Reporting Genetic Results in Research Studies

Introduction

Studies of heart, lung, blood and sleep disorders routinely include genetic tests to identify new genetic risk factors in the population and the inclusion of these tests is increasing rapidly. These findings may provide opportunities for early detection of disease and presymptomatic diagnosis which in turn can provide opportunities for successful treatment and/or prevention. Research studies have an obligation to report research findings of definitive clinical value to study participants when the potential benefits of such information outweigh the potential harm. Although results of genetic tests have considerable potential for risk assessment and appropriate targeting for preventive strategies, genetic tests usually do not predict the development and severity complex diseases. Furthermore, psychological and social harm as well as financial costs may result from introducing information to research subjects and their families about diseases that cannot be prevented or treated.

Even when a genetic mutation conferring increased risk is present, there may be other factors such as the interaction with other genes, variation in exposure, and population stratification, that may make disease risk uncertain. The clinical validity of genetic tests is also affected by small and potentially biased study populations, low penetrance, variable expressivity, lack of understanding of phenotypic modifiers, and ambiguous clinical endpoints. Genetic results are likely to vary in their potential to direct prevention and treatment, and in personal and social consequences. As a result, the task of determining appropriate transmittal of genetic results to research subjects will require careful consideration of a variety of factors, including the analytic validity, clinical validity, clinical utility, and ethical, legal, and social implications of the results.

The goal of population-based genetic research is ultimately to identify genetic variants that indicate increased risk of disease or disability with particular attention to those conditions where risk can be reduced. Although relatively few such variants may currently be identified with certainty, it is hoped that the extensive body of ongoing research in this area may detect many more of such variants. The existing literature has no clear statement on reporting such results to participants in research studies, independent of whether public health or population screening measures are eventually implemented. Thus, there is a need for a consensus to identify genetic findings that have met an acceptable threshold for individual results reporting to study participants.

The NHLBI Working Group on Reporting Genetic Results in Research Studies was held July 12, 2004 in Bethesda, MD. Working group members included experts from scientific, medical and public health communities, and persons with expertise in ethical, legal, and social issues. The main objective of this working group was to discuss and make recommendations for reporting individual results from genetic tests to participants of Heart, Lung, Blood and Sleep research studies involving genetics. The meeting began with welcoming remarks from the chair of the working group, Dr. Russell Luepker, and presentations by each of the working group members.

Presentations

Dr. John Eckfeldt provided information on the Clinical Laboratory Improvement Act (CLIA) and its later Amendments and clinical laboratory regulations generally. The original CLIA regulations (1967) applied only to labs involved in interstate commerce. In 1988, the Clinical Laboratory Improvement Amendments expanded the authority to all laboratories testing human samples for patient care regardless of where testing is done. The following labs are excluded from CLIA: research laboratories that do not report individual results to participants for diagnosis, prevention, treatment, or assessment of health, disease or impairment; National Institute of Drug Abuse (NIDA) or Substance Abuse and Mental Health Services Administration (SAMHSA) labs performing employee drug testing; federal laboratories (VA, DoD, etc.); and forensic labs. CLIA applies to all samples drawn from any patient within the U.S.; if samples collected in the U.S. are sent outside the U.S. for testing, CLIA regulations still apply.

There are about 75,000 laboratories with CLIA certification, of which 50,000 are physician office laboratories. CLIA regulates the management of specimens and the testing process including specimen handling, requisitions, records, reports, and the referral of specimens, quality systems, proficiency testing, and personnel standards (www.phppo.cdc.gov/CLIA). The principal sanction for violating CLIA is the suspension, limitation, or revocation of a laboratory’s CLIA certificate and suspension of all Medicare/Medicaid payments. Secondary sanctions are a directed plan of correction and on-site monitoring at the laboratory’s expense; civil suit and monetary penalty; and/or criminal sanctions for any individual who is convicted of intentionally violating any CLIA requirement. Laboratories have improved quality control procedures since CLIA was implemented. The CDC is currently heading a working group to amend CLIA regulations for genetic testing.

Dr. Kathleen Cranley Glass gave an overview of the ethical issues related to reporting individual genetic results to research participants. These include: the unchangeable nature of accurate personal genetic information which affects the individual, family and community; the moral obligation to protect research participants; the participants’ comprehension of the purpose of the study, the nature of their participation and the genetic results, including the potential for misinterpretation or exaggeration of the meaning of the results; the risk of discrimination (individual and group); the difficulty of drawing health/life choice conclusions from early study results or incompletely understood data; and the potential psychosocial effects caused by knowledge of genetic results. Dr. Glass also stressed the ethical implications of scientific validity and clinical relevance of genetic tests and the consequent importance of their careful evaluation before reporting genetic research results. In addition, pre- and post-test genetic counseling should be available for participants who will receive personal results.

Dr. Gail Jarvik discussed reporting of genetic research results in studies of complex disease. A genetic variant’s effect is modified by other genes and the environment, thus are often weak and may not reliably predict disease. Dr. Jarvik remarked that the odds ratios for most genetic associations are not more than 2 and many associations are not replicated in other studies. Note that an odds ratio of 2 means a doubling of disease risk as compared with healthy controls. The significance (absolute risk) of an odds ratio or relative risk of 2 or more will depend on the frequency of the disease. Thus, even if a relative risk in the test population is very high (i.e. 100), the actual or absolute risk of a condition with a frequency of 1/10,000 in the general population would only be 1/100. Dr. Jarvik emphasized that individual genetic results should only be reported when there is proven accuracy and clinical utility. Currently, Institutional Review Boards and individual genetic studies held primary responsibility for deciding when disclosures should be made. She stressed the importance of standardization or IRB rules for genetic/family studies, including results disclosure. She also noted that informed consent should always be requested with an option to opt in/out of receiving information and that the person disclosing the genetic results should be trained to provide genetic counseling.

Dr. Michael Klag gave an overview of perspectives in reporting genetic results in his research studies and suggested points to consider when deciding to communicate genetic results. Two examples were given, one of participants not asking for genetic results and the other of a participant writing a letter requesting that he be informed of genetic results even though he understood that the clinical utility was unproven and clinical judgments on the basis of the results would be questionable. Dr. Klag suggested that the disease risk, the complexity of the trait, the degree of penetrance, age of onset, disease severity, reproductive implications, and availability of therapeutic or preventive interventions be considered when deciding to report genetic results to individuals. Dr. Klag also suggested that a written document explaining the implications of the results be provided to the participant and/or to the participant’s physician to assist in interpreting risks for the participant and family members. He noted that IRBs in his institution are leaning toward the view that, in some situations, disclosure of non-CLIA certified results will be more ethical than non-disclosure.

Dr. Gregory Koski gave an overview of the decision-making process of the Framingham Heart Study’s Ethics Advisory Board in deciding whether to report research results to participants. He explained that many policies are designed to protect only the individual, but there were repercussions for the family as well. He noted that non-genetic tests and genetic tests have similar implications for risks to the participant, but the genetic information is often considered to be unique in that it also provides information on parents, offspring, siblings and the community. He also noted that genetics is often considered to be unique because of its predictive value but often such predictions are far less definitive than those obtained by more routine clinical tests. Currently there are social and legal reforms in the area of genetic discrimination that may simplify the interpretation of risks and benefits of reporting.

Dr. Russell Luepker discussed the rationale in reporting genetic research results to individuals including the unique complexities of population studies, ethical considerations and the negative effects of reporting results. Dr. Luepker gave suggestions of items that should be included in the informed consent form before genetic results may be reported and reiterated that analytic validity, clinical validity and clinical utility should be taken into account before deciding to give genetic research results back to the participant. He also talked about the huge impact that the media hype surrounding the human genome project has had on the public and its understanding of genetic determinism. Dr. Luepker agreed that proven interventions need to be available before a specific genetic test is reported to the participant.

Dr. Arno Motulsky discussed the ambiguous meaning of ‘genetic’ results and general considerations when deciding to disclose individual results in genetic studies. He explained that many tests other than DNA tests may provide results that portray the underlying genotype. Examples of these tests include tests for hemoglobinopathies (HbS), enzyme deficiencies (G6PD), and clotting protein abnormalities (hemophilia). He also explained that a test generally considered to be non-genetic such as a cholesterol level usually has strong genetic determinants in addition to environmental factors (diet). Dr. Motulsky noted that over 1600 Mendelian traits with a definitive molecular basis have been found, yet only about 10 genes had been detected for human complex disease traits (Glazier et al. 2002). He suggested that the magnitude of the research study, the type of contact with the study subjects (direct or indirect), the certainty of the new information affecting health and disease, the qualifications of the investigators, and the availability of referrals to outside physicians be considered when deciding to report genetic results to research participants. He also noted that informed consent before testing should allow participants to decline to receive results.

Dr. Benjamin Wilfond gave his perspective on the criteria for the discretionary decision of disclosing genetic results in research studies. He described situations in which disclosing genetic research results would be prohibitive, discretionary, and obligatory. Dr. Wilfond remarked that the relationship of validity to benefits and harms is complex when deciding to report results. Uncertain data can be beneficial because it can give information regarding risk although the uncertainty of results can also exaggerate harms. He noted that reasons for disclosing results include participants’ contribution to research and collaboration and trust of communities in which the research is performed. He also suggested that simply disclosing results may not be sufficient and that providing intervention or remediation may also be needed. Dr. Wilfond suggested ways to improve the benefit/harm ratio of result disclosure including consenting for disclosure/non-disclosure, ensuring analytic validity, communicating results effectively by including health care providers and follow-up support, and reviewing of the research protocol and disclosure guidelines by the study’s institutional review board.

Guidelines for Reporting Genetic Research Results

The working group unanimously agreed that there are conditions in which genetic results should be reported to research participants. Guidelines and criteria for analytic validity, clinical validity, clinical utility and ethical, legal, social issues were discussed.

Analytic Validity

Dr. Eckfeldt began this discussion by acknowledging multiple facets of diagnostic accuracy, both pre- and post-analytical. He noted that most proficiency test programs identified most errors made in laboratories as clerical. He suggested that genetic results that are reported back to subjects should be performed in a CLIA certified laboratory and if available, the laboratory should also be certified to perform the particular test. Dr. Eckfeldt explained that certification is not specifically available for many genetic tests because the genetic mutations are relatively rare and usually about 50 or more laboratories must be performing a given genetic test before the typical proficiency testing agencies will offer a proficiency testing program for the given genetic test. Working group members agreed that research investigators should give a study participant guidance on where and how they might seek clinical care when reporting results. However, it was noted that some participants may not have insurance and/or be able to afford clinical care and that a specialist may not be in reasonable proximity.

Clinical Validity

Dr. Motulsky introduced the discussion by defining clinical validity as the accuracy by which the test predicts clinical outcome. He noted that sensitivity (the probability that a person positive for a test will get disease and specificity (the probability that the test will be negative in people without disease) are evaluated when assessing clinical validity. Dr. Motulsky noted that the genetic and epidemiologic factors affecting clinical validity include analytical validity of genotyping, presence of genetic and other modifiers, heterogeneity in etiology, statistical power of studies used to identify genetic associations, selection bias, and gene-environment interaction. Clinical factors affecting clinical validity are penetrance of genotype, variable expressivity, phenotype description, environmental factors, and various clinical or other endpoints. The group discussed genetic markers that are currently clinically valid and criteria by which to judge new markers. They suggested that tests currently offered in clinical CLIA certified laboratories are a good start to identifying clinically valid tests.

Clinical Utility

Dr. Klag began the discussion by stating that the issue of the magnitude or risk and outcome is essential. He also suggested that the determination of clinical utility includes the availability and effectiveness of an intervention. The group noted that the physician’s understanding of results can affect the utility of the results and suggested that the principal investigator include a one page summary sheet, written for lay persons, to educate physicians about the results. The group agreed that a relative risk of greater than 2.0 with consideration of a significant absolute risk should be met before results should be reported to the research participant since in many studies, early results of increased relative risk cannot be replicated.

Ethical, Legal, and Social Issues

Dr. Glass started the discussion by stating that consideration should be given to how genetic results will be reported to participants and how genetic counseling will be implemented. The group agreed that it was the obligation of the investigational team to have a consultant who understands the genetic information and possible benefits and harms explain possible findings as well as the actual genetic results and their meaning to the participant if such expertise is not available within the investigative team. The group also agreed that a research study is obligated to provide reportable results to participants as long as the study is active, even if primary data collection is finished. The group was less clear on the appropriate intensity of efforts to contact participants (that is, is one letter or phone call enough?), or on the responsibilities to participants once contact has ceased, particularly if some contact information is still available. NHLBI was advised to seek additional legal advice on this issue.

Recommendations

1. A primary finding of this group was that some genetic test results from research studies may be offered to subjects, using similar guidelines as other tests, such as cholesterol levels, which are routinely shared with subjects. There are conditions in which genetic results should be offered to study participants. Examples include homozygous Factor V Leiden, cystic fibrosis transmembrane conductance regulator (CFTR) and breast cancer BRCA1/BRCA2 mutations.

2. In general, genetic markers should not be withheld if they meet key criteria described below.

3. Genetic tests should meet three key criteria before they can be reported to participants and their physicians:
   a. The risk for the disease should be significant, i.e. relative risk >2.0. Variants with greater penetrance or associated with younger age of onset should receive priority.
   AND
   b. The disease should have important health implications, i.e. fatal or substantial morbidity or should have significant reproductive implications.
   AND
   c. Proven therapeutic or preventive interventions should be available.
4. Genetic test results should not be reported to study participants and their physicians as clinically valid tests unless the test(s) was performed in a CLIA certified laboratory. If the test was performed in a non-CLIA certified laboratory, a CLIA certified laboratory should be sought to confirm results by redrawing a sample and performing the test within the CLIA certified laboratory. If a genetic test is performed only in one research laboratory and thus is unable to be performed in a CLIA certified laboratory, the test needs to be run by two different methods and/or the research laboratory should work under direct supervision of a CLIA certified laboratory to confirm results. Results reported by a research laboratory should be identified as ‘research’ results.
5. Final decisions regarding reporting of research results should not be made by the investigator alone, and should be done only with IRB approval after careful consideration of risks and benefits.
6. A process should be developed for educating non-geneticist members of the research team (investigators, IRB members, subject advocates, etc.) on the difference between highly penetrant monogenic genetic diseases as compared to genes of small effect contributing to complex traits. Such understanding is required for the evaluation of the risks and benefits of reporting results to participants.
7. NHLBI should develop a list of widely available genetic tests and a subgroup of the working group will suggest those appropriate for consideration for reporting. This list cannot be considered inclusive, given the changing nature of the field, but should provide examples and guidance for deciding which tests should be offered. These suggestions should be reviewed by investigators from individual studies for appropriateness for reporting in their study. This process should be repeated on a periodic basis by a group with sufficient expertise to judge the evolving scientific foundation for reporting these results.
8. Appendix A lists examples of genetic tests/markers which should or should not be reported. Appendix B lists genetic diseases in which clinical testing is available in more than two U.S. laboratories (supplied by GeneTests).
9. Consent forms should address results with personal implications and reproductive implications separately, as by a two part question such as, “We will be studying genes that affect cardiovascular disease but may find other genetic disorders. Do you want results reported that have significant health implications for yourself or your family members? Reproductive implications for yourself or your family?” Persons administering informed consent for genetic tests should be trained to explain the potential implications of reporting, both personal and reproductive.
10. A counselor/consultant should be provided to explain the nature of the study, implications of participation, and the potential relevance of the genetic results, including any risks of harm or potential for benefits for participants, their families or communities; this person need not be a licensed genetic counselor, but must be properly qualified by training and experience to execute this responsibility appropriately.
11. Legitimate and brief information, preferably on a single page, should accompany test results to inform clinicians about what to do with the genetic test/marker results. Ideally, these information sheets should be standardized and available from a repository, perhaps as part of a website relating this information. Findings with reproductive implications, including implications for the relatives or offspring of the subject, should follow the same guidelines as other tests.
12. Standard of care/clinical practice guidelines should be followed.
13. Recommendations regarding reporting of genetic results arising from this NHLBI working group should be coordinated and harmonized across all DHHS agencies (NIH, FDA, CDC, HRSA, etc.) and other federal agencies funding such research if possible.
14. Consensus panels by professional organizations (American Society of Human Genetics, American College of Medical Genetics, etc.) may be valuable in establishing or reviewing the criteria, so that recommendations developed by NIH are not viewed as designed to serve its research agenda.
15. DHHS should issue formal, uniform guidance for IRBs, institutions, investigators and sponsors with respect to best practices for testing and reporting genetic results in human research studies.

Meeting Roster

Reporting Genetic Results in Research Studies Working Group Meeting
July 12, 2004

MEMBERS

Russell V. Luepker, M.D., M.S., Chair
Mayo Professor 
Division of Epidemiology
School of Public Health
University of Minnesota
1300 S. Second St.
Minneapolis, MN, 55454, USA
(612) 624-6362 office
(612) 624-0315 fax
luepk001@umn.edu

John Eckfeldt, M.D.,Ph.D.
Professor
Departments of Laboratory Medicine and Pathology
University of Minnesota
763-1 Mayo
420 Delaware Street SE
Minneapolis, MN 55455
(612) 626-3176 office
(612) 626-3176 fax
eckfe001@umn.edu

Kathleen Cranley Glass, D.C.L. 
Director, Biomedical Ethics Unit
Associate Professor, Departments of
Human Genetics & Pediatrics
McGill University
3647 Peel Street
Montreal, Quebec H3A 1X1
(514) 398-6945 office
(514) 398-8349 fax
kathleen.glass@mcgill.ca

Gail P. Jarvik, M.D., Ph.D.
Associate Professor of Medicine
Division of Medical Genetics 
University of Washington Med Center Box 357720 
Seattle, WA 98195-7720
(206) 685-9069 office 
(206) 616-7186 fax
pair@u.washington.edu

Michael Klag, M.D.
Professor
Department of Internal Medicine
School of Medicine
John’s Hopkins University
Building 2024, Suite 2-200
600 North Wolfe St.
Baltimore, MD 21287-1824
(410) 955-0496 office
(410) 955-0315 fax
mklag@jhmi.edu

Greg Koski, M.D., Ph.D. 
Associate Professor
Department of Anesthesia
Massachusetts General Hospital
32 Fruit Street Clinic 3
Boston, MA 02114
(617) 726-8980 office
(617) 726-5985 fax
gkoski@partners.org

Arno Motulsky, M.D. 
Professor
Depts of Medicine & Genome Sciences
University of Washington
Health Sciences Bldg K-343A
Seattle, WA 98195
(206) 543-3593 ext. 357730 office
(206) 685-7301 fax
agmot@u.washington.edu

Benjamin Wilfond, M.D.
Medical Genetics Branch, NHGRI
National Institutes of Health
Bldg 10, Rm. 1C-118
Bethesda, MD. 20892 
(301) 435-8728 office
(301) 496-0760 fax
wilfond@nhgri.nih.gov