Presentation - Lisa Eckstein

Presentation - Lisa Eckstein
Community Participation

Tarrytown Talk—Racial Justice and Community Engagement

Lisa Eckstein, currently an SJD candidate at GULC, writing my dissertation on the regulation of medical research involving minority racial and ethnic groups.

Quantifying the use of race and ethnicity in biomedical research and its defining characteristics

In August 2011, I searched the National Clinical Trials Registry for race-based biomedical research projects. I defined “race-based research” as studies that placed racial or ethnic restrictions on study enrollment. Of the 259 race-based studies I found, 125 were not biomedical, and therefore were not investigated further. For the remaining 134, I extracted information about: the study sponsor; study type (clinical drug, other interventional, observational, and genetic); race/ethnicity being investigated; definitions of race/ethnicity; any public funding disclosed for the research; and other information being collected which may explain any racial or ethnic differences, such as genetic ancestry markers and socioeconomic status.

Brief snapshot

  • Close to two-thirds interventional in nature—64 clinical drug trials; 19 other interventional trials; and 51 observational (including genetic) studies.
  • Approximately one-third were conducted by the NIH or listed the NIH as a collaborator. Universities and medical centers conducted the bulk of race-based biomedical studies in the U.S., mostly without any disclosed public collaboration. Pharmaceutical companies conducted and sponsored ten domestic biomedical studies and a further ten international biomedical studies.
  • African-Americans were by far the most common race/ethnicity on which biomedical studies were conducted, and African-Americans and Caucasians by far the most common racial or ethnic comparison.  A smaller number of studies were conducted on “Asians” and“Hispanics.” No biomedical research studies specifically involved American Indian/Alaska Native persons. Caucasians were used as a comparator in virtually all racial comparisons.
  • Very few studies indicated the collection of information that could explain the reasons for racial or ethnic differences. Four clinical studies collected potentially relevant biological factors, including two which assessed genetic ancestry markers. Seven observational studies collected lifestyle information including, for example, participant and family medical history, nutrition, tobacco and alcohol consumption, work history and education, income, insurance access, physician utilization and patient trust-in-physicians. Only two genetic studies specified the collection of genetic ancestry markers.
  • Wide range of definitions of race and ethnicity,
    • Approximately half failed to provide any objective guidance regarding what race or ethnicity means for study purposes and even one which expressly gave investigators decision-making power as to whether a potential subject met the inclusion criterion of “Japanese.”
    • Approximately one-third provided extensive guidance—interestingly, many of these seemingly connected to international drug regulatory requirements. Eg, “Of Japanese descent who has resided outside of Japan for no more than 5 years and whose parents and maternal and paternal grandparents are Japanese, as determined by participant's verbal report. Japanese participants must have a valid Japanese passport.” Not coincidental that the Japanese Pharmaceutical and Medical Devices Agency has been one of the most stringent regulators as regards requiring drugs to be approved for its national population.
    • No guidance at all re Caucasians, even when a study is comparing Caucasians with another racial or ethnic group and defines the minority group. Eg, For example, a clinical drug trial that enrolls Chinese, Japanese and Caucasian volunteers. The study requires that “Chinese and Japanese subjects must have both maternal and paternal grandparents of Chinese or Japanese origin, respectively.” Chinese and Japanese subjects must not have lived outside of China or Japan, respectively, more than five years or, since leaving their country of origin, have changed their diet significantly. No guidance is provided on defining “Caucasian” volunteers.

Community engagement strategies as a potential regulatory solution

Strategies for incorporating community engagement in the biomedical research pathway: in particular, NIH funding decisions, IRB review, and FDA marketing approvals. Exposing investigators to racial justice issues, and improving the legitimacy of research regulation. Principally based on the concept of race impact assessments. Discuss in brief the main regulatory features I put forward.See this as a means of opening up space in the regulatory pathway for investigators to engage with many of the racial justice issues that currently are being sidelined.

(1) threshold for when regulatory action is warranted, (2) responsibility for undertaking an RIA, and (3) consequences of RIA findings.

  • (1) Threshold for when regulatory action is warranted: Need to distinguish between, for example, a project that impacts on minority groups because it enrolls disproportionate numbers due to study location, and far more direct effects arise from studies such as those that supported BiDil’s race-specific registration. In other words, in what circumstances regulatory action is warranted.Borrowed from requirements already in place under the Executive Order that requires gov’t agencies to undertake a “tribal impact assessment” for actions that affect American Indian/Alaska Native tribes—namely, projects that have “substantial direct effects.”Typically that would include studies that condition enrolment on race or ethnicity.


  • (2) Responsibility for undertaking an RIA: Obligations to undertake impact assessments commonly placed on gov’t agencies, including environmental impact assessments, tribal impact assessments, and most privacy impact assessment requirements. Could have placed obligation on relevant agencies, eg NIH, FDA. HOWEVER, preferable option to encourage individual investigators to develop RIAs and empowering regulator decision-makers such as the NIH and FDA to direct an investigator to conduct an RIA in limited circumstances.
    • Advantages:
      • multiple steps of biomedical research pathway, minimize duplication
      • Investigators best placed to understand the project, potential to amend protocol in accordance with community preferences
      • Proactive investigator role—foster culture of racial awareness and sensitivity, expand investigator’s capacity to design suitable research projects in the future.
    • Cautions: Ongoing need for agency guidance—won’t minimize duplication if investigators are left on their own to determine how to conduct community engagement and the implications of engagement outcomes.


  • (3) Consequences of RIA findings: An important consideration is what should be the consequences of minority racial and ethnic group member preferences elicited. Consistently with impact assessments in other contexts, I suggest giving deference to investigators in the extent to which, if at all they incorporate such views. In this way, RIAs offer a tool to ensure good-faith consideration of the implications of such research for minority racial and ethnic group members. Provided investigators have made bona fide efforts to seek diverse feedback and can justify their decision whether to incorporate views expressed, then regulatory agencies should accept the RIA process.One point of variation to point out is in the context of IRB review—some issues raised in an RIA process will feed directly into principles for ethical acceptability.
    • Eg, hypertension research involving Hmong Americans for which the survey instrument was drafted in a way that participants may perceive as threatening. [An original survey question asked: "Do you think the following statement is true, false or don't know? If your parents have high blood pressure, you have a higher chance of getting it?" After some participants perceived the question as “potentially threatening” and “a curse,” the question was reworded to lessen study risks from participants’ perspectives.] Should an investigator decide to retain the original wording, an IRB would be entitled to factor into the study risks potential emotional harm.
  • Seek to open up regulatory spaces for community engagement, promote partnership, expose researchers to issues associated with race-based research that they otherwise may be oblivious to. However, the workability of my model depends upon sufficiently educated and engaged minority group members—that is moving beyond demands of access and inclusion. Focus of education likely to be different depending on the racial or ethnic group at issue. I would love to say I’ve found the solution—unsurprisingly I haven’t.