Recruiting diversity where it exists: The Alabama Genomic Health Initiative

Abstract

Lack of diversity among genomic research participants results in disparities in benefits from genetic testing. To address this, the Alabama Genomic Health Initiative employed community engagement strategies to recruit diverse populations where they lived. In this paper, we describe our engagement techniques and recruitment strategies, which resulted in significant improvement in representation of African American participants. While African American participation has not reached the representation of this community as a percentage of Alabama's overall population (26%–27%), we have achieved an overall representation exceeding 20% for African Americans. We believe this demonstrates the value of engagement and recruitment where diverse populations reside.

1 ∣. INTRODUCTION

Representative sampling is essential for good science. Failure to include sufficient diversity among genomic research participants will both deny some populations full participation in the benefits of research findings and inhibit richer understanding of variant-disease associations among all populations. Despite this, individuals from diverse ethnic backgrounds have been underrepresented or even absent in genomics research. Efforts to address this have been extensive, including funding prioritization, targeted efforts to diversify the scientific community, and broadening access to the limited data we do have inclusive of underrepresented populations (NHGRI, 2017). Despite these efforts, only marginal progress has been made toward diversifying the populations represented in genomic research data, and almost all of what limited progress has been made has occurred among participants of Asian ancestry (Auton et al., 2015; Lek et al., 2016; Popejoy & Fullerton, 2016).

In the United States, lack of representation for African Americans among genomic research participants is of particular concern. This population has historically been denied full participation in the goods of U.S. society: a phenomenon continuing to reflect in a number of recognized health disparities resulting in lower life expectancy and more numerous health problems among African Americans than their white counterparts–even after accounting for socioeconomic factors influencing health (Smedley, Stith, & Nelson, 2003). Many of these health disparities are attributable to historical segregation and ‘racially divided health care’ among other factors (Hammonds & Reverby, 2019), phenomena that directly inhibit African Americans' ability to benefit from advances in biomedical science and health care.

Similar disparities are inevitable in genomic medicine if diversity among genomic research participants is not achieved. Persons of African descent constitute 12%–14% of the US population; estimates of the global population of African descent range from 16% to 25% (the exact figure is complicated by the ambiguity of the African diaspora). Yet persons of African descent continue to constitute fewer than 4% of genome-wide association studies (GWAS) participants (Popejoy & Fullerton, 2016). Just as historical failures to include women, children, and minorities in pharmaceutical research denied those individuals an ability to fully participate in the goods resulting from such research (May, 2007; Society for Women's Health Research & the U.S. Food & Drug Administration Office of Women's Health, 2011), failure to include diverse populations among genomic research participants will deny some populations an ability to benefit fully.

Lack of diversity among genomic research participants has led to at least two major points of disparity in access to benefits from genetic testing and the interpretation of genetic variants: first, people from underrepresented minority populations are less often recipients of genetic testing that might benefit them; and second, individuals from diverse backgrounds who do receive genetic testing are less likely to receive informative results. Each of these disparities has a direct impact on clinical diagnoses and medical management.

Several studies have shown that individuals from diverse backgrounds are not appropriately referred for genetic evaluation and testing, even when they qualify based on medical guidelines. A ret-rospective study at four academic institutions showed that African American and Hispanic patients were less likely than white patients to be recommended for genetic evaluation or to undergo testing for Lynch syndrome, despite similar rates of colorectal tumor analysis (Muller et al., 2018). Furthermore, a study of 250 African American breast cancer patients showed 58.8% who met National Comprehensive Cancer Network (NCCN) genetic testing guidelines did not receive it as part of routine care. In addition, 4.8% of patients that did not meet NCCN genetic testing guidelines had a pathogenic variant (Ademuyiwa et al., 2019). This study demonstrates that a large portion of African American breast cancer patients are not receiving recommended genetic testing and indicates that the current breast cancer guidelines may not be sensitive enough for inclusion of African Americans at risk.

Lack of diversity also leads to disparities in the interpretation of genetic variant pathogenicity. A striking example of misinterpretation of genetic variant pathogenicity occurs when multiple false-positive genetic testing results are reported to individuals of African ancestry with hypertrophic cardiomyopathy as a direct result of not including African Americans in control cohorts (Andreasen et al., 2013; Manrai et al., 2016). More prevalent is the relatively lower detection rates of pathogenic variants and relatively higher detection rates of variants of unknown significance in underrepresented minorities (Caswell- Jin et al., 2018; Landry & Rehm, 2018; Raymond, Walker, Dave, & Dedhia, 2019).

2 ∣. THE ALABAMA GENOMIC HEALTH INITIATIVE

Addressing the above-described disparities through greater diversity in genomic research participation will require targeted efforts in areas where populations of African descent are well represented. In the United States, Alabama is well positioned to address this need, with an African American population constituting ~26%∓27% of the State’s overall population. The Alabama Genomic Health Initiative (AGHI) has specifically strived to achieve a representative population distribution among its population screening cohort.

Alabama Genomic Health Initiative is funded by the State of Alabama as a genomics research program to engage a diverse group of citizens from all 67 counties of the state. The overall goals are to develop a ‘genomics ready’ community, including both citizens and health providers; to utilize whole genome sequencing to establish diagnoses in children and adults with undiagnosed chronic disorders; and to establish the efficacy of a population-wide program to return results of medically actionable genomic variants to adult volunteers. Medically actionable genomic variants were derived from the American College of Medical Genetics and Genomics v2.0 gene list (Kalia et al., 2016). These genes are associated with a number of cancer/tumor predispositions, cardiovascular conditions, and metabolic disorders for which a potential for prevention or treatment exists (Table 1). At the time of enrollment, volunteers complete a health history questionnaire and can elect to participate in a biobank, have results shared with a healthcare provider, and/or be contacted about future research. Individuals with an actionable result receive a phone call from an AGHI genetic counselor to describe the results, implications, and next steps. Individuals that do not have an actionable result receive a negative result letter by mail.

Table 1.

AGHI population cohort gene list and associated medical conditions

Medical conditions	Gene list
Tumor predisposition Breast/ovarian, Li-Fraumeni, Peutz-Jeghers, Lynch, Polyposis, Von Hippel-Lindau, MEN1/2, Medullary thyroid cancer, PTEN hamartoma syndrome, Retinoblastoma, Paraganglioma/pheochromocytoma, Tuberous sclerosis complex, WT1-related Wilms' tumor, NF2	BRCA1/2, TP53, STK11, MLH1, MSH2, MSH6, PMS2, APC, MUTYH, BMPR1A, SMAD4, VHL, MEN1 RET, PTEN, RB1, SDHD, SDHAF2, SDHC, SDHB, TSC1, TSC2, WT1, NF2
Connective tissue dysplasia Ehlers-Danlos vascular type, Marfan, Loeys-Dietz, Familial aortic aneurysms and dissections	COL3A1, FBN1, TGFBR1, TGFBR2, SMAD3, ACTA2, MYH11
Cardiac Hypertrophic cardiomyopathy, dilated cardiomyopathy, Arrhythmia	MYBPC3, MYH7, TNNT2, TNNI3, TPM1, MYL3, ACTC1, PRKAG2, GLA, MYL2, LMNA, RYR2, PKP2, DSP, DSC2, TMEM43, DSG2, KCNQ1, KCNH2, SCN5A
Metabolic Hypercholesterolemia, Wilson disease, Ornithine transcarbamylase deficiency	LDLR, APOB, PCSK9, ATP7B, OTC
Pharmacogenetic Malignant Hyperthermia	RYR1, CACNA1S

In order to recruit a participant cohort reflective of Alabama's rich diversity, we recognized the need to address the fundamental challenges to recruitment of African American participants, challenges that Alabama is in a unique position to explore and address. Foremost among these is trust (Precision Medicine Initiative, 2015). Since the infamous U.S. Public Health Service Study in Nature of Syphilis, conducted through Alabama's Tuskegee Institute, all biomedical research involving African American recruitment has faced skepticism of purpose and suspected willingness to sacrifice the in-terests of study participants for broader societal and scientific goals. Although a 2005 study in which researchers conducted a telephone survey of African American and white adults in Baltimore, MD, indicated that knowledge of the Tuskegee study was not a predictor of trust of medical care (Brandon, Isaac, & LaVeist, 2005), and that race differences in mistrust likely stem from broader historical and personal experiences, more recent studies (Alsan & Wanamaker, 2018; Carroll, 2016) have found that ‘the disclosure of the study in 1972 is correlated with increases in medical mistrust and mortality and decreases in both outpatient and inpatient physician interactions for older black men’ (Alsan & Wanamaker, 2018). Memory of this shameful research study is still vibrant in Alabama, and we knew from the earliest stages that effective recruitment of a diverse and representative participant population would require significant attention to community engagement.

The Henrietta Lacks story may add a further layer of distrust of researchers among the African American population. One of the most frequently used research cell lines is HeLa cells, which were originally derived from an African American female cancer patient, without consent and without the knowledge of her family members and offspring. The lines of research conducted on HeLa cells may not necessarily be unethical perse, but the missteps and neglect of transparency has resulted in this story highlighting the priority of science over individual rights for many, especially in terms of concerns about genetic research on biological samples done without the knowledge or consent of African American participants (Lee et al., 2019). Genetic counseling, transparency, informed consent, and ongoing interaction, then, are key to maintaining trust that is established.

3 ∣. COMMUNITY ENGAGEMENT AND AGHI

One of the key goals of AGHI is to build strong relationships with key stakeholders in diverse communities in Alabama to evaluate recruitment, education, and communication strategies associated with the study. Toward this end, an Engagement Working Group was established that helps to serve as a liaison, communicating the unique needs of communities to the broader AGHI leadership team. The AGHI community engagement working group developed a multifaceted approach to achieve its goals; these efforts include the establishment of a community advisory board, town hall meetings, outreach events, and facilitated deliberative process (FDP) groups.

The AGHI community advisory board was created early in the study. AGHI leadership selected board members because they represented the community, clinicians, and health system executives across the state of Alabama. The board launched in the summer of 2017 and initially focused on reviewing participant recruitment strategies, providing guidance to improve visibility, understanding, and support for the AGHI throughout the state. Web conference meetings continue quarterly for the advisory board with an annual face-to-face meeting each year. On average, 6/15 board members attend the meetings. Most members reported not being able to attend due to family or professional obligations. Since the board was created, four members were replaced due to inability to fully participate. Guidance received from the advisory board has helped not only to improve participant recruitment efforts, but also resulted in revision to print and online recruitment materials, and helped to simplify the return-of-result process. In year three of the project, efforts focused on how to increase enrollment of African American participants in order to mirror the current Alabama demographics by race. Members provided broad recommendations as well as contacts to facilitate AGHI engagement with faith-based organizations and outreach at community health fairs.

A second engagement strategy was use of the facilitated deliberative process group meetings. In these groups, we used methods of democratic deliberation to engage stakeholders in education, discussion, and public opinion (Carman et al., 2015). As it relates to AGHI, participants in these meetings were provided education on the project, time for clarifying questions, opportunity for open discussion, generation of priorities, and individual voting on the action items determined through group discussion. In each group, the facilitators presented questions of importance to the researchers and potentially to community members, such as ‘What results from genome sequencing would you and other members of your community want to receive as a part of AGHI?’ Other questions considered in the AGHI groups focused on methods of delivering genome sequencing results and approaches to reach underserved populations. Participant recommendations and feedback were recorded using video and audio and printed on newsprint sheets posted in the meeting room. At the end of these meetings, participants were given an opportunity to edit the recommendations listed on newsprint and to vote on the action items or priorities that they thought AGHI should adopt.

Facilitated democratic deliberation was chosen as an engagement strategy because it is intended to encourage partnership between community members and researchers, it has the potential to make the research goals and methods more transparent to communities, and it is thought to build community trust in the researchers and in the research process. Facilitation is used to engage everyone participating in the meeting, contributing to joint decision-making between stakeholders and researchers and allowing stakeholders to provide and consider diverse perspectives on questions of importance to communities and researchers.

Three of these large meetings were held in cities that spanned the state (Huntsville, Birmingham, Selma), reaching over 90 participants representing community members and leaders of businesses, faith- based organizations, and advocacy groups, clinicians, and health system executives. Potential participants were invited to each group with the goal of representing diverse backgrounds and those underrepresented in biomedical research. Because the early AGHI enrollment demographics did not represent the state, the results from one of these meetings that raised this topic were particularly informative. The top ten recommendations from this meeting are listed in Table 2. Several of these recommendations were implemented in further engagement and recruitment efforts (detailed below).

TABLE 2.

Facilitated deliberative process group recommendations for reaching underrepresented populations in Alabama

Top ten recommendations	Votes (#)
Pop-up enrollment clinics	17
Partner with other initiatives with similar methods	15
Station in public health department in all counties	14
Use Social Media (ex: FB, Twitter) to increase	13
awareness
Partner with community and healthcare organizations	12
Engage more hospitals	11
Partner with employers (e.g., factories, plants)	10
Offer as a part of annual check up	10
Television advertisement	6
Partner with schools (area colleges and universities)	6

In subsequent years of the AGHI study, the engagement working group leadership participated in health fairs hosted by churches and community organizations throughout the state. In addition, presentations were made to healthcare advocacy groups, professional organizations, and neighborhood associations. Digital marketing strategies were implemented to increase visibility about AGHI in social media, geotargeting locations around the state that were not strongly represented in the first year of AGHI enrollment.

4 ∣. PARTICIPANT RECRUITMENT

The population cohort includes self-selected adult volunteers, regardless of personal or family histories of any specific medical condition, with an ultimate goal of 10,000 residents in the state of Alabama. The AGHI Recruitment Working group was established and charged with strategically planning recruitment and enrollment activities. With the approval of the project steering and executive committees, the Recruitment Working Group sought to extend the enrollment opportunities throughout the state with the goal of having participants from all 67 counties with the additional goal of mirroring the state’s demographics.

Given the magnitude and complexity of this task, AGHI partnered with the established UAB Recruitment and Retention Shared Facility (RRSF). Multiple strategies were implemented, using innovative enrollment solutions, including multiple ways to enroll (walk-up, phone, and online enrollment appointment scheduling), hosting ‘pop-up’ one-day enrollment events in targeted areas of the state where participation was not yet represented, and by establishing four permanent enrollment clinics around the state. These strategies led to innovation in both location and process. Locations included four permanent enrollment clinics, a traditional research unit enrollment location, and one-day ‘pop-up’ events strategically located across the state through local community partnerships in each location. Process options included both walk-up access and appointment scheduling by telephone and online access.

The UAB RRSF has extensive experience in the recruitment and retention of research participants with emphasis on diverse populations. Given the challenges of participant recruitment, especially those from diverse population groups, UAB established the Recruitment and Retention Shared Facility (RRSF) as a core facility to serve UAB investigators to recruit and retain participants and patients in research studies with emphasis on the recruitment of minorities and elderly populations.

Once partnered, the AGHI Recruitment Working Group consisted of two recruitment leads, a study and data manager, study navigators and a phlebotomist. Experienced study navigators from diverse backgrounds perform the informed consent process and enroll all study participants. The working group was trained extensively by the AGHI Leadership, Bioethics and Genomic working groups to learn the charges of other groups and open communication between groups to improve the overall participant process as the initiative began and progressed.

To begin recruitment and in order to establish a more visible presences, AGHI was launched in UAB's largest clinic, The Kirklin Clinic of UAB Hospital, which serves patients from across the state of Alabama. Working with clinic leadership, the AGHI Recruitment Working group obtained physical floor space and laboratory space in order to conduct consenting and enrollment on site in a timely manner. This allowed consent and enrollment in a convenient location where patients and their caregivers, as well as employees at UAB and nearby businesses, were presented with the study enrollment opportunity. By locating in a large, busy multispecialty clinic, this model removes the necessity for a participant to travel to a traditional research clinic to enroll. This model requires no appointments, accepting ‘walk-ups’ at any time during the hours of enrollment.

In the first year of the study, enrollment was offered only at UAB clinics in Birmingham. Identifying additional locations to provide access to the study resulted in the selection of additional clinic sites in Huntsville, Selma, and Montgomery in year two of the study. These locations were selected because of both their geographic and demographic diversity. Enrollment at these new sites helped the study to better represent the state population. Modeled after the successful model at UAB’s Kirklin Clinic, this approach also proved effective. (Table 3).

TABLE 3.

Percentage of African American enrollees at different enrollment sites, 2017–2019

	UAB Birmingham	UAB Huntsville	UAB Selma	UAB Montgomery	Pop-up	Total
Number Enrolled	3,704	316	413	394	639	5,466
Caucasian	2,920	269	133	210	441	3,973
African American	509	20	271	174	150	1,124
AA%	13.74%	6.33%	65.62%	44.16%	23.47%	20.56%

The ‘pop-up’ enrollment clinic model was developed to be mobile and includes all components of the permanent enrollment clinics. This provides flexibility for enrollment in unique locations throughout the state. Offering convenient, well marketed enrollment opportunities (flyers and post cards) at a pop-up site versus UAB clinical-based operations has proven successful. Additionally, this model allowed the experienced team to travel to underrepresented communities to offer information and enrollment into the study. A successful example of combining engagement and recruitment was a town hall meeting in Selma the night prior to an enrollment event. A total of 37 participants consented and enrolled as a result, 24 (65%) of whom were African American. The AGHI's enrollment efforts are summarized in Table 4.

TABLE 4.

Enrollment efforts timeline

January 2017–AGHI Retreat with initial timeline for enrollment began. Initial discussions regarding submission, potential engagement, and recruitment strategies were discussed including potential launch date

April 2017–Decision made to enroll approximately 100 participants in a one-month pilot, then stop for review of workflows, participant feedback, etc

May 2, 2017–Training for navigators–including AGHI overview, Genetic Counseling Overview, Bioethics, and workflow and question and answer session

May 3, 2017–Town Hall Meeting at UAB Minority Health Research Center - Health Disparities Research Symposium - Recruitment Leadership was in attendance at this event

May 18-June 1, 2017–105 participants enrolled in the pilot, in just

2 weeks at The Kirklin Clinic (TKC)

Proven recruitment strategies of postcard mailings with mission, enrollment steps, phone number and email address, local presentations to communities, word of mouth, UAB Monday Mailing, Health Fairs, flyers in local community

February 2018—New recruitment Strategies: ‘On Hold’ Messaging at UAB, Social Media, Online Scheduler, Pop-Up Model

Multiple presentations to AGHI stakeholders who forms the

AGHI Community Advisory Board. Board members included healthcare providers, community members, and healthcare administrators from around the state

Stakeholder recommendations and feedback were incorporated into recruitment and enrollment process

October 2019–Cumulative population cohort enrollment cover All 67 counties

5 ∣. ETHICS

The AGHI Bioethics working group has been charged with ensuring the highest standards of ethical review and practices in all areas of the initiative, thereby promoting trust and trustworthiness of this community-engaging project. During the first year, the bioethics working group established four core values that are crucial for the ethical operation of the AGHI: Transparency/Honesty, Respect, Confidentiality, and Fairness.

Based on these core values, the bioethics working group elaborated on the informed consent process and the ways to maintain participant confidentiality. Also provided were ethical recommendations on participant health questionnaires, return-of-result letters to participants, and information pages on the AGHI website, as well as ethical consults concerning participant re-contact and the deletion of research data following participant withdrawals.

Bioethics training is offered to each and every patient navigator and new site coordinator as the AGHI expands its recruitment sites. This training consists of reflecting on reasons why trust issues are critical for genomic research projects especially in the state of Alabama, reconfirming the AGHI's core values, and going through, point by point, the key components of the informed consent doc-ument that every patient should be aware of during the consent process (e.g., which part of the initiative is of experimental nature, limitations to benefits, confidentiality risks, options that participants have for receiving results and participating in biobanking, among others).

6 ∣. RESULTS

The results of our recruitment efforts have been remarkably successful. To date, over 5,500 individuals have enrolled in the population cohort since enrollment commenced in May 2017, and the race/ethnicity data of participants can be seen in Figure 1.

FIGURE 1.

Population cohort enrollment demographics. Enrollment in the Alabama Genomic Health Initiative by Race, as of October 3, 2019

These numbers represent a significant improvement over GWAS numbers concerning African American representation, wherein fewer than 4% of participants are African American. Although representation of Asian ancestry is lower than GWAS representation (14%), and other minorities are similar or only very slightly improved over GWAS representation, we believe that this reflects the demographic make-up of Alabama itself; for example, Alabama’s proportion of Asian population is about a quarter of the national average, while African American representation significantly exceeds the national average. In this, it should be noted that while the 20% rate of African American enrollment exceeds the general U.S. population proportion of 12%–14% African American, it falls below the Alabama state population proportion of 26%−27%. Nonetheless, this represents significant improvement over GWAS representation of less than 4% African American participants and confirms the strategy of recruiting diversity where it exists.

7 ∣. RECOMMENDATIONS

Based upon our initiative’s success in recruiting African American participation in genomic research, we offer several observations and recommendations for other efforts aimed to recruit greater numbers of African American participants. First is to utilize facilitatedw Table 2, our own processes were able to identify a number of effective mechanisms which proved useful to our own efforts in these regards.

Chief among those strategies recommended were pop-up enrollment clinics, which ease logistical obstacles to enrollment by making enrollment information and opportunities more easily accessible to those populations targeted. In addition, taking such opportunities to the community in question raises awareness of both the opportunity, as well as the benefits of participation both for individuals and the community as a whole.

Second, it is important to note our initiative’s improved minority participant enrollment over the life of the study so far. AGHI has increased African American participant enrollment from 13% at the close of year one to 20% at the close of year three, with a steadily increasing rate of African American enrollment each year (Table 5).

TABLE 5.

Alabama Genomic Health Initiative African American participant enrollment by study year

2017	13.02%
2018	16.53%
2019	30.39%
Cumulative	20.56%

We believe the increase in rate of African American enrollment each year reflects long-term strategies to promote trust and increasing confidence in AGHI commitment to ethical conduct of research as the project continues forward. This suggests that time, patience, and ongoing effort can contribute to improved minority participation in research; a phenomenon that is centrally recognized within Community Based Participatory Research approaches that emphasize the establishment of long-term relationships with subject (partner) communities.

Study personnel responsible for facilitating informed consent should also be carefully considered. These are roles commonly performed by genetics professionals, specifically genetic counselors. The AGHI has an advantage in the specific resource of the UAB RRSF to staff recruitment sites with study staff of diverse racial and ethnic backgrounds. However, the genetic counseling profession is known to be extremely racially homogenous, with the latest Professional Status Survey estimating that 90% of the profession identifies as White/Caucasian (National Society of Genetic Counselors & Inc., 2019). This is certainly not a new phenomenon and dates back to the profession's inception when primarily white, middle-class females were initially trained as genetic counselors (Stern, 2009). This, coupled with the aforementioned difficulties in recruiting minority populations to genomic health research, presents an unfortunate intersection of historical trends that will require innovative approaches at both the genetics workforce development level and study recruitment protocol design. An increase in both the numbers and the diversity of the genetic counseling workforce is the ideal solution, but one that will likely require long-term solutions involving a variety of stakeholders. One possibility, as demonstrated here, is the adaptation of existing institutional resources designed to recruit and retain research participants across a breadth of specialties for use in genomics research. Genetic counselors can be utilized to assist in developing appropriate training for these personnel to address genomics-specific concerns. This may also serve to increase the efficiency of the genetic counselor(s) on the protocol.

Though our study did not specifically examine the effect of race concordance between participants and recruitment staff, this strategy is often utilized to increase minority participation. However, researchers have demonstrated that establishing trust with communities, relationship-building, and culturally sensitive communication can have a greater impact on successful minority recruitment (Frierson, Pinto, Denman, Leon, & Jaffe, 2019; Fryer et al., 2016; Kraft & Doerr, 2019). Given that many of these concepts are central to genetic counseling practice-based competencies (Accreditation Council for Genetic Counseling, 2019); this presents an opportunity for the profession to highlight these skills when assisting in the recruitment of diverse populations.

8 ∣. CONCLUSION

Targeted efforts to increase diversity among genomic research participants have been successful over 3 years of the Alabama Genomic Health Initiative. Concerted efforts at community engagement, use of local and regional recruitment expertise within areas highly populated by targeted groups, and a comprehensive emphasis on ethics to build trust, have all combined to see enrollment of African American participants that approaches a representative population. Although African American enrollment to this point is slightly below state population demographics, it exceeds national demographic proportions by a similar amount and represents a very significant increase over prior African American enrollment in genomic research studies (by more than 500%).

We believe this success demonstrates that increasing diversity among genomic research participants is possible. We believe it also argues for recruitment strategies that seek to increase minority enrollment where this exists, both at a regional level and in terms of enrollment opportunities, like our own ‘pop up enrollment clinics’, that bring such opportunities to the populations targeted. To ensure the success of these recruitment strategies via community engagement, we also advocate for thoughtful consideration of the diversity of the workforce responsible for recruiting and retaining minority participants in research.