Diversity in clinical trials typically refers to population groups characterized by race and ethnicity, though other groupings, such as those associated with gender, age, geography, and socioeconomic status, also are components of diversity. Yet race and ethnicity are inexact concepts, and the associations among race, ethnicity, and health depend on a wide array of factors.
At the workshop, three speakers examined the role of race and ethnicity in both health disparities and in clinical trials. Race and ethnicity have both biological and social origins and consequences, which creates a rich and complex arena for policy.
Historically, the concept of race in the United States has had two main components, said Charles Rotimi, chief of the Genomics of Metabolic, Cardiovascular, and Inflammatory Disease Branch and director of the Center for Research on Genomics and Global Health at the National Human Genome Research Institute of NIH. One component has been based on biogenetic variation determined in part by a person’s biogeographic ancestry. The other component has blended social, cultural, and genetic factors into a poorly understood construct used to sort people into a few predetermined categories (Smedley and Smedley, 2005). These categories were always problematic. For example, in the United States, having just a single black ancestor was enough for a person to be considered black. Yet in
South Africa, people with different degrees of African ancestry were sorted into distinct categories.
Many aspects of racial categories are arbitrary at best, yet race “is the fundamental basis for inequality” in the United States, said Rotimi. The concept of race has been used to drive differences that disadvantage some people and advantage others. As the National Research Council (NRC, 2001) has observed, “The idea of race and its persistence as a social category is only given meaning in a social order structured by forms of inequality—economic, political, and cultural—that are organized, to a significant degree, by race.” According to Rotimi, “That is really why it is so critical that we understand a historical perspective when we are talking about inequity and health disparities.”
The advent of genomics has provided an opportunity to look anew at the socially derived categories associated with ideas of race. The genetic variation among humans provides an unbiased history of the human species while also providing previously unknown information about human health, Rotimi said. This variation reveals that all humans are descended predominantly from a relatively small group of anatomically modern humans who lived in Africa within the past 200,000 years. The descendants of this group have dispersed across the globe over the past 100,000 years. As modern humans encountered more archaic populations of humans outside Africa, they sometimes mated with members of these groups and added fragments of their DNA to the DNA they carried out of Africa. As Rotimi said, “One of the things that humans do very well is, whenever we travel, we are very generous in sharing our DNA.”
Because of this history, most of the oldest genetic differences among people are found in all human populations. At the same time, the development of agriculture in the past 10,000 years and of urbanization in the past 700 years has led to rapid population growth and to the origin of new variants that are rare and specific to one population or even to one family. As a result, most genetic variation occurs in all populations, though some is local.
Some people have used the portion of genetic variation that is structured geographically to justify traditional racial groupings. But that interpretation is a “misunderstanding of the concept of human genetic variation,” Rotimi said. For example, Rotimi and his colleagues recently published a study of 3,500 individuals from 163 ethnolinguistic groups around the world (Shriner et al., 2014). A computer program analyzed the genetic differences among these individuals and distinguished 19 major ancestral components—more than the traditional number of racial groups, and a number that varies depending on the size of the sample. Furthermore, most of the individuals—94.4 percent—showed mixed ancestry among these components. Said Rotimi, “Trying to use genetics to define race is like slicing soup. You can cut all you want—the soup stays mixed.”
These results have implications for individuals who identify with particular racial groups. For example, a study of disease-associated genetic variants in self-identified African Americans found percentages of African ancestry ranging from 0.6 percent to 99.7 percent, with an average of about 80 percent (Shriner et al., 2011). Similarly, people who might self-identify as Mexican American or Puerto Rican can have very different combinations of ancestry. Thus, if the action of a drug being tested in such individuals is influenced by a specific genetic variant, their race or biogeographic ancestry is likely to say little about whether they have that variant. “You cannot use group data to say something about [an] individual,” Rotimi said.
Genetic diversity needs to be studied, Rotimi concluded. It can help shed light on ancient human population migrations, the biological relationships among human populations, and why some disease-causing variants occur in higher frequencies among some populations, such as variants that protect against diseases common in particular parts of the world. But genetic diversity “does not coincide or overlap in any kind of systematic way in terms of the way we try to define ourselves,” he said. “Individuals who look alike cannot be used as representing their genetics.” Similarly, if people want their genetic variants to be probed in clinical trials, they need to participate in those trials, said Rotimi, not assume that the participation of other members of their social groups will suffice (Rotimi, 2012; Rotimi and Jorde, 2010). “My genetics cannot work for you,” he concluded.
Otis Brawley, chief medical officer for the American Cancer Society and professor of hematology, oncology, medicine, and epidemiology at Emory University, took what he described as a somewhat contrarian view of race and ethnicity in clinical trials, while agreeing with the overall need to broaden participation. He said that he worries when legislators require analyses of subgroups in clinical trials. The advocates of representation mean well, but they need to be more scientific, he added. Subgroup analyses are often wrong because they include small numbers of people. For example, there are not currently enough women with stage-three ovarian cancer in the United States to get a statistically significant answer about the possible differences in drug action between white women and black women.
Race, as was pointed out by Rotimi, is a sociopolitical characterization, not a biological characterization, and these characterizations have changed over time. For example, someone born in India who came to the United States would have been characterized in three different ways by U.S. censuses since the passage of the NIH Revitalization Act legislation in 1993.
Populations can differ in their susceptibility to disease, yet simplified racial thinking can be misleading. Sickle cell anemia is found among people
throughout the Mediterranean region, not just among those with ancestors from sub-Saharan Africa. A mistaken finding that azidothymidine (AZT) was more effective in white populations than black populations took a decade to overcome; whereas, the actual difference was in adherence to the drug regimen. The problem with requirements for inclusion, said Brawley, is that they can send the message that human population groups are biologically different.
Legislation mandating the inclusion of racial groups has “moved the emphasis off the real problem,” said Brawley. Disparities in outcomes are generally due to social issues, not biological issues. “For most diseases, equal treatment yields equal outcome among equal patients,” he said. Race may help determine the quality of care, but it does not necessarily determine whether a person has a genetic variant that will influence a drug response. “There is not enough concern or emphasis on the fact that there is not equal treatment,” he said.
Studies have revealed many of the reasons behind disparities in outcomes, including cultural differences in acceptance of a therapy (for example, beliefs about the causes of an illness), disparities in comorbid diseases that make aggressive therapy inappropriate (such as differences in obesity rates among groups), lack of convenient access to therapy, racism, and socioeconomic discrimination. For example, Haggstrom et al. (2005) found that 33 percent of blacks and 23 percent of Hispanics got less than minimum expected care compared with whites. Lund et al. (2008) showed that 7.5 percent of black women in metropolitan Atlanta in the year 2000 who were diagnosed with a localized, curable breast cancer did not get surgical treatment within the first year of diagnosis, compared with about 2 percent of whites. Disadvantaged African Americans who have colon cancer tend to get treated in hospitals that are overcrowded and stressed, where pathologists will look at just 3 to 5 lymph nodes for signs of cancer rather than 18 to 24, Brawley said. As a result, the belief took shape that black cancer patients had more aggressive colon cancer. “It was actually a problem with the staging because of economics,” he explained. In military health systems, in contrast, where the members of different groups are treated the same way, diseases in blacks progress the same as in whites.
Using data from the early 1990s, Tejeda et al. (1996) showed that only 2.4 percent of non-Hispanic whites, 2.6 percent of non-Hispanic blacks, and 4.2 percent of Hispanics who had cancer were participating in clinical trials. “There is a shortage of people going into clinical trials of all races,” Brawley said. More recent data similarly show that participation in clinical trials depends more on the nature of the study (for example, treatment versus prevention trials), the study population (for example, pediatric versus adult), and the reputation in the community of the enrolling center than on race or ethnicity. In addition, many minorities get their care from institu-
tions that cannot afford to offer clinical trials. He further noted that “In a situation where clinical trials are offered and available, if offered and available, minorities are just as likely to say yes as majorities. There is no racial difference if you are from an institution that that person trusts.”
As health care enters an era of precision medicine, the genetic markers a person has will be much more important than a person’s race or ethnicity, said Brawley, concluding that “It is not the width of one’s nose or the color of one’s skin. It is what genes are active.”
Amelie Ramirez, professor of epidemiology and biostatistics, founding director of the Institute for Health Promotion Research, and associate director of cancer prevention and health disparities at the Cancer Therapy and Research Center, all at the University of Texas Health Science Center at San Antonio, briefly reviewed the history of clinical trials and efforts to broaden participation in trials.
The origins of clinical trials date back to early ideas on experimentation in the classical era, but the first generally recognized clinical trial was conducted in the middle of the 18th century by the Scottish physician James Lind, who compared three therapies to treat scurvy—a major health problem for the British navy. Over the next 200 years, clinical trials continued to evolve as the concepts of blinded studies, placebo controls, and informed consent were developed.
Following the human experimentation conducted by the Nazis in World War II, the Nuremberg Code was developed in 1947 to protect human subjects through informed consent and the concept of minimizing harm. Nevertheless, abuses continued to occur, such as the Tuskegee syphilis experiments that were initiated in the early 1930s and continued until 1972. The Declaration of Helsinki1 in 1964 and the establishment of institutional review boards2 (IRBs) in 1974 marked major shifts in the consideration of how to balance the risks and benefits of participating in clinical trials.
In 1979 the Belmont Report identified minority populations as vulnerable research participants and recognized that many vulnerable groups may be excluded from research.3 Since that time, improving minority and female
1 The 18th World Medical Assembly adopted a series of guidelines to guide physicians and medical researchers to protect human subjects involved in research.
2 IRBs are organized standing committees of researchers, administrators, physicians, and so on that review all proposed research projects to ensure the protection of human subjects.
3 The Belmont Report was a set of ethical principles and guidelines to ensure protection of human subjects in research. The report was created in 1979 by the National Commission for the Protection of Human Subjects of Biomedical and Behavioral Research for the Secretary of Health, Education, and Welfare.
representation in clinical trials has been a focus of attention, as highlighted by the 1993 NIH Revitalization Act. However, said Ramirez, “While we have continued to make it a priority to include women and minorities in clinical trials, we still struggle to figure out the best way to accomplish this.”
As an example of the progress that has been made and how much remains to be done, Ramirez cited a report by the Public Health Service Task Force on Women’s Health Issues (1985), which raised concerns about the lack of research on women’s health, the poor quality of health information available to women, and the poorer quality of health care that results. A 2010 report from the IOM (2010) documented major strides that have been made in the areas of breast cancer, cervical cancer, and heart disease. But that report also identified conditions where progress is still needed, including unintended pregnancy, maternal morbidity and mortality, autoimmune diseases, alcohol and drug addiction, lung cancer, gynecologic cancers (non-cervical), nonmalignant gynecologic disorders, and Alzheimer’s disease.
In the past, minority underrepresentation in clinical trials has been attributed to many factors, including unwillingness to participate, lack of opportunity, medical ineligibility, lack of flexibility in child care or employment, and distrust. However, a meta-analysis conducted by Wendler et al. (2006) of 20 health research studies involving more than 70,000 individuals found that individuals from minority groups were more willing to participate in health research than the majority population. However, individuals from minority groups were less likely to be offered enrollment. “This is something that we need to continue to look at,” said Ramirez. “We must continue to understand why minority groups are underrepresented and examine our biases around this issue.”
At the time of the workshop, noted Ramirez, the National Cancer Institute (NCI) had more than 8,000 clinical trials that were accepting participants. However, though Latinos represent 17 percent of the U.S. population and are the largest minority group in the United States, they represent only 5 percent of the participants in NCI treatment trials.
Less is known about effective strategies to recruit Latinos into clinical trials than for other groups, said Ramirez. Differences in eligibility criteria, expected clinical outcomes, and geographic availability all raise questions. Furthermore, the issues can be different with early-phase clinical trials than with later phases.
With funding from NCI, Ramirez and her colleagues have studied early-phase clinical trials to identify cultural, economic, and environmental barriers to participation in these trials and to identify key components of an intervention to reduce these barriers and increase participation. Among
health care providers, the barriers for recommending participation in trials include logistical factors, such as the time and effort involved in explaining protocols to patients, and personal factors, such as worries about the loss of control over patient care (Ramirez et al., 2012). Among patients, prominent barriers were the lack of knowledge about the disease and treatment; cultural, language, and literacy issues; a lack of discussion with their doctors about the option; and the costs, travel, and insurance issues involved in participation (Chalela et al., 2014). For example, many trials are offered at centralized cancer centers, and people living in rural areas do not necessarily have access to these trials. For this reason, local providers need the education and training to provide quality treatments, Ramirez said. Also, people without the insurance to cover basic screening costs usually cannot participate.
Yet when patients are told about the option of participating in a clinical trial, many are eager to do so. Ramirez quoted one patient who said, “I did not know cancer clinical trials were an option for me until my doctor told me about them. . . . I thought they were for very ill people. But now I know there are clinical trials for all stages of breast cancer.”
Ramirez explained that when patients were asked about what would enable them to participate in a clinical trial, several factors were prominently cited, including
- Trusting the doctor
- Trusting the trial center
- Feeling that joining a trial will give hope and help future cancer patients
- Having clear information
- Encouragement from family members
For example, one patient said, “To know that every new medicine goes through a clinical trial puts me and my family more at ease and gives us more options if we get sick.”
In a test of three registry recruitment methods among South Texas Hispanics into the Cancer Genetics Network, one randomized group received a letter from their doctors asking them to participate, another received the letter plus a culturally tailored bilingual brochure, and the third received the letter, brochure, and interpersonal contact to urge them to participate. The result was that extra information and interpersonal contact increased accrual, reported Ramirez.
In another experiment, an in-clinic patient navigator/clinical research associate sought to increase accrual to pediatric cancer trials in south Texas (Wittenburg et al., 2010). The result was that accrual rose from 38 in 2007 to 118 in 2010, which is the highest local total ever.
Finally, in a study ongoing at the time of the workshop, Latinas were empowered to make informed decisions regarding breast cancer clinical trials through enhancement of knowledge, attitudes, and skills; increases in self-efficacy; and encouragement to discuss clinical trials as a potential treatment option with doctors and family members. Preliminary results showed that the proportion of Latina breast cancer patients taking steps toward participating in a clinical trial—through asking their doctors about clinical trials, talking with family and friends about participating, and considering the pros and cons of participating—was significantly higher in the intervention group than in the control group. Particularly important, said Ramirez, was informing patients about clinical trials before they talked with their doctors, because providers have limited time to discuss the possibility of clinical trials. Empowering interventions that enhance patients’ awareness and self-efficacy foster a sense of control and provide patients with the knowledge and skills they need to make informed decisions regarding treatment options.
Ramirez made two additional suggestions to increase Latino participation in clinical trials, based on her experiences and observations. Computer-based videos, if specifically tailored to Latina breast cancer patients, are a particularly effective strategy to increase patients’ knowledge and understanding of clinical trials and to promote their participation in clinical research. Also, other populations are underrepresented in clinical trials and need to be considered for inclusion, including individuals with disabilities, the aging population, and gender difference groups.
During the discussion, workshop participants and the presenters continued to talk about what will happen as precision medicine continues to advance. Brawley noted that, today, someone who is dark skinned or of Mediterranean ancestry is more likely to have a genetic variant that would suggest how best to treat a urinary tract infection. “That is what I would call benign racial profiling,” he said. However, as genomic testing spreads, more people will receive tailored therapy depending on exactly what genetic variants they have. Treatments will then be based not on skin color but on an individual’s genomics. One drawback to a genomic approach is that it will be more expensive, Brawley said, than making judgments based on a person’s appearance.
As Rotimi noted, appearance is a proxy that can be useful, but it gives only a vague idea of what genetic variants a person might have. “It would definitely have been better if you had the genetic variant and you made a decision and you do not have to use the phenotypic characterization to do that. We are not there yet. That is where we are moving,” he said.
According to Ramirez, increasing minority accrual in clinical trials is critical to the development of novel therapeutics, including personalized therapies. Without adequate representation of minorities, researchers cannot assess the differential effects among groups, nor ensure the generalizability of trial results.
Rotimi also pointed out that the optimal population to enroll in a clinical trial depends on the questions being asked. For some questions, a genomically uniform population might be better, while for others a heterogeneous population would be better. But representation needs to be done carefully. Having just a few members of a population group in a trial does not provide sufficient statistical power to say anything meaningful about that population. The trial organizers may satisfy a political ideal but not produce scientific results. At the same time, diversity is not an illusion. “If we want to represent [diversity] from a genetic point of view, a social point of view, a cultural point of view, we need to broaden the representation of the people who are at the table,” he concluded.
This page intentionally left blank.