Ethnicity & Health

2012, 1!15, iFirst article

RESEARCH ARTICLE
Concepts of race and ethnicity among health researchers: patterns and
implications
Roberta D. Baera*, Erika Arteagab, Karen Dyera, Aimee Edena, Rosalyn Grossc,
Hannah Helmya, Margaret Karnyskid, Airia Papadopoulosa and Doug Reesera
a
Department of Anthropology, University of South Florida, Tampa, FL, USA; bFederación de
Organizaciones Campesinas, Indı́genas y Negras del Ecuador (FENOCIN), Quito, Ecuador;
c
School of Nursing, Florida Gulf Coast University, Ft. Myers FL, USA; dDepartment of
Downloaded by [University of South Florida] at 19:56 18 August 2012

Anthropology, San Diego Mesa College, San Diego, CA, USA

(Received 23 August 2011; final version received 10 July 2012)

Objectives. This study adds to the discussion of appropriate categories of analysis

in health research. We contribute data based on actual interviews about the
concepts of race and ethnicity, conducted among a broad range of US health
researchers.
Design. In-person qualitative interviews were conducted with 73 scientists at two
health research institutions, one that focused on public health research, and one
that focused on research about a specific disease. This represents a larger and
more interdisciplinary sample of health researchers than has been previously
interviewed about these topics.
Results. We identify a core model of how race and ethnicity are understood. The
respondents were confused about the concepts of race and ethnicity and their link
to genetic differences between populations; many treated these concepts as
interchangeable and genetically based. Although ethnicity was considered some-
what more socially constructed, it was often felt to cause unhealthy behavior. In
addition, the situation is not improving; the younger health researchers tended to
put a stronger emphasis on the genetic aspects of race than did the older health
researchers.
Conclusion. Unlike reviews of how these concepts are used in scientific
publications in which race and ethnicity are often undefined, our face-to-face
interviews with these researchers allowed an understanding of their concepts of
race and ethnicity. Building on their actual perspectives, these data suggest
alternative approaches to formal and continuing educational training for health
researchers. We recommend beginning with discussions of human diversity, and
then moving on to what race and ethnicity are ! and are not.
Keywords: race; ethnicity; health researchers; qualitative research

Introduction
Recent years have witnessed a re-emergence in the USA of a genetic concept of race
in both clinical and public health research on health and health outcomes (Frank
2007). This is due in part to contemporary genetic findings, including new discoveries
about the human genome. This trend of interpreting racial categories as reflecting
genetics has led to a call for the National Institute of Health (NIH) to re-evaluate the

*Corresponding author. Email: baer@usf.edu

ISSN 1355-7858 print/ISSN 1465-3419 online
# 2012 Taylor & Francis
http://dx.doi.org/10.1080/13557858.2012.713091
http://www.tandfonline.com
 2 R.D. Baer et al.

use of the term race, due to the concept’s lack of scientific validity (Oppenheimer
2001). Support for this position comes from the work of anthropologists who claim
that the US system of categorization is based on ‘folk’ categories, which have no link
to genetic variation (Goodman 2006, Marks 2006). Other support for the re-
evaluation of the use of this system of categorization comes from studies of how
these concepts are used in journal articles reporting health research (Anderson and
Moscou 1998, Comstock et al. 2004, Lee 2006, Fullerton et al. 2010, Megyesi and
Hunt 2011, Rachul et al. 2011). A final line of evidence comes from actual interviews
with health researchers about their understandings of these concepts (Hunt and
Megyesi 2008b, Bliss 2011).
This research adds to the discussion of appropriate categories of analysis in
health research. We contribute data based on actual interviews about the concepts of
race and ethnicity, conducted among a wide range of types of US health researchers.
Downloaded by [University of South Florida] at 19:56 18 August 2012

We interviewed a larger and more interdisciplinary sample of health researchers than

has been done previously, and found that respondent understandings of race and
ethnicity were confused and inconsistent. Based on these findings, we also call for a
re-conceptualization of appropriate analytical categories for health research on the
part of the NIH and other research agencies in the USA. In addition we support the
efforts of a number of scientific journals to require definitions and justifications of
how racial terms are used in published reports (Kaplan and Bennett 2003). However,
in order to do this, present and future generations of health researchers will need
training on the nature of human diversity. Our findings suggest specific recommen-
dations as to appropriate content of education programs for health researchers on
these topics.
In some areas of Europe (such as the UK), classification of populations in health
research has moved to a focus on social class, and more recently, ethnicity (although
there are some who emphasize the problems with the use of ethnicity as a category
for health research [Bradby 2003]). However, in the USA, the focus has historically
been on race/ethnicity, as classified by the US Census categories. We evaluate the
views of US health researchers on these topics in light of contemporary anthro-
pological perspectives on the role of race and ethnicity in health problems. These
perspectives criticize the use of race as a category with any genetic/genotypic basis;
however, they do affirm the role of race in phenotypically based health differences
and outcomes. In contrast, the term ethnicity, as used by anthropologists, refers to
groups of people who share learned aspects of their culture, such as language, place
of origin, religion, values, and beliefs, and who view themselves as a group (Smedley
and Smedley 2005). Ethnicity is considered to be fluid and not related to human
genetic variation.
We begin with some caveats ! our goal here is not to enter into the debate on the
appropriate term for categorization of human differences for health research, race, or
ethnicity, nor do we wish to address more theoretical aspects of cultural competency
training (see Gregg and Saha 2006). We do consider the extent to which, and the
reasons why, health researchers consider race and ethnicity useful categories for
health research. Our focus is on those involved in health research, as we assume a
relationship between their assumptions and the design of their research, the
subsequent findings, and ultimately, policies and programs based on those research
results. However, much of the previous research on this topic has not been based on
actual interview data with health researchers. Nor have those not involved in
 Ethnicity & Health 3

genetically focused research been interviewed, despite the fact that the US Census
categories are used widely in many different types of health-related research. Thus,
we do not infer the perspectives of health researchers based on examination of their
scientific publications, and we expand the scope of types of health researchers
studied. Our data on the actual perspectives of health researchers on these concepts
(as opposed to how their assumptions shape their research) contribute to the growing
critique of the use of US Census categories for health research. In addition, these
data will make it possible for both formal and continuing educational programs for
these scientists to be based on their actual views. This will facilitate the creation of
more effectively designed programs to address aspects of their perspectives that are
inaccurate.
Downloaded by [University of South Florida] at 19:56 18 August 2012

Background
The genetic concept of race is usually operationalized in health research in the USA
based on groups as categorized by the US Census (Outram and Ellison 2006). These
groupings follow the 1997 Office of Management and Budget (OMB) standards on
race and ethnicity. Five minimal racial groups, related to geographical origins, are
offered: White, Black or African-American, American Indian or Alaska Native,
Asian, and Native Hawaiian or Other Pacific Islander. In addition, respondents
choose one of two ethnicities: ‘Hispanic or Latino’ and ‘Not Hispanic or Latino’ (US
Census Bureau 2012).
Using these categories, studies focus on racial links to disease, often in the
context of concern over increasing health disparities (Frank 2007). Another factor
contributing to the revival of a focus on race is the concern in the USA with the
inclusion of minorities in health research. In 1993, the NIH began requiring that all
researchers report the racial/ethnic backgrounds of study participants based on the
US Census categories (Office of Extramural Research, NIH 2001). As such, studies
are commonly designed with the assumption that the US Census categories are
meaningful units of analysis for the examination of health problems (Stevens 2003).
Paradoxically, and parallel to these developments in health research, social
scientists (particularly anthropologists) have increasingly challenged the concept of
race as a useful unit of analysis (Goodman 2006, Marks 2006). Anthropologists have
argued for a sociocultural perspective of race, suggesting that race is not a genetic
category, but rather, a ‘folk’ category, i.e., one that may be in popular use,
particularly in the USA, but which has no scientific basis (Goodman 2006, Marks
2006). Anthropologists also continue to stress that categories based on the OMB
divisions of the US Census are not a basis on which to do health research (Goodman
2006, Marks 2006). With respect to patterns of health disparities that seem to follow
the lines of these categories, Gravlee (2009) and Kuzawa and Sweet (2009) note the
role of patterns of discrimination based on racism. The key argument is that the basis
of differences in health outcomes lies predominantly in the social experiences, rather
than genetic similarities among people so classified. Thus, while race is not
genetically based, due to epigenetic processes driven by experiences of racism and
discrimination, these social classifications may come to have biological effects
(Kuzawa and Sweet 2009, Gravlee 2009).
However, despite the evidence discrediting the meaningfulness of race as an
analytical category, it continues to be used extensively in health research, and its
 4 R.D. Baer et al.

meaning and definition are rarely clear. The second type of critique of the analytical
use of race comes from reviews of how this term is used in published health research.
A review of articles published in two premier public health journals between 1996
and 1999 found that 77% referenced race or ethnicity, and that an ‘enormous
diversity of terms was used to describe the concepts of race and ethnicity’ (Comstock
et al. 2004, p. 611). Health researchers rarely differentiated concepts of race and
ethnicity, nor did these researchers report the criteria used to assess these variables
(Comstock et al. 2004). Other content analyses of biomedical research journal
publications (Anderson and Moscou 1998, Lee 2006, Fullerton et al. 2010, Megyesi
and Hunt 2011) also indicate that race and ethnicity are rarely defined, nor are
differences between conceptions of race and ethnicity addressed. Particularly
troubling is the extent to which racial/ethnic differences are attributed to genetic
differences between groups, despite the fact that the studies reporting these findings
Downloaded by [University of South Florida] at 19:56 18 August 2012

did not include genetic data (Megyesi and Hunt 2011). Finally, Rachul et al. (2011)
find similar patterns in published journal articles, and even more distortion when
these results are reported in the popular media. A limitation of all of these studies is
that the researchers were not interviewed; analyses were based solely on their
published articles. Since race and ethnicity are usually not defined in the publications
reviewed, we cannot be sure how the researchers actually understand these concepts.
Studies that actually interview researchers to explore their understandings of race
and/or ethnicity are somewhat limited. In a study of professors of biology and
anthropology and the textbooks they utilized, Lieberman et al. (1992) found that
67% of biologists accepted the concept of human genetically based races, as opposed
to 50% of physical anthropologists. However, a content analysis of texts between
1936 and 1984 showed significant changeover time in physical anthropology texts
when compared with those used in biology. Morning (2006) conducted a similar
study in 2001 and 2002, interviewing biology and anthropology professors. Forty
percent of the professors considered each race to have innate inherited genetic traits,
while 60% viewed racial groups as being socially constructed for political, economic,
and other purposes.
Actual studies based on interviews with health researchers have been limited to
geneticists. One study of geneticists associated with the editorial boards of the
leading genetics journals (Outram and Ellison 2006) found that these scientists had
difficulty in differentiating between race and ethnicity and they were hard-pressed to
define these terms. The geneticists were not ‘convinced that the criteria used to
classify racial or ethnic groups provide the most appropriate markers of genetic
variation’ (Outram and Ellison 2006, p. 167). They viewed these categories as
‘unreliable, invalid, and politically sensitive’ (Outram and Ellison 2006, p. 168). Yet
curiously, the geneticists advocated continued use of these categories for scientific
study of human variation due to ‘very good evidence of biological differences
between racial and ethnic groups’ (Outram and Ellison 2006, p. 168). Hunt and
Megyesi (2008b) interviewed a convenience sample of 30 human genetics scientists
and focused specifically on how the geneticists classified the individuals they studied.
The findings indicated that the geneticists used arbitrary and colloquial terms to
describe the racial/ethnic groups in their research samples. And most recently, Bliss
(2011) interviewed 36 of the world’s most elite genomics researchers. She found that
they ‘paradoxically advance social explanations for race, while asserting genomics as
a plausible solution to racial dilemmas’ (Bliss 2011, p. 1019).
 Ethnicity & Health 5

In this paper, we address some of the research needs remaining in this area.
Gravlee and Sweet (2008) call for additional ethnographic research on the hidden
assumptions health researchers have about race. In addition, they cite the need for
studies that clarify the relation between race and ethnicity. As such, this study builds
on earlier work, but also goes beyond it and looks more broadly at the range of
scientists involved in health research. We conducted in person interviews with an
interdisciplinary group of health researchers who conduct a wide range of public
health and specific disease-focused health research. Using open-ended, ethnographic
research, we explored their actual concepts of race and ethnicity, as well as what
relationships, if any, they saw between them.
We assumed that their understandings do shape their research and practice, but
we did not focus on how they operationalize these concepts in their own research.
Instead, our goal was a more basic exploration of the beliefs that health researchers,
Downloaded by [University of South Florida] at 19:56 18 August 2012

other than geneticists, hold about these concepts and their utility. While information
on how these concepts of race and ethnicity are used in research is helpful, these
concepts are usually not defined in the publications that report the results of those
studies (Anderson and Moscou 1998). How then can we know what the researchers
mean by these terms? How would they define them? Understanding how health
researchers actually view these concepts makes it possible to design more effective
educational programs for them. Our data also provide additional justification to
support the call for a re-conceptualization of how human diversity is categorized in
health research in the USA, particularly in studies funded by institutions such as the
NIH.

Methods
Samples and sampling
We studied two health research institutions in a multi-ethnic/racial metropolitan area
in the state of Florida, one that focused on public health research and programs
(‘PH’), and one that focused on research and patient treatment of a specific disease
(‘DF’). We chose these institutions because of the interest in the project by senior
administration at both institutions who were committed to improving the abilities of
their institutions to do research with diverse populations. In addition, these two
institutions provided a very broad group of health researchers to study; both
institutions were extremely interdisciplinary, and our sample included clinicians,
basic scientists, and researchers involved in community-based work.
Respondents were chosen by random sampling so that the results could be
generalized to the institutions as a whole. For the DF participants, three types of
work activities were identified: basic research, clinical activities, and community
outreach. The desired N was 50 out of a total population of 380. For the PH
participants, a list of all researchers (N "74) was randomly sampled according to
department, including one focused on community outreach.
Semi-structured interview guides were developed in cooperation with the
participating institutions. Topics that were addressed through open-ended questions
included definitions of diversity, race, and ethnicity, and perceptions of the role of
these factors in health care and health outcomes. Due to the schedules of these
researchers, we chose semi-structured face-to-face interviews as the most effective
 6 R.D. Baer et al.

way to gather the data with the least respondent time burden and inconvenience.
Participants were interviewed in their offices/laboratories for approximately 30 min-
utes. All of the interviewers were ‘White’ graduate students, with the exception of one
who was ‘African-American.’ We did not observe any affects of interviewer
background on responses.
In total, 73 interviews were completed, 49 from DF (67%) and 24 (33%) from PH.
Half of the total interviewees were male. Ages ranged from 29 to 74, with DF
researchers tending to be younger than their PH counterparts. Ethnicity of the
interviewees was self-identified. Half of the total sample identified as Caucasian and
did not specify any additional information about their heritage. The others gave
more detailed information about their backgrounds; half of this group was Asian or
South Asian (n"15), and a fifth were from Latin American or the Caribbean (n"8).
The remainder (n"14) were members of other ethnic groups that included African,
Downloaded by [University of South Florida] at 19:56 18 August 2012

Jewish, Greek, Italian, Eastern European, Ukrainian, and American Indian. The
majority of the total sample held master’s or doctoral degrees, and a third were
physicians (MD/DO). A third reported direct involvement in research related to
diversity, a quarter reported indirect involvement in such research, another quarter
were involved in unrelated research, and five individuals reported no involvement in
research. Those in clinical positions commented that the lack of diversity in the
patient population at their institution was the reason for this lack of research related
to diversity. The community outreach groups at each institution had the greatest
focus on research involving diverse populations.

Data analysis
Data were hand-coded based on recurrent themes in the responses. After all authors
agreed on an overall coding scheme, each interview was then reviewed by at least two
of the authors to maintain consistency in coding. As such, inter-rater reliability was
achieved by consensus, but not formally measured. Responses were coded using the
categories as shown in Table 1.

Results
Concepts of diversity
We began by asking the health researchers to define diversity. This was difficult for
them, and their responses focused on the categories of the US Census:

[Diversity] is a term adopted by the Supreme Court so they didn’t have to use the word
race.

Diversity means the major racial groups that we know about, classically White,
Hispanic, a variety of Asians, and African Americans.

Concepts of race
We next moved to an exploration of concepts of race, as well as perceptions of how
racial differences affect health care, and the utility of the concept of race in health
research. One quarter of the respondents described race as genetically based, ‘. . . race
 Ethnicity & Health 7

Table 1. Response categories for questions.

Code Possible inclusive responses

Genetics Genetics, ancestry, family history, personal medical history

Socioeconomic status Education, schools, income, poverty, money, economics, health
insurance
Individual choice Lifestyle, diet, nutrition, sexual orientation, personal practices,
habits, personal choices, exercise, drinking, smoking, drug use,
promiscuity, culture as a barrier to healthy behavior
Environment Birthplace, geography, residence, occupation, exposure to
toxins, availability of clean water
Culture in the Religion, social resources, family influence, community, socio-
anthropological sense psychological factors, beliefs, values, acculturation
Downloaded by [University of South Florida] at 19:56 18 August 2012

Structural Providers, access, transportation, location, quality of care,

health system biases, lack of diversity in health system
Racism Discrimination, unequal/unfair treatment
Demographics Age, gender, marital status, disability
Census Categories as defined by the US Census Bureau
Not race Racial bias does not cause racial differences in health

and ethnicity are the same, but genetic predispositions to diseases occur based on a
racial background.’
Another quarter of respondents defined race as synonymous with physical traits,
‘the official government definition of the color of your skin.’ Finally, approximately
one quarter of respondents felt race was a social construct, ‘. . . I don’t think race is
genetically determined. It’s socially determined and I think we need to rethink how
we define race because often times it is looked upon as the color of one’s skin.’ While
some respondents agreed with Census Bureau definitions of race (for example:
‘Census defines it as Black, White and Asian, and that’s also the way I define it’),
10% of respondents were unable to define race.
Most respondents felt that racial differences affect health care, particularly
through structural issues (health care providers and institutions) and socioeconomic
status (SES). Individual behaviors associated with different racial groups were also
mentioned:

Certain races don’t tend to present until they are much sicker, and they don’t get
appropriate screening. Poorer populations have poorer diets and tend to eat more fast
food.

Sometimes certain races avoid certain medical procedures ! a certain belief system may
include avoiding going to the doctor.

Three-quarters of the sample felt health services should be provided differently to

people of different racial backgrounds under some circumstances. A quarter of these
respondents interpreted racial differences as cultural differences; ‘In many cases, the
physician should be aware of ethnic differences, and how people look differently at
diseases.’ About a quarter focused on genetic differences:
 8 R.D. Baer et al.

Genetics is very important as it is linked to race.

Genetic changes; maybe some genes have differential expression depending on race.

Some researchers felt that race was not a justification for differences in health
services, ‘Differences [in health services] should be based on different needs and risk
factors . . . but not on races.’ Only a few mentioned racism or the environment as
reasons why health services should be provided differently to people of different
backgrounds.
Of the 73 participants at the two institutions, most felt that the concept of race is
useful in health research. However, 86% at DF found race was useful in health
research, compared with only 58% at PH. Among DF respondents who felt that race
is a useful concept in health research, most identified a genetic basis for their views.
Downloaded by [University of South Florida] at 19:56 18 August 2012

Some expressed the idea that ‘It depends on the research; if the research is based on
genetics, then race is important.’ Others noted that ‘you have to consider that certain
races have a genetic predisposition to certain diseases.’ Still others simply stated that
‘races are genetically different.’
Only a few DF respondents did not feel that race was a useful concept in health
research. Among the reasons they noted for their views were that, ‘[Race] serves as a
proxy for socio-economic status.’ Others suggested ethnicity as a unit of analysis:

We should be using ancestry, because it gives you a more individualized perspective to

look at differences between people.

Ethnicity is far more important than what is defined by others as the color of your skin.

Finally, others worried about political correctness:

Politically, it is a sensitive topic in the press ! no one wants to be quoted on the evening
news saying that race doesn’t matter because there will always be a segment of the
population that doesn’t understand what you’re trying to say ! no matter how correct it
may be.

Half of the PH respondents felt that race is a useful concept in health research;
however, their reasons differed from those given by DF respondents. There was less
of a focus on genetics in general, as well as on genetic predispositions to disease.
When genetics was mentioned at PH, it was presented as one of multiple variables to
consider:

We should tend to integrate not just race, but manage race and ethnicity.

We should make health research more comprehensive ! not just genetics, but also
cultural backgrounds.

The importance of SES was also mentioned by a quarter of PH respondents, ‘We

need to look at how well our systems serve our populations. Socio-economics is a
much more valuable factor in research, but it does correlate highly with race and
ethnicity.’ Differences in responding that race was a useful concept in health research
were related to age, with all respondents under the age of 45 responding affirmatively
 Ethnicity & Health 9

that race was a useful concept in health research, versus only about half of those 45
and older.
PH researchers who did not feel that race was a useful concept (about a third)
also mentioned a variety of reasons to explain how and why, such as explanations
related to power relations: ‘Race is one of the ways that the mainstream culture
continues using others for their own gain ! it is more an issue of power than anything
else.’ Another respondent noted that race ‘puts people in categories artificially
because we have so many blended ethnicities now. If I put people in categories, it
would be based on education level and income.’ One noted that researchers do not
have much choice but to use race in their studies:

We’re sort of stuck with it after so many years. The only surrogate we have for income is
education, which may or may not be true. It’s pretty messy and our governmental
Downloaded by [University of South Florida] at 19:56 18 August 2012

organizations still report all their health statistics by race and not by education. We’re
trying to get rid of it.

In summary, more respondents at DF saw race as a useful concept and tied to

biology and genetics. Fewer of those at PH viewed race as a useful concept; they
focused more on variables such as cultural background and SES. The older age
group (age 45 and older) at the PH institution was the least likely to find the concept
of race useful. When looked at by gender, we found that almost all DF males found
the concept useful. Finally, PH respondents were more likely to feel that health
services should be tailored to those of differing racial backgrounds.

Concepts of ethnicity
Our next goal was to determine if the researchers distinguished differences between
race and ethnicity. Respondents were asked to define ethnicity, as well as discuss the
effects of ethnic differences on health and health care, and the utility of the concept
in health research. For most respondents, it was not easy to distinguish race and
ethnicity:

Race and ethnicity are blurred concepts.

I know it is different from race, but I am not sure as to how.

There is a fuzzy line between race and ethnicity.

Some people try to use ethnicity rather than race, [be]cause race is a very political term.

I don’t differentiate between the two [race and ethnicity] . . . there is a political tone at
play here and there is need to be careful.

[Ethnicity is] more cultural than race but is mostly four categories: Hispanic, non-
Hispanic, Jewish and non-Jewish.

About half of the definitions of ethnicity included some reference to the social
construction of the term:
 10 R.D. Baer et al.

Ethnic groups are defined by common culture and origins distinct from race.

Religious and social background, belief system, religious affiliation.

The cultural upbringing of the patient.

However, genetic differences, physical traits, and race were mentioned as part of the
ethnicity concept in a quarter of the total responses:

Genetic background of people . . . who were the parents and grandparents and what sort
of ethnic groups did they have in their origins.

Same as race, redundant. [The] rules[are]: Black, White, Hispanic, etc.

Downloaded by [University of South Florida] at 19:56 18 August 2012

Most respondents felt that ethnic differences affect health. About half related the
effect of ethnicity on health to genetic differences, a fifth to physical traits, and about
a third related it to individual behavior:

Ethnicity is part of culture and culture affects behavior and diet, which affect health.

Eating, lifestyle, religion, Christian background ! use of alcohol can cause a

predisposition to certain types of cancer.

And some respondents felt that ethnic differences affect health through culture in an
anthropological sense:

We all have different beliefs about what causes illness, how to treat illness, we all have
some tradition in that but also how we access the health care system and what that
system’s expectations of us are and all combinations therein.

Ethnicity affects health when people perceive they should get healthcare and how, what
is normal in that population, when they are ill and when it is appropriate to seek
attention.

Most interviewees indicated that ethnic differences do affect health care. About a
quarter attributed this to individual behaviors, ‘Everything about a person’s culture
affects their health.’ Another quarter of responses focused on socioeconomic issues,
‘Ethnic groups are related to education and SES. If you do not have the money to
pay, then you do not get good health care.’ And another quarter of responses cited
structural concerns, ‘Patients are likely to identify well with doctors and to obtain
more continuous healthcare from a same-ethnicity and race provider.’ Some cited
racism as the basis for ethnicity affecting health care, ‘People discriminate based on
cultural differences. Societies discriminate, and this affects healthcare.’ And a few
respondents noted that ethnic differences affected health because of genetic
differences between groups.
A majority of respondents felt that health services should be provided differently
to different ethnic groups, focusing on the role of behavior patterns, ‘To an extent,
ethnicity results in different behavior patterns and [because] certain health issues [are]
more relevant to certain ethnic groups.’ A third of responses noted structural issues,
‘The provider needs to have an understanding of an individual’s belief systems, values,
 Ethnicity & Health 11

and attitudes . . . to provide appropriate care and complete care.’ Some noted genetic
differences perceived among groups, ‘Health services should be provided the same to
everyone, but the genetic piece needs to be looked at, as it makes a difference.’
Most respondents attributed the usefulness of ethnicity in health research to the
need for better understanding of people’s behavior, ‘We need to understand it. There
is no one size fits all. We need to be sensitive to beliefs, behaviors, meanings of illness,
language, interpretation of behaviors.’ One participant noted the role of diet, ‘The
particular diet that a group may eat can influence health or disease. Certain ethnic
groups eat certain diets.’ A quarter of responses cited genetic differences among
populations as affecting health care:

Ethnically speaking, there are certain populations that have a predisposition to certain
diseases based on the disease’s genetic expression.
Downloaded by [University of South Florida] at 19:56 18 August 2012

It involves the genetic backgrounds of people so yes, it is.

A few thought structural issues were important, but only a few noted racism or SES,
‘So we can better understand why people don’t get treated as well as others, and we
can try to understand differences to intervene and hopefully reduce or eliminate
those differences.’

Summary ! a core model

These data suggest a core model of how this population views these issues.
Conceptions of diversity were vague, centering on US Census Bureau classifications.
Race and ethnicity were also not well understood; respondents were not clear on the
differences between the concepts of race and ethnicity, the extent to which they
overlap, and the extent to which either of them is genetically based. Ethnicity was
considered somewhat more socially constructed, while race was considered to be a
genetic fact. Ethnicity was often felt to affect health either through genetics or by
culturally influencing individuals to unhealthy behaviors. The view that individual
behavior and/or culture are barriers to positive health outcomes was greater among
PH researchers. Overall, ethnicity was considered a useful concept because it enables
an understanding of why and how individual behavior and/or culture can be barriers,
as well as a deeper understanding of genetics. Interviewees did not mention the role
that ethnically based beliefs and behaviors can play in a healthy lifestyle; the
perspective that ethnically based behaviors can also be health protective was largely
absent. And very few respondents noted structural factors as an important variable.

Discussion and conclusions

Our results contribute to the small body of interview-based data on how health
researchers understand race and ethnicity, as well as relationships between the two
concepts. Building on the work of Hunt and Megyesi (2008b) and Bliss (2011), we
broadened the types of health researchers interviewed to a multidisciplinary group of
respondents working in both public health and more disease-focused research, and
included clinicians, basic scientists, and researchers involved in community-based
research.
 12 R.D. Baer et al.

Our first finding is a core pattern of confusion and inconsistency in how race and
ethnicity were understood. Elements of this pattern parallel previous work on these
topics. In a pattern similar to that found by Megyesi and Hunt (2011), half of the
researchers in our study considered race to be related to genetic background; some
expressed this idea in terms of phenotypic characteristics. These results are also
consistent with the studies of Lieberman et al. (1992) and Morning (2006), that
found race to be genetically based. Eighty-two percent of our respondents considered
the concept of race useful for health research due to the link between race and
genetics; Outram and Ellison (2006) found similar beliefs among geneticists. The
arbitrary nature of how race and ethnicity were categorized, as described by Hunt
and Megyesi (2008b) and Bliss (2011), was also seen among our respondents.
Although some research has suggested a switch in focus from race to ethnicity in the
Downloaded by [University of South Florida] at 19:56 18 August 2012

USA (Afshari and Bhopal 2010), in contrast, we found that our respondents
generally considered ethnicity to be the same as race. In addition, the perspective of
many of these researchers tends to focus on the individual as opposed to the social
context. The emphasis on the behavior of the individual and the role of genetics in
health, including applying genetic explanations to population-level health differences
(Pearce et al. 2004), is likely related to the low salience of issues such as racism,
discrimination, the role of environmental toxins, and other structural issues (Megyesi
and Hunt 2011).
A second contribution of this study is that the data indicate that the situation is
not improving; rather, the reverse seems to be the case. Younger researchers tended to
put a stronger emphasis on the genetic aspects of race, as did those at DF. These
patterns suggest the need for somewhat different approaches to continuing training
for these different types of scientists. DF researchers, in particular, need a deeper
understanding of the lack of links between the US Census categories and the
genotype. PH researchers tend to focus on the importance of individual behavior and
culture; therefore, programs for this group should address the extent to which culture
does not necessarily cause unhealthy behavior, i.e., culture is not a barrier to good
health. Both groups of researchers also need a deeper understanding on the role of
discrimination and other social factors in differences in health outcomes.
We do not believe the lack of a contemporary perspective on these issues is in any
way the ‘fault’ of the particular researchers we studied. To the contrary, we see the
root of the problem in the formal education of health professionals and researchers.
Many of those who have gone on to become health researchers received little
background in social science perspectives on race and ethnicity as part of their
professional training, but are forced by national funding agencies to use and
operationalize these concepts in their research. Based on the work of Lieberman
et al. (1992) and Morning (2006), the biologists they most likely studied with as
undergraduates did not impart an accurate view of these concepts either. Thus,
problematic US Census Bureau-based categories continue to be used as the basis for
health research.
A limitation of this study is that, in its focus on the views of the researchers
themselves, it did not explore how the abstract views of race and ethnicity expressed
in the interviews related to the research in which the researchers were actually
involved, and how they designed that research. This research was not designed to
explore these topics due to logistical and other constraints. Given our findings, future
 Ethnicity & Health 13

research is needed to examine how these views are related to the research carried out
by these scientists.
However, these data, based on actual interviews with health researchers, do
enable a better understanding of the diversity of concepts surrounding race and
ethnicity. Based on the confusion and inconsistency we found in the understanding
of these concepts, we echo the calls for a re-conceptualization of the categories
appropriate for health research in the USA. The current stipulation that investigators
fit their study populations to the current categories of the US Census is not having its
intended effect. While this requirement began with the well-intentioned goal of
including of minority populations in health research, the unintended result has been
a further reifying of race, as understood by the Census categories ! as a genetic fact.
Continuing use of these categories perpetuates stereotypes of a link between the
categories and patterns of human genetic variation.
Downloaded by [University of South Florida] at 19:56 18 August 2012

A third contribution of this study is that based on the inaccuracies in the core
model of race and ethnicity that our data identified (as well as the interview data of
Hunt and Megyesi [2008b] and Bliss [2011]), we can identify specific gaps in the
educational training of health researchers. The data on how race and ethnicity are
operationalized in journal articles reveal that conceptions are vague. But the actual
words and concepts expressed in the face-to-face interviews provide important
information about what they think these concepts actually mean. We found that the
categorizations of these researchers are using at present to understand human
diversity are inappropriate. But if they are to be expected to use an alternative
system, such as that proposed for scientific journals (Kaplan and Bennett 2003),
including the definition and justification of terms used to describe study populations,
they will need a solid grounding in the constructivist critique of the race concept.
Training of health professionals and researchers at every level, including pre-
medical, medical, and continuing education, should include more and better training
in these sociocultural concepts, and must be built on the actual perspectives of this
population. A key concept to be addressed is the socially constructed aspects of the
concepts of race and racial categories, as used in the USA, and the inappropriateness
of using these kinds of categories in isolation for analyzing genetically based health
issues (Kaplan and Bennett 2003). In addition, our data suggest that an appropriate
beginning point for these discussions is an exploration of how human diversity is
understood and classified, and what those classification systems mean ! and do not
mean. Discussions should then consider what race and ethnicity are ! and are not
(Baer 2008). Educational programs must address the issues of perceived genetic basis
of racial categories, distinguishing between race and ethnicity, and differences
between the genotype and the phenotype. Relationships and differences between
genetic and sociocultural dimensions should also be explored, as should the validity
of race as a research construct.
Training should enable researchers to critically evaluate the literature on these
topics, and differentiate among what is genetic, biological, and social. In some
critiques of the use of race as an analytical tool made by the social constructionists,
the term ‘biological’ has often been used, when what is really meant is ‘genetic.’ Thus,
assertions that race has ‘no biological basis’ actually refer to the lack of overlap
between the categories of the US Census and genetically based patterns of diversity.
Recent research has actually stressed that although race is not genetic, it can indeed
come to be biological (Gravlee 2009). Health researchers must also be given the
 14 R.D. Baer et al.

analytical tools to evaluate reports in scientific publications that have previously used
inappropriate categories to characterize study populations, such as the literature
noted by Anderson and Moscou (1998), Comstock et al. (2004), Lee (2006),
Fullerton et al. (2010), Megyesi and Hunt (2011), and Rachul et al. (2011). Finally,
training must also include discussion of proposed alternative nomenclatures, such as
geographic or continental ancestry groups.
These issues are of serious import; it has been suggested that the scientific validity
of research done using vague notions of race and ethnicity may be in question (Hunt
and Megyesi 2008a). Other countries, considering requirements of categorization of
study populations in a manner similar to those currently being used in the USA
might well reflect on the effects this has had in the USA. Health research must be
based on appropriate understandings and categorizations of human diversity, and be
Downloaded by [University of South Florida] at 19:56 18 August 2012

clear on differences between the genetic and the social. The results of this study not
only support the critical need for changes in categorization of subjects in health
research in the USA, but will also contribute to the development of appropriate
educational programs for health researchers on these topics.

Key messages
(1) Respondents were confused about the concepts of race and ethnicity and
their link to genetic differences between populations; many treated these
concepts as interchangeable and genetically based. Although ethnicity was
considered somewhat more socially constructed, it was often felt to cause
unhealthy behavior.
(2) More and better training in these concepts is necessary in formal and
continuing educational training of health researchers. Such training should
build on the actual perspectives of this population, beginning with discus-
sions of human diversity, and then move into what race and ethnicity are !
and are not.

Acknowledgements
The authors appreciate the editorial assistance of Jeannine Coreil and Susan McCombie.

References
Afshari, R. and Bhopal, R., 2010. Ethnicity has overtaken race in medical science:
MEDLINE-based comparison of trends in the USA and the rest of the world, 1965!
2005. International Journal of Epidemiology, 39 (6), 1682!1683.
Anderson, M. and Moscou, S., 1998. Race and ethnicity in research on infant mortality.
Family Medicine, 30 (3), 224!227.
Baer, R.D., 2008. Approaches to changing perspectives about the role of race in health. Paper
presented at the Society for Applied Anthropology, 68th annual meeting, 25!29 March 2008.
Memphis, TN.
Bliss, C., 2011. Racial taxonomy in genomics. Social Science and Medicine, 73 (7), 1019!1027.
Bradby, H., 2003. Describing ethnicity in health research. Ethnicity and Health, 8 (1), 5!13.
Comstock, R.D., Castillo, E.M., and Lindsay, S.P., 2004. Four-year review of the use of race and
ethnicity in epidemiologic and public health research. American Journal of Epidemiology,
159, 611!619.
 Ethnicity & Health 15

Frank, R., 2007. What to make of it? The (re)emergence of a biological conceptualization of
race in health disparities research. Social Science and Medicine, 64 (10), 1977!1983.
Fullerton, S., et al., 2010. Population description and its role in the interpretation of genetic
association. Human Genetics, 127 (5), 563!572.
Goodman, A., 2006. Seeing culture in biology. In: G. Ellison and A. Goodman, eds. The Nature
of Difference: Science, Society and Human Biology. London: CRC Press, 225!242.
Gravlee, C., 2009. How race becomes biology: embodiment of social inequality. American
Journal of Physical Anthropology, 139 (1), 47!57.
Gravlee, C. and Sweet, E., 2008. Race, ethnicity and racism in medical anthropology, 1977!
2002. Medical Anthropology Quarterly, 22 (1), 27!51.
Gregg, J. and Saha, S., 2006. Losing culture on the way to competence. Academic Medicine, 81
(6), 542!547.
Hunt, L. and Megyesi, M., 2008a. Genes, race and research ethics. Journal of Medical Ethics,
34, 495!500.
Hunt, L. and Megyesi, M., 2008b. The ambiguous meanings of the racial/ethnic categories
Downloaded by [University of South Florida] at 19:56 18 August 2012

routinely used in human genetics research. Social Science and Medicine, 66, 349!361.
Kaplan, J.B. and Bennett, T., 2003. Use of race and ethnicity in biomedical publication.
Journal of the American Medical Association, 289 (20), 2709!2716.
Kuzawa, C. and Sweet, E., 2009. Epigenetics and the embodiment of race: developmental
origins of U.S. racial disparities in cardiovascular health. American Journal of Human
Biology, 21 (1), 2!15.
Lee, S.C., 2006. Rethinking race and ethnicity in health disparities. Anthropology News, 47 (3),
7!8.
Lieberman, L., et al., 1992. Race in biology and anthropology: a study of college texts and
professors. Journal of Research in Science Teaching, 29 (3), 301!321.
Marks, J., 2006. The scientific and cultural meaning of the odious ape-human comparison. In:
G. Ellison and A. Goodman, eds. The Nature of Difference: Science, Society and Human
Biology. London: CRC Press, 35!51.
Megyesi, M. and Hunt, L., 2011. A critical review of racial/ethnic variables in osteoporosis
and bone density research. Osteoporosis International, 22, 1669!1679.
Morning, A., 2006. On distinction. Social Science Research Council [online]. Available from:
http://raceandgenomics.ssrc.org/Morning [Accessed 25 May 2012].
Office of Extramural Research, National Institutes of Health (NIH). 2001. NIH policy and
guidelines on the inclusion of women and minorities as subjects in clinical research*amended,
October, 2001 [online]. Available from: http://grants.nih.gov/grants/funding/women_min/
guidelines_amended_10_2001.htm [Accessed 15 February 2009].
Oppenheimer, G.M., 2001. Paradigm lost: race, ethnicity, and the search for a new population
taxonomy. American Journal of Public Health, 91 (7), 1049!55.
Outram, S.M. and Ellison, G.T.H., 2006. Improving the use of race and ethnicity in genetic
research: a survey of instructions to authors in genetics journals. Science Editor, 29 (3),
78!81.
Pearce, N., et al., 2004. Genetics, race, ethnicity and health. British Medical Journal, 328,
1070!1072.
Rachul, C., et al., 2011. Tracking the use and source of racial terminology in representation of
genetic research. Genetics in Medicine, 13 (4), 314!319.
Smedley, A. and Smedley, B., 2005. Race as biology is fiction: racism as a social problem is
real. American Psychologist, 60 (1), 16!26.
Stevens, J., 2003. Racial meanings and scientific methods. Journal of Health Politics, Policy and
Law, 28, 1033!1085.
US Census Bureau 2012. About race [online]. Available from: http://www.census.gov/
population/race/about [Accessed 7 February 2012].