Second Report of the NIGMS Populations Advisory Group on Pharmacogenetics Research

Bethesda, Maryland

Start Date: 8/2/2000 8:00 AM

End Date: 8/2/2000 5:00 PM


The Populations Advisory Group heard overviews of the research studies involving human subjects that are being conducted by the Pharmacogenetics Research Network. They learned about the nature of information being stored in the Pharmacogenetics Knowledge Base (PharmGKB), and the efforts being taken to ensure people's privacy. Plans were presented to probe the public's perceptions of this research by using a telesurvey as a part of one of the funded research projects. The Populations Advisory Group provided specific advice on each of these activities, and additionally, recommended that researchers should: be respectful, but not afraid to proceed; recognize that community members are experts on communities; allow people to describe themselves; and provide post-study feedback in simple, direct language.


The purpose of the Populations Advisory Group is to consider the implications of pharmacogenetics research for members of identified populations. The group met previously in May 1999, and a report including the group's recommendations can be viewed at NIGMS is working to convey the advice of this panel to researchers and is considering this input while the operating policies for the Pharmacogenetics Research Network and Knowledge Base (see are being developed through the Steering Committee.

The group was tasked with providing feedback regarding the specific research projects that are currently funded through the Pharmacogenetics Research Network. The request for applications describing the network was recently reannounced, and NIH is preparing for the second round of competition. NIGMS considers the comments and advice from lay and professional people who are not necessarily pharmacogenetics experts to be extremely helpful. NIGMS wants to promote and fund research that is consistent with these recommendations.


Pharmacogenetics Research Network Study Designs
Rochelle Long, Ph.D.

The first presentation outlined the range of research protocol designs for the pharmacogenetics awards (abstracts for the awards are available at; additionally, a copy of the handout for this presentation is available upon request). In general, the projects employ either of two strategies: a genotype to phenotype approach or a phenotype-to-genotype approach. In the first approach, investigators are concentrating on first discovering the varying genetic sequences, followed by determining the functional consequences of such genetic variability. In the latter approach, patients are identified first, in particular those who have unusual or distinctive presentations and/or therapeutic drug responses compared to others. Genetic analyses of patient samples can then yield information that is correlated with the clinical responses. At any time (before or after genotype/phenotype relationships are established), "control"--meaning those not selected for any particular disease or medication response--human samples can be studied to learn the frequency of the occurrence of genetic characteristics in the general population or in isolated groups.

For these types of research studies, study participants are generally told that they will not learn the results about their genetic characteristics, for several reasons: 1) scientists do not yet know exactly how to interpret this information; 2) genetic counseling standards have not yet been developed, because in most cases the results cannot yet be interpreted; 3) these are not standardized clinical laboratory tests being offered in certified labs; and 4) information is sometimes encrypted from both the investigator and the patients. This is an appropriate way to conduct basic research studies, where the objective is to learn how to correlate clinical presentations with the observed genetic variability.

Some of the investigators will be examining groups of samples derived from human materials (e.g., DNA that has been isolated using immortalized cells created from human blood cells). These samples may have been obtained from a repository where they were anonymized and appropriately consented for sharing and distribution for research purposes. In some cases the self-identified racial and/or ethnic backgrounds of the samples are known. Using current technologies, genetic sequence variation can be identified in these samples for the molecular pathways under study.

Sometimes patients undergoing treatment for various diseases, such as neuropsychiatric disorders or different cancers, agree to participate in studies where their biological samples (often blood cells or cells lining the cheek) are collected and analyzed to search for genetic variation. The patients may already be participating in research studies, including clinical treatment trials or other genetic studies (not necessarily genetic variation studies). It is important to note that the goals of the pharmacogenetic studies must be explained to the participants, who have the opportunity to accept or decline to join this study. The subjects are informed that this will not affect the course of medical care they are currently receiving. They are also told that their interest in participating must be based on altruism. Patients receive a token reimbursement, usually around $35-$50 per event (a lab test or an interview).

Of the research proposals funded to date, most studies incorporate participants from a broad range of ethnic and racial groups, and most often they are selected because of the disease condition under investigation. The members of study populations are typically referrals from ongoing clinical treatment trials, and often are patients seeking medical care from centers located in and around urban areas. The group compositions of the study populations have been described (taken from the applications) variously as: Caucasians or Whites or European-Americans; Blacks or African-Americans; Asians or Asian-Americans; Hispanics or Hispanic-Americans or Latin Americans; and Native Americans. The balance between men and women in these studies is reflective of the predominance of the diseases being studiedreflective of the predominance of the diseases being studied (e.g., twice as many women as men seek treatment for depression, and less than 10 percent of a breast cancer population is comprised of men).

For one proposal, a limited Mexican-American population in the greater Los Angeles area is being studied, because the group was identified as an underserved population. In this case, the investigator's motivation is both to minimize the fluctuation in the genetic background by more narrowly circumscribing the group, as well as to offer the opportunity for people to participate in clinical research studies and thereby receive medical care. In this particular study, after completion of an 8-week component, the participants are informed of their drug responder status (how well a drug appeared to work according to the standard psychological instruments), although they will not be informed of their genetic status for the reasons listed previously. In this study, great care will be taken to employ bilingual and acculturated staff, and to offer supportive counseling and referral. Although this particular population is being examined, the genetic variation-disease correlations that are found will likely hold for individuals of any ethnic and racial background.

PharmGKB: The Pharmacogenetics and Pharmacogenomics Knowledge Base
Mildred Cho, Ph.D.

This presentation described the research project directed by Russ Altman (Stanford University) that is funded to create the Pharmacogenetics and Pharmacogenomics Knowledge Base, or PharmGKB. This is an integrated data resource that will focus on how genetic variability relates to the observed range of responses to medications, and will interlink genomic, molecular, cellular, and clinical information about gene systems important for modulating drug responses. PharmGKB is based on a hierarchical data representation system that allows the data model to evolve as new knowledge is gained, while ensuring the security and stability of the information. It is a work-in-progress; the data model is being developed at the present time.

Right from the start, PharmGKB is being designed to respect the absolute confidentiality of genetic information from individuals. The standards for maintaining privacy and confidentiality will be tied to the Health Insurance Portability and Accountability Act of 1996, and there will be policy statements posted accordingly (see Certain identifying data (e.g., names, addresses, phone numbers) will never be accepted into PharmGKB. Because this will be a repository that has free and open access, raw medical records data will not be deposited, in order to protect the identity of the people participating in the research. Most of the clinical data content will be aggregated into "rankings along a range" (e.g., percentiling) to be expressed in PharmGKB.

In addition to limits on the content and format of data that are submitted by the investigators and entered into PharmGKB, there will also be limits on the retrieval of data by PharmGKB users, both for the data values alone and in combination. Computational transformations will take place called "binning," an adjustable process of combining data so that "bins of not smaller than x" are created when data are displayed, to ensure that sufficient information to identify an individual cannot be released to the general public. For example, a current standard is a "cell size" of not less than five, meaning that the information released should not be sufficient to narrow down to a group of five people or fewer. This cell size may need to be adjusted, or even be variable as part of a dynamic process, depending upon the kinds and values of data being retrieved.

Frequently Asked Questions (FAQ) will be posted at the PharmGKB Web site, in order to anticipate people's concerns and address them in plain language (see the PharmGKB FAQ attachment, which was revised following the Populations Advisory Group meeting). The Pharmacogenetics Research Network is also working on developing model informed-consent language that will mention a public data repository, among other considerations (e.g., lack of significant financial benefits for study participants). This way, people will know in advance about PharmGKB (and its access on the Internet), and can consider this when deciding whether to participate in the research. Plans for the protection of human subjects and for computer security measures have been presented to and approved by the Stanford University Institutional Review Board, even though this was not a Federal requirement. The nature of the computer access and types of queries entered by users of PharmGKB will be continuously monitored. This is consistent with the overall objective of being "state-of-the-art" in ethical protections and cognizant of all the possible ramifications of a computer data repository.

The Populations Advisory Group discussed the FAQ and offered the following advice:

Suggestions for the PharmGKB FAQ

  • Who has ownership of data (including sale of data and any benefits to patients)?
  • What about any future clinical discoveries?
  • Could the data be used for other purposes?
  • How reliable is the data?
  • What if something goes wrong with the protections?
  • Who will access the data?
  • How long will the data be stored?
  • What are the anticipated benefits for communities?
  • Who else can obtain the data under Federal "sunshine" laws?
  • How will the users be told they are being monitored?
  • FAQ answers should be checked and re-checked, to see if answers can be "twisted"!

Survey of Ethical Implications of Pharmacogenetics Research
Phyllis Griffin Epps, Ph.D.

This presentation outlined the objectives of a project headed by Mark Rothstein at the University of Houston that is designed to anticipate the impact of pharmacogenetic differences related to ethnicity. The research will focus on four elements: 1) current and future use of information intended to establish the safety and efficacy of certain drugs; 2) the impact on legal standards applicable to the recruitment and use of human subjects in drug trials, including issues of diversity, informed consent, and confidentiality; 3) providing access to appropriate health care for certain groups and the anticipated response of managed care organizations; and 4) the impact on legal and ethical standards of care, applicable to physicians and pharmacists advising patients about genetic intolerance to, or compatibility with, certain drug therapies.

As part of this project, a telephone survey will be conducted to determine how members of various racial and ethnic groups view pharmacogenetics/pharmacogenomics, and what they might want to know before agreeing to a diagnostic test or to take medicines that are prescribed based upon genetic variation. The plan then is to host a conference that brings together key players to examine and discuss the ethical, legal, and social issues. The conference speakers will write a book that considers the implications, especially as they relate to race, ethnicity, and gender, and will set forth their collective opinions of guidelines for the proper use of the science.

The telephone survey component was discussed extensively (an updated copy of the survey questionnaire is available upon request). The objective is to probe on the following topics: people's attitudes and whether they might choose to participate in research studies; what concerns they have; whether other beliefs might have an impact; their perception of the genetic literacy of physicians; and who might be trusted to conduct research among pharmaceutical companies, academia, or government. The plan is to make 1800 randomly placed calls in selected areas and to direct survey questions in the language of choice to those who describe themselves as Whites, African-Americans, Hispanics, and Asian-Americans. People will be categorized by self-identification, which the investigators believe will predict habits and lifestyle choices, as well as attitudes toward participation in research studies. The aim of the proposal is to be comprehensive, although not necessarily totally inclusive, in reaching different racial and ethnic groups. There was some discussion of whether self-identification will be valid. Many members of the group thought it would be. One member pointed out that answers describing a multi-ethnic background must be expected and interpreted correctly.

The Populations Advisory Group discussed the telephone survey and offered the following advice:

Suggestions for the telephone survey

  • Consider the biases in conducting a telephone survey.
  • A significant number of people living in rural areas have no working telephones.
  • Expect that some people will be untruthful in their responses.
  • There will a high non-responder rate.
  • The survey instrument must be translated and back-translated.
  • There are limitations to self-identification.
  • Why was Mandarin Chinese selected over Cantonese?
  • Allow for multiple ethnicities within a family tree.
  • Do not discount people not included in one of the four major identified groups.
  • Know the "cultural context" (questions can be misinterpreted easily).
  • Misunderstood questions will lead to misunderstood results, or worse.

Progress Report--Other NIGMS Activities

Next, NIGMS staff summarized some related activities for the Populations Advisory Group. Alison Davis described the creation of the brochure "Medicines for YOU", and thanked the group members for their input. This educational brochure is available in single or bulk copies upon request.Judith Greenberg described the activities of the NIGMS Human Genetic Cell Repository at the Coriell Cell Repositories. At a recent meeting, a working group assembled by NIGMS recommended the creation of special oversight groups to provide guidance on new and existing sample sets derived from human tissues (see a report at This led to NIH convening a "Community Consultation" meeting, which occurred in September 2000. Members of the Populations Advisory Group indicated their interest in learning the results of the meeting. An NIH Record news article describing the meeting c​an be found at, and a report will be posted at in January 2001.

There were other miscellaneous points raised during the day's discussion that are worth noting for the record. The Populations Advisory Group considered the legal ramifications (e.g., liability for diagnostic tests not being administered, or misinterpreted results of tests) of scientific advances made in the pharmacogenetics arena. They discussed the current state of the law, as distinct from the state of the science. The latter is forward looking, while the former is based upon precedent. Finally, in one exchange, a group member clarified the following: the Tuskegee experiment is not the same as the Tuskegee study (the experiment is a positive thing, and refers to building a university; the study is a negative thing, and refers to monitoring the development of syphilis without proper treatment).

Open Discussion

The first part of the general discussion period focused on community responses to pharmacogenetics research. One member said that he showed the educational brochure to community-based individuals, and got one strong message: biomedical researchers do not provide enough feedback to the community. He reported that the reactions ranged from concern to anger. Participants in research projects need to hear about the results in easy-to-understand language. Both risks and benefits to individuals and groups must be considered. A scientist pointed out that universities do not allow researchers to contact study participants directly--in fact, some institutions have committees with lay membership that serve as buffers to decide whether subjects can be re-contacted. These barriers make it difficult, even when interventions are possible, such as providing advice to change a medication, or to change the dosage. The group discussed to whom the information would go--to the person? to the family? to the physician? The Populations Advisory Group believed that it should be stated clearly in the informed consent process that participation in these studies must be based solely upon altruism, and that no clinical information will be provided, nor will medical recommendations be made at the time the research is being conducted. They felt that people truly do want to contribute to science. It was strongly suggested that simple steps such as an immediate thank-you note (sample text is below), as well as access to a lay summary at a later date with a digest of the experiment's conclusions, would help significantly.

One group member stated that it is absolutely unacceptable to collapse groups under the heading "Other." She stated that it is insulting to people to be recruited based upon their racial and/or ethnic background and then not be identified as such. All groups should be broken out in reporting research results. Communities need to feel appreciated and understood by scientists. Also, it must be recognized that members of an impoverished community do not have a constant relationship with a single physician. Some do not even have a consistent medical facility. Some go to emergency clinics or health fairs or schools to obtain medical care.

The discussion then turned to legal responsibilities. If an adverse effect is identified 5 to 10 years after a research study is conducted, does the researcher have a legal obligation to notify the study participants? What is the due diligence? Phyllis Griffin Epps replied that litigation is reactive, not proactive. The basis for decision-making would be whether it was reasonable to inform someone on the basis of what was known at the time the research was conducted. The statute of limitations for legal action is 2 to 5 to 10 years, depending upon the state. The first few cases will set the precedent for whatever else comes along. The group thought that because technology is always changing, it might be wise to recommend that a research subject keep in contact with a center if he or she wishes to know anything in subsequent years.

The group then chose to discuss the differences between race and ethnicity. Language is confusing, and people want to be respectful, as well as politically correct. One definition was volunteered stating that a difference between race and ethnicity is that ethnicity can be more fluid, can be a social category, can be political or geographical, and can change. The group asked: is Black a race? is African-American an ethnicity? Another view was expressed that simply to use the word race is racist. The group put forth that societally, we are a visually based culture in this country. There is a concern that people will receive different medical treatments based upon race. One group member stated that race is the most analyzed and the least understood attribute, and there are no species differences. The group observed that before 1990, the term race was used more often, but now, in 2000, the word race and ethnicity are both used, and a prediction was made that after 2010, the word ethnicity will be used exclusively.

However, the focus for pharmacogenetics research is to understand drug responses, and sometimes to understand disease itself. The group recognized that people cannot always be lumped (e.g., collapsing all Koreans, Japanese, Chinese as Asians--what about different Chinese groups or grouping all tribes as Native Americans without thinking of their distinctions?). There are Federal requirements to report the distribution of patients enrolled in clinical studies (see the definitions of terms in the OMB Directive 15, at These categories have been revised periodically (see, and were expanded for the 2000 Census (see This is an extraordinarily difficult issue.

The goals for pharmacogenetics research have nothing to do with determining the basis for race or ethnicity, but grouping people according to their origins can lessen the background genetic variability against which a particular drug response characteristic appears. From a public health perspective, it could be a positive outcome if significant differences in the distribution of genetic variants are found, because different policies could then be recommended according to existing political-geographic lines (e.g., deciding whether to screen the population in an Asian country for a life-threatening drug response polymorphism).

The group agreed that there are social and historical concepts of race (in contrast to biological descriptors). What really matters to researchers is descent, or geographic origin. Under no circumstances should racial/species arguments be permitted. Descent, geography, and family background are relevant. This could mean collecting more detailed information than the regulations require, by obtaining an oral history and allowing self-identification. This may need to be tailored according to group. For example, a question about blood quantum may be suitable for Native Americans, where tribal registries exist and are valued, but this could be viewed differently by other groups. Researchers should be respectful but not afraid. Cultural differences will lead to distinct and different medical approaches being sought and/or accepted. Operational definitions may be different depending upon scientific, social, or public health goals. For example, one member pointed out that very specific questioning about family background takes place when newborn screening tests are offered. This occurs for practical reasons, and in a sensitive environment. Because of the financial burden, all tests cannot be offered to all people, but reasoned decisions must be made without raising people's anxiety levels.

In summary, the group recommended that questions about ethnicity, geography, and descent may be asked. They felt strongly that people must be allowed to describe themselves. Multiple answers are acceptable, allowing for multi-ethnic identities. Synonymous questions might work (e.g,. asking about the dominant part of a family's origin rather than geography alone). The historical and linguistic aspects of group identity should be honored. Cultural differences between groups must be respected.

The Populations Advisory Group recommended that the work of the Pharmacogenetics Research Network and Knowledge Base proceed, and the group asked to be updated regarding other related activities as they occur. The group agreed that its next meeting will tentatively be planned for Spring 2001, ideally to coincide with the open scientific meeting that the network is planning for April 25, 2001 on the NIH campus.

Overall Recommendations

  1. Researchers should be respectful, but not afraid to proceed.
  2. Recognize that community members are experts on communities.
  3. People must be allowed to describe themselves.
  4. Provide post-study feedback to the community in simple, direct language.


  • Welcome and Introductions, Charge to the Group
    Marvin Cassman and Dan Kevles (Chair)
  • Pharmacogenetics Research Network
    Rochelle Long (with comments from Dick Weinshilboum)
    • Overview of the applications
    • What kinds of research are being conducted?
    • Which populations will be studied and how?
  • Questions & Answers
    All Participants
  • Pharmacogenetics Knowledge Base
    Mildred Cho
    • What information will be contained?
    • How will people's privacy be protected?
  • Questions & Answers
    All Participants
  • Ethical and Social Issues in Pharmacogenetics Research
    Phyllis Griffin Epps
    • Plans to learn of people's reactions
    • Plans to stimulate exchange between experts
  • Questions & Answers
    All Participants
  • Progress Report--Response to advice provided at the last meeting
    NIGMS Staff
    • Focus group testing, development of the brochure
      Alison Davis
    • Feedback from the Steering Committee
      Rochelle Long
    • Community Consultation meeting planned
      Judith Greenberg
  • Group Discussion
    Dan Kevles, Moderator
  • Wrap-up, Concluding Remarks
    Dan Kevles and Marvin Cassman

Sample "Thank You Note" Text

Thank you for joining this research study. It is part of a larger scientific effort called "pharmacogenetics" that is designed to understand the connection between genes (your DNA) and medication responses. It may take many years to understand the results of these studies. The scientific and medical advances will take place because of the participation of willing volunteers like you.

If you would like to know more about the results of this research, watch your local television or newspaper or contact the university's office for the public at (phone number and/or address) periodically to ask if any findings are ready to be shared with the public, or if any special programs are planned for the community.

Part of the money used to perform studies comes from the National Institutes of Health (NIH). You can learn more about NIH at You may want also to read the brochure "Medicines for YOU."

Your comments are welcome. Please give the number of the study you have joined, which is (number), and feel free to tell us what you think at:

Give university office's name or IRB and address

For Further Reading

(These are selected current articles identified at the meeting and afterward.)

Jones, C.P. "Levels of Racism: a Theoretic Framework and a Gardener's Tale." American Journal of Public Health , 90 :1212-1215 (2000).

Sharp, R. R. and Foster, M. W. "Involving Study Populations in the Review of Genetic Research". Journal of Law, Medicine, and Ethics , 28 :41-51 (2000).

Weijer, C. and Emanuel, E. J. "Protecting Communities in Biomedical Research". Science 289 :1142-1144 (2000).

"Census, Race, and Science (editorial)". Nature Genetics , 24 :97-98 (2000).

"Perceptions of Genetic Research: A Focus Group Study. A Report for the National Health Council."

This brochure can be requested from the National Health Council:
1730 M Street, NW, Suite 500
Washington DC, 20036
Tel: 202-785-3910
Fax: 202-785-5923

Comments and feedback on this report are welcome and may be directed to:

Rochelle M. Long, Ph.D.


The author thanks members of NIGMS staff and the Populations Advisory Group members for their helpful suggestions.