MOSAAIC CCFCs are located in New York, New York; Philadelphia, Pennsylvania; Chicago, Illinois; San Francisco Bay Area/Palo Alto, California; and Honolulu, Hawaii (Table 1). The MOSAAIC CC is based at the Fred Hutchinson Cancer Center (Seattle, Washington), in collaboration with the University of Washington (Seattle, Washington) and Albert Einstein College of Medicine (Bronx, New York). The MOSAAIC central laboratory and biorepository is located at University of Minnesota (Minneapolis, Minnesota). Specialized reading centers oversee collection of electrocardiogram measurements (Wake Forest University) and spirometry tests (University of Pittsburgh). The Fred Hutchinson Cancer Center provides the single institutional review board for the study. A list of contributing investigators from all funded sites and sponsors appears in the supplemental materials (Supplemental Table 1).

Table 1.

MOSAAIC Study Sites and Lead Investigators

	Institutions	PIs/ Multiple PIs
Philadelphia CCFC	Fox Chase Cancer Center; Temple University; Harvard T. H. Chan School of Public Health	Carolyn Fang, PhD; Grace Ma, PhD; Frank Hu, MD, PhD
New York CCFC	New York University Grossman School of Medicine; Langone Health; Perlmutter Cancer Center	Jiyoung Ahn, PhD; Yu Chen, PhD
San Francisco Bay Area CCFC	Stanford University; Stanford Medicine	Ann Hsing, PhD, MPH; Clete Kushida, MD, PhD; Joseph Wu, PhD; Paul Wang, MD
Chicago CCFC	University of Chicago; Penn State University	Brisa Aschebrook, PhD, MPH; Habib Ahsan, MD; Karen Kim, MD
Hawaii CCFC	University of Hawaii at Manoa; John A. Burns School of Medicine; University of Hawaii Cancer Research Center	Marjorie Mau, MD, MS; Alika Maunakea, PhD; Lani Park, PhD, MPH
Coordinating Center	Fred Hutchinson Cancer Center; University of Washington; Albert Einstein College of Medicine	Garnet Anderson, PhD; James Floyd, MD, MS; Gary Chan, PhD; Robert Kaplan, PhD
Central laboratory and biorepository	University of Minnesota	Bharat Thyagarajan, MD, PhD, MPH
ECG reading center	Wake Forest University	Elsayed Soliman, MD, MS
Spirometry reading center	University of Pittsburgh	Juan Celedon, MD, DrPH

CCFC = clinical and community field center; ECG = electrocardiogram; MOSAAIC = Multiethnic Observational Study in American Asian and Pacific Islander Communities; PI = Principal Investigator.

MOSAAIC is governed by a Steering Committee comprising Principal Investigators from each CCFC and the CC, representatives of NHLBI, and an external chairperson selected by NHLBI. The Steering Committee is supported by subcommittees responsible for particular areas, such as design and analysis, survey development, biospecimen protocols, community engagement, publications and presentations, and quality control (Supplemental Figure 1). An Observational Study Monitoring Board made up of external experts is charged with program oversight and monitoring of participant burden and safety. All CCFCs collaborate with local community advisors and organizations to implement the study in an effective and culturally appropriate manner. This community-centered approach has been adapted to the needs and preferences of each CCFC community, with the Hawaii CCFC having a formalized process for obtaining community approval for study activities and other CCFCs using more informal or ad hoc processes of community consultation.

To fulfill the mission of MOSAAIC as an accessible research resource, all de-identified data will be made publicly available through data repositories or in collaboration with MOSAAIC investigators. Following the precedent of major NIH-initiated studies, public-use versions of the MOSAAIC data will provide cleaned, derived variables as well as original variable codings with missingness indicators. Participants are able to choose restrictions; for example, during the informed consent process, they may refuse to make their data and specimens available to commercial, for-profit enterprises. In the case of Indigenous populations, a Community Review Board at the Hawaii CCFC site will provide an additional community level of review on the use of NHPI data and/or biospecimens. MOSAAIC’s Publications and Presentations Committee ensures that study publications meet a high standard of scientific validity and cultural appropriateness. Likewise, the Ancillary Studies Committee reviews all proposed ancillary (add-on) studies.

Protocol Development Process

The investigator teams and funding agencies collaborated to create a uniform study-wide protocol. First, the MOSAAIC Steering Committee agreed by consensus to assemble a cohort of adults with broad age eligibility that had adequate sample sizes for disaggregated analyses among the 5 most populous AsA groups⁵ (Chinese, Asian Indian, Filipino, Korean, and Vietnamese) and Native Hawaiians. We considered U.S. Census demographic data, local population sizes around the 5 study centers, community engagement experience of each field center with these populations, and gaps in data availability among these groups¹ to guide decisions about participant enrollment strategies.

Clinical measurement protocols and survey instruments were adopted from prior studies, both for efficiency and to facilitate between-study comparisons. Meanwhile, several innovative approaches were required to achieve the MOSAAIC study goals, for example: 1) a flexible participant recruitment methodology that accounted for unique circumstances and community resources in the 5 CCFC regions; 2) multilingual surveys that would enable broad inclusivity, while maintaining cultural appropriateness, acceptability, and standardization across all participants; and 3) an electronic data capture system that would support both in-person and online methods of participant eligibility screening and data collection in multiple languages.

Enrollment and Eligibility Criteria

Eligibility is inclusive of adults aged 18 years and older, who themselves or whose ancestors originated from one of the study’s designated geographic regions within Asia or Oceania (Table 2). Participants must be able to complete data collection in person at one of the MOSAAIC CCFC sites, they should have no immediate plans to move away from their region, and they must not be currently institutionalized or in active military service. Pregnancy is not an exclusion criterion but necessitates delay of study enrollment until the postpartum period.

Table 2.

MOSAAIC Study Participant Eligibility for 4 Sectors Within the Asia-Pacific Geographic Region

Geographic Sector	Country of Origin Groups Included Within Regions
East Asia	Chinese, Korean, Taiwanese, Japanese, Mongolian
South Asia	Indian, Pakistani, Bangladeshi, Nepalese, Sri Lankan, Bhutanese, or Maldivian
Southeast Asia	Filipino, Vietnamese, Cambodian, Laotian, Thai, Indonesian, Malaysian, Hmong, Singaporean, Burmese
Oceania	Native Hawaiian, Micronesian, Melanesian, Polynesian

Abbreviation as in Table 1.

Qualifying MOSAAIC participants are those who self-identify as a person of East Asian, South Asian, Southeast Asian, Native Hawaiian, and/or Pacific Islander background, stating either a personal place of origin or ancestry belonging to these population groups. Individuals who have non-Asian or non-NHPI ancestors may be enrolled provided that they consider themselves to be Asian, AsA, Native Hawaiian, and/or Pacific Islander. Multiethnic individuals will be eligible to enroll, and extensive data on birthplace and background of the enrollee and parents will be made available for use by researchers. Participants have the option to complete an Eligibility Screener and all other questionnaires either online or on paper, in English, traditional or simplified Chinese, Korean, Vietnamese, or Tagalog. These languages were chosen after consideration of English language proficiency among the most populous AsA and NHPI groups, as well as local CCFC experience and capabilities, considering feasibility and available resources including staffing.

Each CCFC implements a participant recruitment plan tailored to the characteristics and needs of their local study community and available research facilities. The CC supplies core recruitment support materials, including the main study website, videos and flyers, which are developed through the Community Outreach and Engagement Committee. CCFCs may modify and supplement these materials with site-specific information to enhance their relevance and engagement within their communities. Through the public-facing study website, anyone can request to be considered for study enrollment. Other methods used to identify potentially eligible individuals include community events, direct mail, patient portals, phone outreach, and respondent-driven sampling. To reach older adults, CCFCs will partner with community-based organizations that serve the older population and provide staff assistance to support participants with mobility or sensory needs.

A person is considered to have enrolled in the MOSAAIC study after providing informed consent, attending an in-person examination, and completing a minimum set of data collection activities, specifically: surveys on demographic information, birthplace, and medical history (>80% of questions completed); anthropometry including height, weight, waist circumference, and hip circumference; blood pressure; pulse; and blood collection.

The CC monitors enrollment so that the study achieves a desired distribution across AsA and NHPI populations, with balance across sex strata. The target age distribution of participants is 5,000 aged 18 to 39 years, 5,000 aged 40 to 64 years, and 1,500 aged 65 years and older.

Schedule of Data Collection Episodes

MOSAAIC involves multiple data collection episodes over time (Supplemental Figure 2). First, a brief online eligibility screening survey assigns the potential participant to a MOSAAIC CCFC in their local region and identifies those likely to meet the eligibility criteria. On passing the eligibility screen, an enrollment survey captures contact information and preferred languages; the CCFC then reaches out to schedule an in-person baseline examination. The examination is a multi-hour encounter that may be conducted at clinic, mobile, or temporary (eg, community-based) locations to maximize convenience and comfort of participants. Survey data are mainly collected through a web-based tool. Questionnaires are organized in multiple tranches to spread the participant interactions over a period of several months. Over time, each participant will complete recurring annual follow-up surveys, which will update key information, add new questionnaire content, and ascertain health outcomes. Participants are provided with modest incentives after they complete study activities.

Clinical Measurements

The centerpiece of the baseline data collection phase is the in-person examination. Known, suspected, and novel risk factors and biomarkers are measured so that they can be studied in relation to future cardiovascular disease events and other incident diseases (Table 3). The examination visit is usually scheduled in the morning, beginning with informed consent and intake processes. In a typical sequence, before breaking their overnight fast, participants complete blood pressure measurement; electrocardiogram; and collection of spot urine, saliva, and venous blood. A breakfast break is followed by a medication inventory; a series of questionnaires; spirometry (both basal and post-albuterol challenge); anthropometry measurements including height, weight waist circumference, and hip circumference; and instructions for returning an at-home stool collection kit. Adults aged 65 years and older perform a physical function battery to assess balance, strength, and mobility. We designed a brief, easily administered cognitive assessment battery, comprising tasks that minimized linguistic bias and were used in studies with similar populations.⁶ Prior experience suggests that cross-linguistic variation in response patterns is minimal across our selected cognitive tests, and we plan to compare normative responses across ethnic groups stratified where feasible by language, sex, educational attainment, and age group.

Table 3.

MOSAAIC Examination Components and Instrumentation

Examination component	Instrumentation
Anthropometry (height, weight, waist and hip circumferences)a	Stadiometer: Health-o-Meter Professional Wall-Mounted Height Rod (fixed); HM200P PortStad Portable Stadiometer (portable) Weight, body fat %: Tanita WB-3000 Digital Physicians Scale Body circumferences: Gulick II tape measure
Biospecimen collection
Blooda	Serum, EDTA plasma, buffy coat/packed red blood cells, cryopreserved peripheral blood mononuclear cells, PAXgene Blood RNA tube
Urine	Spot urine sample
Saliva	OGR-500 saliva collection kit
Stool	iSWAB stool collection kit for DNA and stool sample fixed in alcohol for metabolomics
Seated blood pressurea and pulsea	OMRON HEM 907 XL
Electrocardiograma	GE MAC 5
Spirometry	EasyOne (NDD) spirometer
Medication inventory	RxNorm medication database
Cognitive assessment	Semantic Animal Fluency16
	Common Objects Memory Test17
	Symbol Digit Modalities Test18
	Bell Cancellation Testb,19
	Digits Backward20
Physical Performance	Semi-tandem stand
	Side-by-side stand
	Tandem stand
	Single-leg stand
	4-meter walk
	Single chair stand
	Repeated chair stands
	Grip strength

Abbreviation as in Table 1.

^aIndicates fasting component.

^bBell Cancellation Test was substituted for the Clock Drawing Task after evaluation of pilot phase.

On finishing the examination, each participant completes an exit interview. The MOSAAIC central laboratory analyzes biological samples, which have been processed and frozen at −80°C (or −192°C in the case of cryopreserved peripheral blood mononuclear cells) at the CCFC (Table 4). Additional volumes of serum, plasma, buffy coat/packed red blood cells, peripheral blood mononuclear cells, PAXgene samples (for blood RNA extraction), urine, saliva, and stool are held in central long-term storage (Table 4). Clinically relevant results such as body mass index, blood pressure levels, serum cholesterol levels, and diabetes tests are returned to participants. CCFC staff facilitate appropriate clinical referrals as needed for routine clinical findings, and procedures that are tailored to the logistics of each CCFC handle any emergent or urgent medical problems. All staff must complete a rigorous training and certification process before commencing to collect study measurements. Before study launch, an in-person training was conducted with staff members from each CCFC, as a train-the-trainer model, the attendees returning to their CCFC to train the remainder of their research team. The reading centers, Performance Monitoring Committee, and Laboratory & Repository Committee monitor metrics related to standardization and quality control.

Table 4.

Biospecimen Collection, Storage, and Laboratory Measurements

Sample Type	No. Tubes/Volume	No. of Aliquots	Assay Measurements
Saliva	1 × 4 mL	1
Spot urine	1 × 120 mL (maximum)
	Testing aliquot	1	Albumin/creatinine ratio
	Storage aliquots	2
Serum	1 × 10 mL
	Testing aliquot	1	Metabolic panela
	Small storage aliquots	3
	Large storage aliquots	2
EDTA	3 × 10 mL
	Whole blood: testing	1	HbA1C
	Plasma: storage	20
	Buffy coat/RBC:storage	2
PBMCs (EDTA tube)	1 × 10 mL, 5-10 million cells	2-3
PAXGene	1 × 2.5 mL	1
Stool – Metabolomics	1 × 1 mL	1
Stool – Microbial DNA	1 × 1 mL	1

PBMC = peripheral blood mononuclear cell; RBC = red blood cell.

^aGlucose, insulin, creatinine, C-reactive protein, total cholesterol, high-density lipoprotein cholesterol (direct), low-density lipoprotein cholesterol (calculated), triglycerides.

Survey Measurements

Survey instruments were selected by 5 working groups, each focused on a distinct domain: clinical measures, mental health and cognition, health behaviors, psychosocial factors, and social and environmental determinants of health (Table 3). When possible, MOSAAIC adopted questionnaire batteries that had been used in prior studies, favoring especially those in the public domain that were known to have good measurement properties (eg, reliability, Cronbach’s alpha), and were already translated into Asian languages. (The Supplemental Materials describe the provenance and measurement properties of each survey.) Survey instruments are mainly self-administered using online data capture but are also offered for completion on paper or by interview in monolingual or bilingual formats, according to participant preferences and needs.

A Translation and Cultural Adaptation Working Group devised a multistep process to ensure that translations met minimum criteria for between-language equivalency and cultural appropriateness. This involved a panel of experts for each of the 5 non-English study languages (simplified and traditional Chinese, Korean, Vietnamese, and Tagalog). In addition, a subset of the 400-plus survey items was selected for community input, focusing on items that covered sensitive issues, included abstract concepts (eg, emotions), or otherwise were deemed to require careful wording to ensure cultural appropriateness. In each language, CCFCs conducted field testing among layperson volunteers, including both monolingual and bilingual individuals (N = 141). This employed a modified cognitive interview technique, eliciting feedback about comprehension, clarity, and interpretation to ensure that final survey measures were both semantically equivalent and culturally respectful across all target populations. We created a translation dictionary of commonly used words, so that all study communications, surveys, and instructions would have consistent and correct verbiage.

Outcomes Surveillance and Adjudication

To ensure accurate and complete ascertainment of incident health outcomes, MOSAAIC has adapted established surveillance and adjudication approaches used in other NHLBI-funded prospective cohort studies.7, 8, 9 Participants (or their proxies) will be contacted annually to obtain information on new diagnoses from their health care providers, hospitalizations, and deaths. For primary outcomes, which include acute myocardial infarction, heart failure, stroke, and death, CCFC staff will obtain relevant medical records that will be reviewed by study physicians and adjudicated centrally using standardized diagnostic criteria. Deaths will be identified through next of kin reports, linkages with the National Death Index, and local queries performed by CCFCs, and cause of death similarly adjudicated using medical records, autopsy, and death certificate information. Although identification of incident cardiovascular events is based on annual surveillance phone calls with participants, searches of vital statistics databases will be performed for participants who do not complete the follow-up contacts. Language in the informed consent permits the study investigators to continue long-term follow-up of the cohort beyond the end of the current funding period in 2030. In addition to these well-established cohort study outcome surveillance methods, future stages will link with data from electronic health record exchanges, which have the potential to provide rich information on a broad range of health outcomes from diagnosis codes, laboratory and imaging results, and even unstructured electronic health record data.

Retention Strategies

The Community Outreach and Engagement Committee developed a participant retention plan, which includes activities that build on the community-centered approach applied in the recruitment of participants. Informed by the literature10, 11, 12 and extensive experiences and knowledge from the CC and CCFCs, a study-wide retention plan promotes continued participant engagement focused on meaningful dates of individual or community significance (eg, birthdays, enrollment anniversaries, New Year’s greetings, and other holidays with cultural significance). A newsletter provides study updates and showcases participants’ stories, and culturally tailored social media campaigns celebrate cultural holidays throughout the year, including Asian American and Pacific Islander Heritage Month in May. MOSAAIC also strengthens retention through CCFC partnerships with trusted community leaders and local community- and faith-based organizations to relay recruitment and retention messages, reflecting engagement strategies preferred by Asian, AsA, and NHPI communities.¹³ In addition, MOSAAIC uses social organizations as a source of potential enrollees and allows multiple household members to enroll, thus participating alongside others in their social networks can support retention,¹¹ especially relevant because of this population’s orientation toward family and community. For instance, the Chicago CCFC collaborates with multiple community-based organizations in the Greater Chicago area to establish community-based clinics to enroll local community members of AsA or NHPI backgrounds. This approach creates a community-based pathway to enroll participants and to better retain them throughout the study period.

Data Analysis Approaches

Two primary analytic objectives are: 1) to estimate the prevalence, mean values, and distribution of key health and disease risk factors, stratified by background groups and other relevant demographic characteristics; and 2) to evaluate relationships among baseline risk factors and between risk factors and health outcomes. Although the study will include multiple AsA and NHPI backgrounds within the Asia and Oceania regions (Table 2), well-powered disaggregated analyses require sufficient sample sizes for selected background groups. To address this need, while considering CCFC catchment areas, U.S. population sizes, and cost, the MOSAAIC Steering Committee prioritized 6 groups for disaggregated analyses, Chinese, Filipino, Indian, Korean, Native Hawaiian, and Vietnamese, with a goal of at least 1,730 enrolled individuals from each. The study will provide power to identify differences in disease and risk factors across groups (Supplemental Table 2). Analyses also may be conducted at broader regional levels and will typically include participants outside these 6 disaggregated groups. Aside from field center, background, age, and sex, additional variables such as country of birth, preferred language, and socioeconomic position will not have predefined enrollment targets. Rather, the recruitment strategy will focus on ensuring the inclusion of participants across these characteristics, even if their proportions may differ from those in the U.S.-wide population. We will use statistical approaches such as post-stratification weighting to improve generalizability to the national population. In evaluating associations between risk factors and health outcomes, interactions between background group or geographic region and risk factors will be examined to assess differential effects. Regression models appropriate to each outcome will be applied (eg, Cox proportional hazards or generalized linear models, adjusting for confounders). Methods such as generalized estimating equations or mixed and random effects will account for individuals within the same household and field center. The Design and Analysis Committee will develop additional recommendations for appropriate analyses of MOSAAIC data, including modeling choices, suggested sensitivity analyses, and imputation strategies for handling missing data, such as Multiple Imputation by Chained Equations.

Study Limitations, Challenges, and Solutions

The MOSAAIC study design applies largely the same methodologies that have been used in previous large NHLBI population-based cohort studies. However, to overcome the challenges of studying AsA and NHPI populations with varied cultural and linguistic backgrounds (Central Illustration) required both innovations and trade-offs. A rigidly standardized protocol for identifying and recruiting study enrollees, implemented identically at each CCFC, would have offered some theoretical advantages for making between-center comparisons and deriving population estimates of disease prevalence. However, systematic sampling methods were expected to have low participation rates leading to inefficiency. Moreover, it was deemed impossible to identify 5 regional sampling frames that would capture all of the target populations without strong sampling bias. Instead, each CCFC devised an approach to recruitment appropriate to its local population, community partnerships, and institutional strengths. The absence of a single unified sampling frame across CCFCs may introduce selection biases. We therefore designed recruitment such that each CCFC will capture multiple population groups, and each AsA population will be represented in at least 2 CCFCs; this will reduce intercorrelation between Asian background and U.S. region (NHPI populations will be an exception, to be predominantly enrolled at the Hawaii site).

Central Illustration.

Multiethnic Observational Study in American Asian and Pacific Islander Communities

Visual representation of the geographic locations of the field centers recruiting participants and the coordinating center, the health outcomes being measured, and the multicultural nature of the study. Includes the MOSAAIC (Multiethnic Observational Study in American Asian and Pacific Islander Communities) logo with translations of the study slogan into the 5 additional study languages underneath to represent all languages available to participants.

The study has broad eligibility criteria to capture the AsA and NHPI population’s wide-ranging social, lifestyle, and biological characteristics. Accordingly, each CCFC welcomes participation by persons from any of the eligible AsA and NHPI background groups. Yet, strong engagement with local communities is a key element of MOSAAIC’s approach to recruitment and retention, which requires focused, culturally tailored strategies. Logistical considerations, including existing ties with community partner organizations and availability of multicultural and multilingual staff members, also made it desirable for each CCFC to focus on specific population groups. Thus, each CCFC was assigned numerical recruitment goals for 2 or more specific AsA and NHPI populations. Based on careful monitoring of recruitment, mid-course adjustments to outreach and recruitment strategies will be made, as needed to achieve the desired distribution of participants across the 4 regional study populations (East Asia, Southeast Asia, South Asia, and Oceania). The study allows multiple members of a household to enroll, which is helpful for achieving long-term retention and for capturing hard-to-reach individuals (eg, older adults unable to live independently, younger adults with changing residences). This aspect introduces between-subject correlations that must be accounted for in analyses.

Data collection via a custom online multilingual interface overcomes staffing challenges and permits respondents to toggle between English and Asian language versions to enhance comprehension. Although survey instruments were selected based on prior validation and documented translations, we acknowledge that construct validity may vary across the 6 major languages of administration or by cultural orientation. Cross-linguistic variations are mitigated by documenting each participant’s spoken and written language preference and survey administration language. Because a primary goal is to enable comparisons across background groups of AsA and NHPI people, data collection instruments are near-identical for each participant, with a few exceptions. For example, we will use 2 different food frequency questionnaires, one from the Canadian Study of Health Assessment and Risk in Ethnic groups for South Asian participants (eg, Indian),¹⁴ and another from the Multiethnic Cohort Study, which was validated and used in multiple other NHPI, Asian, and non-Asian populations, but had not been validated or used for South Asian diets.¹⁵ A staff guide was prepared that gave culturally appropriate examples of foods with similar nutrient content as the named food frequency questionnaire line items. The guide also gave further description of some line items that may not have been as culturally familiar to some participants.

Cultural adaptations to suit each CCFC community are vital to future retention efforts and trust-building. For instance, to engage NHPI communities in MOSAAIC, the Hawaii CCFC built trustworthy relationships that are community-driven and led by ethical Indigenous leadership. To reflect this, as a branding strategy the Hawaii CCFC adopted an additional study logo and name, the “Pacific Ocean Native Observational (PONO) Health Legacy Study.” “Pono” in the Hawaiian language means righteous, goodness, and balance and reflects the value and the long-term goals of the study. Integration of cultural adaptation and community-led review of study activities through the Hawaii Community Review Board emphasizes respect for Indigenous values, transparency, and community benefit, for example in decisions related to use, sharing, and disposition of specimens. Consistent with these principles, the Hawaii site follows a site-specific stewardship practice whereby any destruction of biospecimens must be coordinated with and returned to Hawaii, reflecting local expectations regarding custodianship and handling of biological materials. MOSAAIC’s integration of cultural adaptation and community-led review of study activities through the Hawaii Community Review Board establishes a replicable framework for ensuring that interpretation is contextual and inclusive of Indigenous Hawaiian and Pacific islander experts in science and culture to be represented and have their perspectives voiced.

Conclusions and Opportunities

Although the Asian, AsA, and NHPI populations to be studied in MOSAAIC may share genetic backgrounds with other large international cohorts, they have likely experienced different and more variable environmental, lifestyle, and societal exposures related to their national origin, immigration histories, and generation. Individuals of these backgrounds have experienced, to varying degrees, the barriers, stresses, and stereotyping experienced by other minoritized groups in the United States, whereas there exist both threats to health and wellness-promoting characteristics unique to them. Understanding the heterogeneity of lifestyles and their societal and environmental circumstances, as well as ancestry, is a critical step in characterizing community health needs and learning how to address them. Through its study protocol, as well as future ancillary studies yet to be developed, MOSAAIC will advance our knowledge of the health conditions of Asian, AsA, and NHPI persons and provide the infrastructure for deeper understanding of their biological, environmental, and social underpinnings.

Funding Support and Author Disclosures

Research reported in this publication is supported by the National Heart, Lung and Blood Institute, National Human Genome Research Institute, National Institute for Mental Health, National Institute for Minority Health and Health Disparities, National Institute for Neurological Disorders and Stroke, and National Institute on Aging under Award Numbers U24HL169645, 3U24HL169645-02S1, U24HL169645-02S2, 3U24HL169645-03S2, 3U24HL169645-03S1, 3U24HL169645-03S3, UG3HL169657/UH3HL169657 (Hawaii Community/Clinical Field Center [CCFC]); UG3HL170036/UH3HL170036 (Chicago CCFC), UG3HL169648/UH3HL169648 (SF Bay Area CCFC), and UG3HL169656/UH3HL169656 (New York CCFC). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. The authors have reported that they have no relationships relevant to the contents of this paper to disclose.