Return of Participants' Incidental Genetic Research Findings: Experience from a Case-Control Study of Asthma in an American Indian Community

Purpose: The proper communication of clinically actionable findings to participants of genetic research entails important ethical considerations, but has been challenging for a variety of reasons. We document an instance of return of individual genetic results in the context of a very rural American Indian community, in hopes of providing insight to other investigators about potentially superior or inferior courses of action. Methods: The original study was a case/control study of asthma among 324 pediatric participants. The study utilized a genotyping microarray assessing over 2 million variants, including one conferring risk for hypertrophic cardiomyopathy for which the American College of Medical Genetics recommends return of results to participants. The study investigators engaged in extensive consultation with the IRB, the Tribal government and local clinicians to better inform our approach. Results: With some difficulty we were able to notify the 2 participants heterozygous for this variant. One participant welcomed this information and proceeded to obtain further clinical work-up; the other participant declined further follow-up. Conclusion: While demanding of considerable time and effort, the return of clinically actionable genetic results is important from both an ethical perspective and to provide an improved trust relationship with the community of research participants.


Introduction
The return of individual, incidental genetic ndings to research participants has been controversial 1,2 and uncommonly undertaken until recent movement in that direction. 3,4 Reluctance on the part of investigators to commit to this practice derives from a number of valid concerns, such as lack of Clinical Laboratory Improvement Amendments (CLIA) certi ed results, frequent uncertainty as to the clinical implications of a particular genetic variant, lack of long-term grant funding to maintain contact with participants, and often insu cient support to provide adequate communication and counseling. 5 In counter balance to these concerns is the inherent wish to provide potential bene t to individual participants who generously engage in research, a desire to prevent harm to a participant that is unaware of a potentially remediable genetic risk, and a respect for the autonomy of participants who wish to have information they feel may be useful.
In routine clinical practice, it is common to encounter "incidental" information that was not sought in the diagnostic or treatment process. This information is routinely conveyed to the patient for consideration of additional diagnostic or treatment options. Indeed, failing to inform the patient of new information of a serious nature would be considered a breach of professional ethics. The American College of Medical Geneticists has provided guidance in this area as it relates to the provision of clinical care.
The research setting is not the same as a clinical encounter, however one might expect some overlap in the ethical evaluation of responsibilities. Certainly the research participant could be forgiven for assuming a certain degree of congruence, even though this issue may have been addressed in the ne print of the consent. To our knowledge, no research professional organizations have produced ethical standards to guide investigators on this issue. The National Human Genome Research Institute funded eMERGE consortium is an effort to utilize genome scale screening to further both genetic research and direct patient care goals. 6 This naturally addresses issues inherent in return of "incidental" genetic results.
In the current report, we report our experience in screening the MEGA EX genotyping array (Illumina Inc.) for clinically actionable genetic variants from a case-control study of genetic and environmental in uences on the risk of asthma in an American Indian pediatric population. 7 We also describe the response of 2 participants to the incidental ndings provided, and outline some of the di culties inherent in this process.

Materials And Methods
Population-based study participants were ascertained through an unbiased query of the Indian Health Service (IHS) electronic medical records system for an array of diagnostic codes suggesting asthma. The Factors In uencing Pediatric Asthma (FIPA) study identi ed potential asthma cases and matched controls, from whom consent was requested for review of medical records to identify 108 cases and 215 controls that met study criteria.
As with all research involving IHS facilities, FIPA was approved by the regional IHS institutional review board (IRB) and the Tribal government, as well as the funding institution's IRB. The consent forms advised that it was unlikely that any individual genetic test results would be returned, but if potentially useful results were developed, the study team would ask if they wished to be informed. The consent also stated that "…only genes that might be involved with asthma." would be tested. Participants were told in lay terms that study genetic tests were not performed in Clinical Laboratory Improvements Amendment (CLIA) approved facilities and that potential test results of clinical signi cance would probably need to be repeated in a "clinical lab" for con rmation. The consent also advised that payment for additional testing would not be available through the study. An explicit "opt-out" option was not provided in the consent.
After completing active enrollment and execution of an interventional trial in the FIPA study, we pursued genome-wide variant genotyping as a cost effective means to identify genetic associations with the many hundreds of variants and genes previously implicated in asthma and asthma-related traits. Prior to this, we had genotyped ten candidate variants and con rmed some, but not all of the documented associations with asthma identi ed from previous studies. Before expanding our genetic studies to a genome-wide variant array, discussions as to whether and how to proceed were held with Tribal members on the research team, local clinical providers, the IRB and Tribal government. The Tribal government approved the transfer of su cient DNA to the University of Colorado for genotyping on the Illumina MEGA-EX microarray including ~ 2.3 million variants. Due to the large number of variants tested, it was anticipated that incidental ndings of clinical signi cance could result. Incidental ndings were determined from the American College of Medical Genetics (ACMG) recommended clinically actionable variants for use in return of secondary results in the context of clinical diagnostic evaluation. 8 Plans for handling this possibility were developed and further consultations began with the same stakeholders as described above. Initially it was suggested that community messaging inform both the general public and resident participants that FIPA now had extensive genetic results, and that there may be participants with clinically actionable results of importance. The rationale was that this would not stigmatize any individual with direct contact from research staff for what could be presumed deleterious information; but that as participants came forward to receive con dential communication about their genetic results, most would be told there were no medically actionable variants to report. The participants would also be re-consented to accept this information prior to communicating their results.
However, our proposal to reach out to all study participants was viewed skeptically by most clinicians, Tribal study staff and the Tribal government. The predominant suggestion was to simply contact affected individuals and ask if they wanted the genetic results returned, in line with expectations of what would occur in the case of an incidental nding in routine medical care. A major advantage of this approach is to avoid missing the very individuals most likely to bene t, if they either didn't receive the community message or didn't consider it su ciently important to contact the research staff.

Results
The demographic characteristics of the research participant community are important factors in fully understanding the challenges and bene ts of implementing the return of incidental genetic results in this setting.
Study participants reside in an area of northern South Dakota covering 4,266 square miles, with a population density between 2 and 3 people per square mile. Most Tribal members live in cluster housing near small towns, on farmsteads or in cluster sites far removed from basic services. Beyond Tribal government and Federally supported work in health care and education, ranching and farming provide the bulk of employment. Out of more than 3,000 counties in America, two of the three counties in this community have the 2nd and 24th lowest per capita income, 9 with 44% and 37% of residents having incomes below the poverty line. 10 More than 15% of adult residents have less than a high school education. 10 Although no "opt-out" option related to return of individual results was provided at the time of consent, there were no potential parents or participants that indicated an unwillingness to enroll due to concern over this issue.
The predominant source of medical care for tribal members in this community is the Federally funded Indian Health Service and a tribal health department. The IHS hospital and clinic are typically staffed by 3 or 4 primary care physicians (often including a pediatrician) and an equal number of mid-level providers.
The Tribe operates an independent clinic with a couple of physicians and a similar number of mid-level providers. The closest specialty care is at least 90 miles and more commonly 170 miles distant (e.g. cardiac or pulmonary consultation). Although all recognized Tribal members are eligible for care at the IHS facility, in order to be paid for with IHS funding, any non-IHS care must be pre-approved and is often deferred or delayed due to lack of funds or insu cient priority.
Both the community at large and individual participants were provided information on the aggregate results of FIPA. This was accomplished by a newsletter to all participants including reference to a YouTube video created conveying our results, a radio broadcast, presentations to both the Tribal council and local clinic staff, as well as Tribal approval of each scienti c publication.
All genetic data was de-identi ed during quality control and analysis as to protect the identity of study participants. Standard quality control for genome-wide association studies was performed, including the removal of variants with low call rates (< 95%) and deviation from Hardy-Weinberg expectations. We also examined individual samples for unexpected genetic relatedness, low genotype call rates, and unexpected levels of heterozygosity (including heterozygosity of the X chromosome for consistency with reported gender). We then performed an intersection of variant identi ers with pathogenic variants listed by the ACGM, followed by visual con rmation of similar genome build, chromosome number, variant position, variant alleles (including strand), and genotype quality. We also compared overall allele frequency and heterozygosity among all FIPA participants for the identi ed ACGM pathogenic variants speci cally, to provide an extra level of more stringent quality control without an arbitrary threshold. Overall, two individuals were identi ed as having clinically actionable genetic variants in the present genetic study. The two were siblings, and both lived 40 miles from the IHS facility and research headquarters. The R870H variant is associated with hypertrophic cardiomyopathy (HCM) and the characterization as pathogenic is based on multiple clinical reports in the literature, from about 1995 as documented in dbSNP citations for rs36211715. 13 Since this variant acts in a primarily dominant fashion (with penetrance of ~ 60%), 14 evidence of pathogenicity is derived mainly from intra-family linkage 15 and is at very low prevalence in the general population (frequency of the pathogenic allele is 2.3 x 10 − 5 in TOPMED 13 ). The phenotype is characterized by otherwise unexplained left ventricular hypertrophy and increased risk of asymmetric septal hypertrophy and ventricular arrhythmias, occasionally leading to sudden cardiac death. [14][15][16] Prior to contacting participants, a plan was needed to offer independent, no-cost CLIA certi ed tests to con rm the presence of ACMG variants, however IHS does not have a mechanism to fund clinical followup of incidental genetic ndings in a research setting. Therefore, we committed to provide discretionary research funds to obtain the needed clinical con rmation should the participants consent. Following this an attempt was made to contact the individuals by phone. Due to unstable cellular connections at the participants' home and the COVID-19 pandemic, there was considerable di culty making initial contact.
However, one of us (LGB) was able to visit with one of the siblings and their father for about 20 minutes and the preliminary results and their basic implications were explained. During this session, the uncon rmed nature of our original results was emphasized, the general characteristics of the phenotype were outlined and a release of information was obtained to review potentially pertinent medical records related to HCM. The participant, who was > 18 years of age at the time of contact, agreed to proceed with con rmatory testing. The Mayo Clinic was the most convenient reference genetic laboratory and arrangements were made to have a new blood sample collected at a local, private clinic. The original results were subsequently con rmed by the Mayo Clinic and the variant was characterized as "pathogenic".
After con rmatory results were available, a video teleconference was conducted with 3 of the authors to discuss the results and implications of this variant more fully. It was recommended (LGB) that the participant have a cardiology consultation and 24 hour ECG recording. Efforts then began to obtain referral from IHS for these services, the results of which showed no current, anatomic evidence of HCM and no arrhythmogenic concerns. Our nal contact with this participant was a written statement reiterating the ndings from the cardiology consultation, the offer to provide additional testing for family members, the fact that the phenotype has limited penetrance, and strongly encouraging regular follow-up of blood pressure, repeat echocardiogram every few years and general lifestyle recommendations for cardiovascular health. Contact information was also provided with encouragement to reconnect at any time.
Concurrent with above, multiple efforts to communicate with this participant's sibling were unsuccessful, including efforts on the part of the participant we returned results to. A parent of the siblings was present intermittently during the initial 20 minute phone consultation and did not decline offered testing for themself, but has not attempted to follow-up and make arrangements for sample collection. The communicating participant has limited contact with their other parent and testing was not pursued. The full process, from the time the genetic variant was identi ed through our last contact with the family, lasted approximately 21 months.

Discussion
We present here a case study illustrating some of the di culties in the return of incidental genetic ndings in a research setting, and in a unique community context. Although consensus appears to be building that medically actionable research genetic ndings should be conveyed to participants, 6,17,18 many of the justi ed concerns of investigators and ethicists about implementation were encountered in the current study. None-the-less our efforts were eventually successful to the degree that one participant wished to engage. Some of these challenges are described below.
Participants' desire to learn of individual genetic results from research participation is not universal, although the evidence accumulated to this point seems to indicate more acceptance than many anticipated. [19][20][21][22] Indeed, participants often reject their rst choice to eschew return of results when given a second opportunity. 1 One of us (LGB) with personal experience obtaining informed consent from over 800 research participants who were provided the option of allowing return of actionable results found only 2 opting out. 23 While investigators have often expressed support for return of aggregate study results, their actual performance has not been consistent or extensive. 24,25 In regard to return of individual research results, investigators have typically been more concerned about the desirability and challenges inherent in this process. Speci c concerns have been related to return of results unauthorized by the participant, di culties determining which variants are "clinically actionable", potential long-term responsibilities of investigators and possible lack of subsequent funding, as well as uncertainty about return of results to children [26][27][28][29] The development of acceptable ethical standards in this context began with consideration of whether and how best to return incidental ndings uncovered during clinical investigation of conditions with a likely genetic origin. The increasing use of broad-based genetic testing, such as genotyping arrays and whole genome sequencing has resulted in a great number of unanticipated results, not directly associated with the primary diagnostic problem. The rst professional body to provide guidance was the American College of Medical Genetics (ACMG) in 2013. 30 This recommendation speci cally listed 56 genes with multiple, likely pathogenic variants that should be communicated to the individual via standard genetic counseling. A number of medical facilities have adopted congruent policies and research is in progress to assess the impact of return of genetic results in both a diagnostic realm and a more general screening approach. 6,19,20,31 Although many governmental consultative groups have suggested that similar standards should apply, at least in certain circumstances, to the return of results to research participants, 17,32−36 to our knowledge no professional body has promulgated detailed standards similar to the ACMG. 37 Institutional review boards have, not surprisingly, taken a cautious approach to the problems associated with return of genetic research results. Many have recognized the key role that IRBs should play in guiding appropriate return of results, 17,18,38 and the valued principle of participant autonomy is often raised in these deliberations. 39 None the less, Wolf and Evans 27 raise serious concerns related to some recommendations to empower IRBs to set standards on an institutional or study-by-study basis.
In the present study, the participant consents were permissive for return of potentially useful information and considerable effort was expended to garner input from the wider community of Tribal leaders, local clinicians and others as suggested by Raymond et al. 40 We are grateful for this community-grounded input and feel that our initial implementation plans would likely have failed, at least to some extent. It is also fortunate that we had the bene t of supplemental research funds to allow CLIA-certi ed con rmation of our results, so this issue was easily resolved.
Although we encountered only two participants that met our quali cations for return of clinically useful results, just one of these (within the same family) was accepting of this information and utilized our recommendations. In spite of clearly informing one of the parents involved, there has been no further indication of interest in additional testing within the family. In this study, we were fortunate to identify a very small and manageable number of participants qualifying for return of results. Had the number identi ed been signi cantly larger, we may not have had the resources to follow our intentions. As it was, the challenges of geography, communications, health care nancing and sparse medical resources were substantial.
It is the authors' strong conviction that return of clinically actionable results, whether genetic or not, is a moral obligation that cannot be overlooked, regardless of the many practical impediments it poses. In the United States, the watershed event precipitating a major reassessment of research ethics, the US Public Health Service Tuskegee Study, primarily failed in its duty to inform participants when treatment became available (ie "clinically actionable"). It is unfortunate that many in the biomedical research establishment continue to put more emphasis on practical impediments than on the underlying moral imperative.
While the experience of this study is limited in size and scope, we feel the unique community context provides useful additional information about the bene ts and challenges related to what we feel is an extremely important ethical issue, namely the return of clinically relevant genetic information to research participants.

Declarations
Disclaimer: The opinions expressed in this paper are those of the author(s) and do not necessarily re ect the views of the Indian Health Services (IHS). American Indian owned organization, which conducts research and provides health care as a durable medical equipment supplier to rural populations. MBIRI's mission is to promote clinical research that will enable communities and individuals to choose healthier lifestyles through support, involvement, and education, while networked to the rural, tribal, scienti c, medical, and psychological communities. MBIRI is a quali ed HUBZone small business concern (SBC), is buy Indian certi ed and is considered a business under NAICS code 541715. Although MOL and ROL have proprietary interest in MBIRI, their contribution to the manuscript was limited to provision of data and editorial comments. MBIRI otherwise played no role in the conception, design or analysis of the original study this report is based on; nor on the conception, analysis or writing of the current manuscript describing follow-up activities of the original study. Dr. Torgerson received no compensation for either the original study this report is based on, or the preparation of this manuscript.
Funding Source: National Institute of Minority Health and Health Disparities, grant U54 MD008164.
Author Contributions: LGB designed the primary study this manuscript relates to, developed and supervised the return of results process, and is the primary author of the text. DGT provided independent investigator funds for the microarray genotyping and important text edits. MOL, ROL, and WL assisted with conduct of the study and were instrumental in developing plans for and conducting the return of results.
Data Availability: The data underlying the results of this study are owned and controlled by the Cheyenne River Sioux Tribe that approved its collection. This fact is clearly stated in the Tribal resolution authorizing the research; and it must be recognized that this Tribal community is an independent, sovereign government, in control over research activities within their borders. Access to data and materials is accomplished by application to Dr.
Lyle G. Best, who will arrange for further consultation with the appropriate Tribal o cial. Approximately 2 to 3 months may be required. The authors received special access privileges to the data due to their relationship with the Tribal Government, however, interested researchers who apply for data access will be able to access the same data as the authors.