The Design and Conduct of a Community‐Based Registry and Biorepository: A Focus on Cardiometabolic Health in Latinos

Abstract

Background

Latinos are disproportionately impacted by obesity and type 2 diabetes but remain underrepresented in biomedical research. Therefore, the purpose of this project was to develop a research registry and biorepository to examine cardiometabolic disease risk in the Latino community of Phoenix, Arizona. The overarching goal was to establish the research infrastructure that would encourage transdisciplinary research regarding the biocultural mechanisms of obesity‐related health disparities and facilitate access to this research for the Latino community.

Methods

Prior to recruitment, key stakeholders from the local Latino community were engaged to develop a broad rapport within the community and seek advice regarding recruitment, enrollment, and follow‐up. Self‐identified community‐dwelling Latinos underwent a comprehensive cardiometabolic health assessment that included anthropometrics, a fasting laboratory panel, and a 2‐hour oral glucose tolerance test with measures of insulin and glucose to estimate insulin action and secretion. Separate consent was requested for future contact and banking of serum, DNA, and RNA. Research collaborations were sought out based on the cultural and metabolic profile of participants, faculty research agendas, and the potential for generating hypotheses.

Results

A total of 667 participants (20.4% children, and 79.6% adults) were enrolled with 97% consenting to the registry and 94% to banking of samples. The prevalence of overweight/obesity was 50% in children and 81% in adults. Nearly 20% of children and more than 45% of the adults exhibited some degree of hyperglycemia. To date, more than 15 research projects have been supported through this infrastructure and have included projects on the molecular biology of insulin resistance to the sociocultural determinants of health behaviors and outcomes.

Conclusions

The high prevalence of obesity and cardiometabolic disease risk factors coupled with the overwhelming majority of participants consenting to be re‐contacted, highlights the importance of supporting research infrastructure to generate hypotheses about obesity‐related health in Latinos. Future studies that stem from the initial project will likely advance the limited understanding regarding the biocultural determinants of health disparities in the Latino community.

Introduction

Participation in clinical research studies provides important benefits to both individuals and society. However, recruitment and enrollment of medically underserved minority groups into biomedical research can be a challenge. There are many potential barriers that account for this difficulty, including geography, sociopolitical factors, and trust in biomedical research institutions.1 Unfortunately, many of the chronic diseases that are becoming a public health and medical care burden disproportionately affect underserved minority groups. For example, Latinos represent the largest and fastest growing minority group in the United States2 and are disproportionately affected by obesity and type 2 diabetes. Approximately three‐quarters of Latino adults in the United States are overweight or obese3 and rates of type 2 diabetes are at least 80% higher compared to non‐Hispanic whites.4 Even more troubling is that approximately 36% of Latino adults with type 2 diabetes are undiagnosed4 further highlighting the need for research focused on this high‐risk population.

Although it is hypothesized that underserved groups may be less willing to participate in biomedical research, several recent reports suggest that minorities, including Latinos, are just as willing and interested to participate in research if given the opportunity.5, 6, 7 Therefore, access to research as well as willingness to participate are important factors that may influence minority participation in biomedical research. In order to improve the health of underserved communities and close the health disparities gap, it is imperative that researchers include minority participation in their studies. To this end, the NIH has mandated that any research they fund have provisions for including minority participants.8 Encouraging minority participation is important for clinical as well as social science and behavioral research as recent evidence suggests that most of the health disparities related to chronic disease are the result of biological as well as social factors.9 Therefore, it is critical that research examines how biological and social factors interact to influence health across the lifespan. These well‐integrated studies will provide a more complete understanding of the mechanisms underlying health disparities and ultimately facilitate the translation of science into clinical practice and public health promotion programs and policies.

In Arizona, Latinos represent approximately 30% of the population or nearly 2 million individuals10 and according to the Behavioral Risk Factor Surveillance Survey, 33% of Latino adults in Arizona are obese and 12.3% have been diagnosed with type 2 diabetes. With these statistics in mind and in an effort to support obesity‐related health disparities research among Latinos living in Arizona's largest city, Phoenix, we developed a community‐based registry and biorepository with the purpose of (1) facilitating access to biomedical research for an underserved community, (2) fostering transdisciplinary research collaborations among investigators, and (3) generating hypotheses regarding the biocultural mechanisms of disease among this high‐risk population. Our overall goal was to provide a diverse group of investigators with an effective mechanism along with the infrastructure to support research in the area of obesity‐related disease among Latinos in Phoenix, Arizona. This paper describes the design, personnel, methodology, clinical presentation of the cohort, and description of projects supported to date.

Methods

The project was designed as part of a statewide CTSA application to test the feasibility of linking basic and clinical research to community engagement activities. State‐appropriated funds were provided to support personnel (Project Director—25%, Research Nurse Coordinator—100%, and Administrative Assistant—50%), participant reimbursement, and the cost of lab testing. Other personnel included the PI, project physician, research nurses, research assistants, and lab technicians. An ad hoc scientific steering committee was developed to meet with investigators interested in collaborating and oversee administrative processes related to specimen and data sharing.

Establishing rapport with clinics and agencies serving the Latino community

Prior to enrolling participants, the study team met with Latino health and advocacy groups as well as local community health clinics serving the Latino population. Initial interactions were developed organically and facilitated through existing relationships and partnerships. The purposes of these interactions were to inform key community stakeholders about the project as well as engage these stakeholders in the long‐term goals. Advice was sought about recruitment and retention as well as how to handle follow‐up care in the event that treatment was required as provisions for medical care were not included in the project. The format of stakeholder interactions varied from formal presentations during regular meetings of community health coalitions and executive councils to informal discussions with community leaders and clinic directors. Although not formalized, ongoing contact and statusupdates were provided during the project period to facilitate the process of community engagement.

Enrollment process (individuals and families). Initially, participants were referred from our clinical partners who posted flyers in their waiting rooms and provided recruitment materials to the family members of their patients. Since obesity and diabetes is prevalent in this population, these initial recruitment efforts were quickly augmented by word‐of‐mouth and the project experienced an influx of family members and friends who were interested in participating. Prior to enrollment, participants were contacted by phone to explain the general purpose of the project and screened for basic inclusion/exclusion criteria. General inclusion criteria were (1) Ethnicity, self‐identified as Hispanic or Latino; (2) Age, initially the project focused on adults between the ages of 18–65 years of age, however, as the project developed over time and both participants and investigator interest grew, children as young as 7 and adults up to age 85 were included; (3) Health Status, in general, eligible participants were to be free of significant health conditions and not on any medication(s) known to effect insulin metabolism or glucose homeostasis. Examples of exclusion conditions were untreated metabolic disease, HIV/AIDS, active cancer or in remission for less than 3 years, acute illness, and currently pregnant. Once these general inclusion/exclusion criteria were met, participants were scheduled for their clinical visit described further.

Phenotypic characterization of participants. Potential participants were scheduled to arrive at the Clinical Research Unit of the Center for Metabolic Biology at Arizona State University after an overnight (8–10 hours) fast. Consent was administered to participants in their preferred language by a bilingual/bicultural research nurse. Participants were asked to consent to (1) procedures in the current study (described below), (2) collection and banking of additional biospecimens for future use, and (3) permission to be recontacted for future studies. Once written consent was obtained, an in‐depth medical and family history was collected and a brief physical examination was performed (Table 1). After the physical exam, an intravenous indwelling catheter was inserted in an antecubital vein and fasting samples were collected for a complete metabolic panel, a complete blood count, HbA1c, and a lipid panel. In those participants who consented to the biorepository, additional fasting samples were collected for storage of serum, RNA, DNA and cell line processing (described below). Participants then underwent a 2‐hour, multisample oral glucose tolerance test (OGTT). Blood samples were taken at ‐15 and ‐5 minutes prior to ingestion of 75g dextrose in solution with subsequent samples collected at 10, 20, 30, 60, 90, and 120 minutes for determination of plasma glucose and insulin concentrations. The multisample OGTT allows for estimation of various assessments of insulin dynamics including insulin sensitivity using the Whole‐Body Insulin Sensitivity Index described by Matsuda and DeFronzo11 and insulin secretion using the insulinogenic index.12 These measures were also used to estimate β‐cell function by the disposition index as the product of insulin sensitivity and insulin secretion.13 Upon completion of the study, individuals were given a snack, provided instructions for caring for the needle insertion site, and completed a form to indicate their preferred method for returning their results (i.e., mail, fax, or in‐person pick‐up).

Table 1.

Measures collected

Medical History, Exam, and Laboratory Measures
Family history of type 2 diabetes
Social history*
Physical activity*
Standing height
Weight
Body mass index
Waist circumference
Hip circumference
Seated blood pressure
Temperature
Urine sample
Fasting blood sample†
2‐Hour oral glucose tolerance test‡

*Social history included alcohol and tobacco use (yes/no, and amount) and physical activity included frequency, intensity, and duration.

^a†Complete blood count, lipid panel, metabolic panel, HbA1c.

^b‡Dosed at 1.75 g/kg to a maximum of 75 g with blood samples collected at −15, −5, 10, 20, 30, 60, 90, and 120 minutes for determination of insulin and glucose.

Specimen biobanking. Fasting serum was collected and stored in individual 0.5cc aliquots at ‐80°C for future use. Whole blood was collected in PAXgene DNA and RNA tubes, as per manufacturer's instructions and preserved at −80°C until future use. Whole blood was collected and shipped to the Tissue Culture Facility at the UNC Lineberger Comprehensive Cancer Center (Chapel Hill, North Carolina) for the immortalization of human B‐lymphocytes.

Results

Recruitment and demographics

Initial recruitment began in September, 2008 and continued through February, 2011. A total of 1,111 inquiries were received, and of these, 667 participants were enrolled. The majority of participants were adults over the age of 18 (79.6%) however, 136 children and adolescents were also enrolled. Of the participants enrolled, 353 had at least one other blood relative who also participated for a total of 92 families enrolled. Demographic characteristics of the participants by age‐group (children vs. adults) are shown in Table 2. Adult participants were predominantly Spanish‐speaking women while pediatric participants were evenly distributed by gender and most preferred English as their primary language. Regardless of age, the vast majority of participants consented to be included in the registry for re‐contact as well as for inclusion in the biorepository for banking samples.

Table 2.

Descriptive characteristics by age

	ADULTS (≥18)	YOUTH (7‐17)
Age (years)	36.3 ± 11.1	14.0 ± 2.6
Gender (M/F)	193/339	68/68
Spanish speaking preference (%)	77.3	26.5
Agree to re‐contact (%)	97.4	97.1
Agree to stored samples (%)	96.4	91.2
Weight status*	—	—
Lean (%)	19.2	50.0
Overweight (%)	36.5	25.0
Obese (%)	44.3	25.0

*Weight Status determined according to BMI in adults: (Lean = BMI < 25; Overweight = 25≤BMI<30; Obese BMI≥30 kg/m²) and age and gender adjusted BMI percentiles in youth: (Lean=BMI < 85th; Overweight=85th ≤ BMI < 95th; Obese=BMI ≥ 95th).

Cardiometabolic risk factors in adults

In general, the prevalence of obesity in adults was high with over 44% of adults considered obese (BMI ≥ 30 kg/m²), 37% overweight (25 ≤ BMI < 30 kg/m²) and only 19% considered lean with a BMI < 25 kg/m². The percentage of adults with cardiometabolic risk factors as defined by the Adult Treatment Panel III14 is depicted in Figure 1. Abdominal obesity and low HDL‐cholesterol were among the most prevalent risk factors. Nearly half of all adults exhibited some degree of hyperglycemia defined according to the American Diabetes Association15 as either impaired fasting glucose, impaired glucose tolerance, or overt type 2 diabetes, and almost 15% met the criteria for type 2 diabetes mellitus. Approximately one‐third had high triglycerides and/or elevated blood pressure. Despite the level of risk exhibited in the cohort as a whole, considerable variability in insulin action was observed (Figure 2). Those with the lowest insulin sensitivity were more likely to exhibit cardiometabolic risk factors compared to those with the highest insulin sensitivity (Figure 3).

Figure 1.

Frequency of cardiometabolic risk factors in adults. Percentage of adults with each individual risk factor for the metabolic syndrome.

Figure 2.

Insulin Sensitivity and Cardiometabolic Syndrome. Mean ± SD of insulin sensitivity in adults according to the number of risk factors for the metabolic syndrome. Risk factors include abnormalities in blood pressure, triglycerides, glucose, HDL‐cholesterol, and waist circumference as defined by the Adult Treatment Panel III.

Figure 3.

Distribution of Insulin Action. Mean level of insulin action as estimated by the whole body insulin sensitivity index across each decile in children (black bars) and adults (grey bars).

Cardiometabolic risk factors in youth

As expected, the degree of cardiometabolic risk was lower in the 136 youth compared to adults. However, 25% of youth were overweight (BMI ≥ 85th percentile according to age and gender and another 25% were obese (BMI ≥95th percentile). Although there is no universally accepted definition for cardiometabolic disease in the pediatric population, the criteria for prediabetes and type 2 diabetes is the same in children as in adults. A striking percentage of youth exhibited hyperglycemia (19.1%) and one case of overt diabetes discovered. Children with prediabetes (impaired fasting or post‐challenge glucose) exhibited 26.5% lower insulin sensitivity (3.6 ± 0.6 vs. 4.9 ± 0.3, p = 0.01) and 60.5% lower B‐cell function as estimated by the disposition index (3.8 ± 0.5 vs. 9.6 ± 1.6, p = 0.08) compared to healthy counterparts.

Use of the registry and repository

Use of the registry and repository is facilitated primarily through word‐of‐mouth among researchers with interests in obesity, diabetes, cardiovascular disease, and health disparities. In general, utilization can be divided into 3 main categories: (1) projects ancillary to the parent study (i.e., individuals consented to the parent registry/repository as well as a separate study and data collection occurred simultaneously), (2) recontact projects (i.e., separate projects that contacted participants in the registry who meet eligibility based upon demographic and/or clinic inclusion), or (3) repository projects (i.e., projects that utilize stored biospecimens linked to phenotypic data already collected). To date, 18 projects have been supported that range in scope from an undergraduate honors thesis to NIH‐sponsored clinical studies. These studies are outlined in Table 3 and represent ongoing research as well as completed projects that have contributed to preliminary data for NIH‐funded grant proposals.

Table 3.

Utilization of registry and repository

Study title	Study type
Quality of Life and Metabolic Risk in Latino Youth*	Ancillary
Latino Lifestyles to Prevent Type 2 Diabetes	Ancillary
Community Based Diabetes Prevention For Latino Adolescents*	Recall
Quality of Life and Physical Activity in Latino Youth*	Recall
Factors related to obesity in Latinos living in the Phoenix area*	Recall
Experiences and Perceptions of Adults of Mexican Origin Newly Informed of Having Hyperglycemic Values that Exceeds the Threshold of Diabetes*	Recall
Diabetes Management and Care Seeking Practices among Recently Diagnosed Adults of Mexican Ethnicity County, AZ*	Recall
Effect of Pioglitazone on Mitochondrial Function in Muscle and Adipose Tissue in Humans	Recall
Molecular Mechanisms of Insulin Resistance in Skeletal Muscle	Recall
Serum Vitamin D and Glucose Metabolism in Obese Latino Youth*	Repository
Novel Markers of Cardiovascular Disease Risk in Latino Adolescents*	Repository
Osteocalcin and Insulin Resistance in Latinos	Repository
Assessing Cardiovascular Disease Risk Factors in Overweight and Obese Mexican‐American Adults*	Repository
Insulin Sensitivity and Insulin Secretion across the spectrum of Glucose Tolerance in Latino Youth and Adults*	Repository
Cognitive Health and Physical Activity in Latino Adults	Repository
Common Genotype Associations with Diabetes and Cardiometabolic Syndrome Traits*	Repository
Whole Blood Gene Expression Profiles of Mexican Americans in the Arizona Insulin Resistance (AIR) Registry*	Repository
Adiponectin Receptor 2 Genotype Associations with Lipid Measures in the Arizona Insulin Resistance (AIR) Registry*	Repository

Note: Ancillary studies are those that collected data simultaneously to the parent study.

Recall studies are those that recontacted participants based upon demographic or phenotypic characteristics.

Repository studies use stored biospecimens in conjunction with collected demographic or phenotypic data.

*Denotes student, trainee, or junior faculty‐led study.

Discussion

Reducing obesity‐related health disparities is a public health and research priority that remains a challenge. In particular, Latinos are the fastest growing minority group in the United States,2 are disproportionately affected by obesity and type 2 diabetes, but are underrepresented in biomedical research and clinical trials. In response to these growing challenges, new models for conducting biomedical research have been proposed to address the cross‐cutting topic of health disparities.16 These models encourage transdisciplinary teams to work collaboratively on this complex issue with the hopes of decreasing the time required to translate research findings into meaningful health benefits for society.17 Finally, operationalizing new research models in order to close the health disparities gap will require investigators to engage with the community to focus on health promotion and disease prevention.18 With these principles in mind, we set out to develop the essential infrastructure to focus on obesity‐related health disparities among the Latino community of Phoenix, Arizona.

We engaged key community stakeholders early on in the process to establish the necessary rapport and develop a better understanding of the potential barriers to working with the Latino community. Stakeholders were diverse with strong ties in the local Latino community and included nonprofit organizations, health agencies and clinics, and local health coalitions. These formative meetings with stakeholders were instrumental to the overall success of the project that enrolled nearly 700 individuals over a 2.5‐year period. In addition to the essential input at the onset of the project, these relationships led to resources for the research participants, many of whom were identified with hyperglycemia/diabetes through the project. Although overt diabetes was an initial exclusion criteria, over 45% of the adults enrolled exhibited some degree of hyperglycemia and 15% met the criteria for diabetes. These individuals were provided with their lab results along with a list of community clinics and agencies that offered low‐cost/no‐cost medical care. Unfortunately, resources were not available to follow‐up with the vast majority of these participants to assess whether care was sought. However, two of the ancillary studies are focused on the perceptions and experiences of participants who were identified as hyperglycemic as they sought care in the community.

Early identification via screening and access to care are thought to be two critical components to addressing diabetes health disparities in underserved communities.9 Therefore, population‐based studies that incorporate laboratory screening that are likely to render results that require follow‐up should consider formal integration of care or counseling when returning these results to participants. A great deal of recent attention in the literature has focused on the topic of returning research results to individuals who participate in similar types of studies. 19 The current state of the science is lacking consensus in terms of common terminologies 20 as well as policies, procedures, and best practices for returning results.21 Moreover, much less is known about the perceptions of minority communities regarding return of research results.22 As population‐based biobanks and research registries continue to gather momentum as important pieces of research infrastructure, the resources and support for the social, legal, and ethical aspects of these projects should mirror the importance of the biological and genomic components.23

In addition to the focus on Latinos, a second novel aspect to this project is the inclusion of children and adolescents. The purpose for including a pediatric component was driven by both investigator interest and community input. From a research perspective, it is clear that obesity and the related health consequences run in families.24, 25 Therefore including multiple generations within family units helps to strengthen the heritability estimates that could be generated for future studies. However, intergenerational genetic studies may be better served by recruitment strategies that employ a proband‐family design. Nevertheless, the prevalence of both obesity and diabetes risk in this study is apparent and extends to the children. Therefore several of the repository studies focus exclusively on the pediatric population and one of the recall studies was to develop and test the effects of a diabetes prevention program in Latino adolescents.26 The need for early intervention programs in children is highlighted by the degree of obesity and prediabetes observed in this age group. The prevalence of overweight and obesity among the children recruited into the study was 50% and almost 20% exhibited some degree of hyperglycemia. These numbers support recent estimates by the CDC that up to 50% of Latino youth born in the year 2000 will develop type 2 diabetes in their lifetime.27

In addition to the scientific rationale, familismo or the emphasis on the family represents an important cultural construct within the Latino community.28 Although the study initially focused on individuals ≥18 years of age, both our community partners as well as the adults recruited requested that we expand the inclusion criteria to include children and adolescents. Special considerations were made in order to accommodate the enrollment of minors into the project. In addition to needing to obtain assent and using age appropriate language during the process, we opted not to perform the OGTT in those under the age of 12 as the increased risk associated with the blood draw schedule did not outweigh the potential benefits. Given the possible challenges with conducting blood draws in children, we employed nurses and phlebotomists with pediatric training and experience and provided age‐appropriate videos and movies to keep children entertained during their visit.

Limitations, lessons learned, and missed opportunities

Despite our attempts to develop a robust research infrastructure that met the needs of researchers and addressed a primary health concern in the community, there are limitations and several learning opportunities that are worthy of comment. First, the data gathered are not meant to be representative of the local community. Although the obesity and diabetes rates observed in our cohort are similar to national estimates in the Latino population, we do not know how closely they reflect the greater Phoenix area from which the sample was drawn. Second, we did not include provisions for regular follow‐up to update contact information of the participants. Therefore, the ongoing utility of the registry may be limited as many participants have since moved or changed phone numbers. In our attempts to re‐contact participants for follow‐up studies, we have found that those enrolled more recently were more likely to have viable contact information in the database. As such, regular contact with participants may be essential for keeping registries viable. Third, despite the formative work with community stakeholders to launch the project, we have not formally returned to present the findings to many of those who initially provided input. Closer integration with community partners and healthcare providers may facilitate ongoing engagement and help to ensure follow‐up medical care is received if needed. All participants were provided with the results of their laboratory tests and a list of community clinics and agencies that offered low‐cost/no‐cost medical services. In the event that treatment was needed (e.g., blood glucose levels in the diabetic range), participants received a call from the study nurse with instruction to see a healthcare provider as soon as possible. Due to limited resources and the fact that the project was not linked with a specific healthcare organization or provider, the inability to ensure follow‐up medical care represents a distinct limitation. Future studies should explore the use of community advisory boards to provide ongoing input, advice, shared governance, and facilitate access to community resources and care if needed.

The lack of bioinformatics support limited our ability to fully integrate methodologies used and data derived from ancillary studies into a common database. Integration of data could help strengthen existing collaborations and foster new transdisciplinary research activities. As mentioned above, collaborating investigators were identified primarily through word‐of‐mouth and included researchers from medicine, genetics, molecular biology, nutrition, exercise physiology, health psychology, social work, and nursing. Opportunities to encourage transdisciplinary research through a seminar series, journal clubs, annual conference, and/or pilot funding program may have further expanded on the number of investigators and disciplines involved.

We describe the design, personnel, methodology, and initial presentation of a cohort of Latinos from Phoenix, Arizona who agreed to participate in a community‐based research registry and biorepository to support biomedical research. The strength of this project includes the inclusion of a high‐risk minority population that is well characterized from a metabolic perspective and has provided consent for both re‐contact and for samples to be banked. Although the initial success of this project can be measured by the number of participants enrolled, the long‐term success is dependent on the number of future studies that are able to utilize this infrastructure to support health disparities research. Ultimately, it is our hope that facilitating research in this area with a high‐risk community will lead to important benefits for both scientists and the research participants through the development and implementation of health promotion and disease prevention strategies.