Muddied Waters: FDA Encourages Use of Race in Clinical Drug Trials

By Phyllis Griffin Epps

The Food and Drug Administration (FDA) is finalizing guidelines that would recommend the use of racial and ethnic categories in clinical drug trials and the submission of data regarding the safety and efficacy of pharmaceutical products subject to FDA approval.  The draft guidance, titled “Guidance for Industry: Collection of Race and Ethnicity Data in Clinical Trials” ( ) and issued in January 2003, has as its premise the 1998 final rule of Investigational New Drug Applications and New Drug Applications (the “Demographic Rule”).  63 Federal Register 6854 (Feb. 11, 1998) (codified at 21 CFR 312.33(a) (2) and 21 CFR 314.50(d) (5)).  A commentary on the Demographic Rule, which requires sponsors of New Drug Applications to include an analysis of data according to demographic subgroups, the draft guidance recommends use of the categories developed by the Office of Management and Budget (“OMB”) to standardize the collection and use of race and ethnicity data by federal agencies.  The draft guidance indirectly mentions pharmacogenomics as one reason for the use of OMB categories on race and ethnicity.  For reasons stated below, the FDA should reconsider its initial recommendations.

OMB Policy Directive 15 encourages the use of a two-part format that treats separately race and ethnicity, and the use of self-reporting to collect such data.  In the world according to the OMB, Americans are divisible into one or more of five races (American Indian or Alaskan Native, Asian, Black or African-American, Native Hawaiian or other Pacific Islander, and White) and one of two ethnicities (Hispanic/Latino or not).  The draft guidance encourages the use of these categories, though with minor revisions to the definition of “Black or African-American.”  OMB has stated that the categories are social-political constructs and not scientific or anthropological in nature.  Across categories, definitions range from skin color (Black) to geography (Asian) to “community attachment” (American Indian).

An increasing number of clinical trials for drugs that ultimately seek FDA approval take place outside of the United States.  During the comment period, several contributors cited difficulties associated with the use of OMB categories and self-reporting techniques with participants in such countries as Japan and Brazil.  The comments urge that the subject be presented before the International Council on Harmonization (ICH).  (See, e.g., ICH, Guidance on Ethnic Factors in the Acceptability of Foreign Clinical Data at

The first reasons cited in support of the recommendation are its strongest.  To be specific: “the use of the recommended OBM categories will help ensure consistency in demographic subset analyses across --- studies used to support certain marketing applications to the FDA --- and across data collected by other government agencies.”  The use of uniform standard categories in data collection may well yield information that is more easily shared with other agencies.  The question is, for what purpose is the information being collected?

The draft guidance states that the OMB categories “may be useful in evaluating potential differences in the safety and efficacy of pharmaceutical products among population subgroups.”  This objective, an indirect reference to pharmacogenomics, or the study of the effect of genetic variations on drug response, drew considerable criticism from the biotechnology industry during the comment period.  Differences in response to drug products occasionally seem to correspond with subgroups of the United States population typically categorized as races.  The heritability of drug responses is the result of, among other things, migratory patterns of early humans.  Many variations in genotype are the result of patterns of human migration from Africa across other continents over thousands of years.  The very, very, very general patterns of genetic diversity attributable to continental migration (or lack thereof) are rife with exceptions, however, and do not support the proposition of biologically distinct races.  More genetic diversity exists on a single migratory branch of the human family tree than between branches.  Importantly, the patterns of genetic variation reflected in migratory history do not correspond well to existing categories of race.  Regarding inherited variants that affect drug response, the continent of ancestral origin yields more useful information than race.  To the extent that the current OMB categories are sociopolitical and not scientific or anthropological in nature, data collected will not facilitate the prediction of drug response to genetic variation.

The collection of data in accordance with OMB categories, particularly through the use of self-reporting, will yield significant information about cultural affiliation rather than biology.  Identity with a particular culture, whether by voluntary or involuntary assignment, can be useful in efforts to examine physical and social environment.  Information about environment is useful in efforts to identify factors that contribute to health status, including the environmental elements relevant to drug tolerance and response.  Several of the comments favoring the guidance were submitted by organizations that advocate on behalf of one or several ethnic groups.  Where the objective of measurement is integral to research regarding the role of environment or efforts to remedy or prohibit unjust behavior, however, use of the term “ethnicity” alone would elicit the same information without the troublesome ambiguities about biology that arise in discussions of race.

The use of OMB categories in clinical drug trials will contribute little of what data is most relevant to the prediction of drug response.  Where patterns of heritable genetic variants must be generalized, continental origin or early ancestry is most valuable.  The OMB categories are slightly better suited to classify elements associated with cultural affiliation, which may prove useful in the interpretation of safety and efficacy data in the post-approval, post-marketing phase of the review process.  The current version of categories does not adequately capture the necessary distinctions between heritable physical traits and shared culture.  The distinctions are necessary to weaken the assumption of biological underpinnings that are inherent in race and to elicit meaningful information in a way that minimizes misinterpretation.