Skip to main content
NCBI home page
UKPMC Funders Author Manuscripts logoLink to UKPMC Funders Author Manuscripts
. Author manuscript; available in PMC: 2013 Apr 22.
Published in final edited form as: Soc Sci Med. 2012 Jun 16;75(8):1400–1407. doi: 10.1016/j.socscimed.2012.05.020

Investigating the potential for ethnic group harm in collaborative genomics research in Africa: Is ethnic stigmatisation likely?

Jantina de Vries a,g,*, Muminatou Jallow b, Thomas N Williams c, Dominic Kwiatkowski d,e, Michael Parker a, Raymond Fitzpatrick f
PMCID: PMC3632260  EMSID: EMS52843  PMID: 22749442

Abstract

A common assumption in genomics research is that the use of ethnic categories has the potential to lead to ethnic stigmatisation – particularly when the research is done on minority populations. Yet few empirical studies have sought to investigate the relation between genomics and stigma, and fewer still with a focus on Africa. In this paper, we investigate the potential for genomics research to lead to harms to ethnic groups. We carried out 49 semi-structured, open-ended interviews with stakeholders in a current medical genomics research project in Africa, MalariaGEN. Interviews were conducted with MalariaGEN researchers, fieldworkers, members of three ethics committees who reviewed MalariaGEN project proposals, and with members of the two funding bodies providing support to the MalariaGEN project. Interviews were conducted in Kenya, The Gambia and the UK between June 2008 and October 2009. They covered a range of aspects relating to the use of ethnicity in the genomics project, including views on adverse effects of the inclusion of ethnicity in such research. Drawing on the empirical data, we argue that the risk of harm to ethnic groups is likely to be more acute in specific types of genomics research. We develop a typology of research questions and projects that carry a greater risk of harm to the populations included in genomics research. We conclude that the potential of generating harm to ethnic groups in genomics research is present if research includes populations that are already stigmatised or discriminated against, or where the research investigates questions with particular normative implications. We identify a clear need for genomics researchers to take account of the social context of the work they are proposing to do, including understanding the local realities and relations between ethnic groups, and whether diseases are already stigmatised.

Keywords: Africa, Kenya, The Gambia, Ethnicity, Stigma, Genomics, MalariaGEN

Introduction

With the rise of genomic studies, literature on the ethical implications of such work has proliferated (Caulfield et al., 2008; Kaye, Boddington, de Vries, Hawkins, & Melham, 2010). Few of these publications are, however, directly relevant to the African context (De Vries et al., 2011). Major concerns have been raised about the possible effects of genomics research in causing harm to the population groups included in the study, with the most specific form of harm being stigmatisation of ethnic groups (Foster & Sharp, 2006; Goodman, 1996; WHO, 2002). But although stigmatisation is frequently said to be a potential harmful consequence of genomics research, hardly any scholarly work has investigated the nature of this challenge. As a result, little is known about the processes of stigmatisation which may result from genomic research, about how researchers and other stakeholders in research conceive them, and about what may be done to ameliorate or avoid such consequences. This paper aims to explore the nature of potential harm for population groups included in a genome-wide association (GWA) study in Africa, when such harm might occur, and which measures could reduce the risk.

Genome-wide association studies aim to systematically compare genetic data derived from individuals with and without a particular disease, in order to identify genetic variants that are associated with the disease under study (for a good description see Fujimura & Rajagopalan, 2011). In this paper, we describe the results of a qualitative research study that was conducted within a GWA study on malaria, MalariaGEN (Jallow et al., 2009).

The majority of genomic studies conducted to date have been on populations of European descent (Need & Goldstein, 2009). Only a fraction has included populations from Africa (Rosenberg et al., 2010). The exclusion of African populations from genomic studies may promote or prolong existing global health inequalities, in particular if such research leads to knowledge that is of clinical relevance (Coloma & Harris, 2009; Newport & Rotimi, 2009). It is therefore very important that genomics research methods are also used for the investigation of diseases primarily affecting patients on the African continent.

Such studies are not without methodological difficulties however. African people are genetically very diverse (Rosenberg et al., 2010), and this diversity causes considerable methodological challenges in expanding GWA studies to the African continent (Teo, Small, & Kwiatkowski, 2010). Population substructure – i.e. when members of a population have a shared genetic background – has the potential to confound GWA analyses and lead to false positive associations between disease and clinical outcome. One approach has been to match samples according to population groups. In African settings, this generally means matching samples by ethnic group.

The effect of including ethnicity in the GWA studies, however, is that two of the main outcomes of such research – research results and genomic datasets – are thereby correlated with ethnicity. For instance, although the MalariaGEN project did not set out to study ethnic groups, researchers considered the collection of information on participants' ethnicity to be essential for analysis. As a result, however, the genomic datasets produced were specific for ethnic groups. A rich literature exists around social and ethical issues raised by the use of racial and ethnic categories in genomics research. Important concerns relate to the measurement of these categories (Hunt & Megyesi, 2008; Lee, 2009), appropriate ways of reporting on group specific research findings (Martin, Ashcroft, Ellison, Smart, & Tutton, 2007; Wensley & King, 2008), and a possible revival of eugenic ideologies (Aultman, 2006). Scholars have also criticised genomic and genetic models of disease that overlook the ways in which diseases, as well as their perceived severity and social impact, are culturally and institutionally embedded within countries and population groups (Fullwiley, 2010). A detailed overview of these issues is not the purpose of this paper. Excellent reviews of this literature can be found in Reardon (2005), Frank (2007) and Koenig et al. (2008).

This literature, however, tends to be silent about the potential for GWA studies to cause stigmatisation for ethnic groups included in studies. When considering the relation between genomics and stigma, authors seem to locate the potential for stigma in the possibility to attribute genetic risk factors for particular diseases to identifiable population groups (Foster & Sharp, 2006; Goodman, 1996; WHO, 2002). Such an understanding seems to be concordant with views of stigma as a ‘mark’ – an understanding of stigma that has been amply criticised as being overly simplistic (Greely, 2010; Parker & Aggleton, 2003; Phelan, 2005). Moreover, genomic studies tend to focus on complex diseases, for which it is unlikely that any one genetic variant will be identified that could explain disease causation.

A small number of studies have investigated the relation between stigma and research on single gene conditions such as some forms of deafness. The majority of such studies have been conducted in the US (Phelan, 2005; Sankar, Cho, Root Wolpe, & Schairer, 2006), although some studies have been conducted outside of the US (Meiser, Mitchell, McGirr, Van Herten, & Schofield, 2005). With the exception of one study on consent for a genomic study on podoconiosis (Tekola et al., 2009), no such studies have been conducted on the African continent. But with the expansion of genomics methodology to Africa, it is increasingly pertinent to understand the potential of genomics research in Africa to impact on stigma.

Methods

MalariaGEN

To investigate the nature of potential harm for ethnic groups arising out of GWA research in Africa, we conducted qualitative interviews with 49 stakeholders involved with a current GWA study in Africa. MalariaGEN was a partnership between malaria researchers in over 20 countries across the globe (The MalariaGEN Consortium, 2008). The network was composed of research groups that each conduct their own research on topics relating to malaria, and all participated in some component of the GWA study. For the purposes of this study, we were interested in each group's role in and perspectives on the use of ethnicity in the GWA study.

On the African continent, samples were collected by 13 research groups from 11 countries. With the exception of only a small subset of samples, all samples were collected with ethnic identifiers, which were used to account for population stratification. Data on the ethnicity of MalariaGEN research participants was mostly self-defined, and sometimes triangulated with information about language and the ethnicity of parents or grandparents. Samples were donated by participants from at least 30 different ethnic groups from across Africa. MalariaGEN participant recruitment did not initially target members of particular groups but included all malaria patients presenting at selected healthcare facilities. However, the research design did subsequently require that samples from healthy volunteers – the controls in the experiment – were matched in terms of age, gender and ethnicity. The recruitment of samples from healthy volunteers was therefore ethnic-group specific. Although samples were collected from only a small number of research participants per group, in presentations and publications the results were treated as if they were relevant for the entire ethnic group – an interesting and important issue that we intend to discuss in a future paper. The MalariaGEN consent process did not generally include a description of the need to collect information about participants' ethnicity, or any possible harm arising out of the use of ethnicity in the research project. In some cases, however, this was discussed during the community consultation process. For an example of a community consultation process around a MalariaGEN project at one particular research site, see Tindana et al. (2007).1

Selection of interview participants

Interviews were conducted at two MalariaGEN project sites in Africa and in the UK, between June 2008 and October 2009. The two sites in Africa – the KEMRI-Wellcome Overseas Programme in Kilifi, Kenya and the MRC Labs in Fajara, The Gambia – had experience with genomics research prior to MalariaGEN, and had also hosted other social science and ethics research projects (Fairhead, Leach, & Small, 2006; Marsh, Kamuya, Rowa, Gikonyo, & Molyneux, 2008). Both these research programmes receive primary funding from institutions in the UK. These two sites were chosen because they were amongst the few in the MalariaGEN study that had prior experience with genomics research, and because they were expected to contribute the largest number of samples to the MalariaGEN study. The two MalariaGEN research groups in Kenya and in The Gambia both conduct a broad range of malaria research but neither of these groups is primarily composed of genomic researchers.

The main objective of the interviews was to identify ethical issues in the use of ethnic data in genomics research in Africa. Four categories of stakeholders were identified that were considered important for this project. These were: MalariaGEN researchers (20 interviews); MalariaGEN fieldworkers (15 interviews); members of 3 research ethics committees that reviewed MalariaGEN project proposals (12 interviews); and members of the funding bodies that supported the MalariaGEN project (2 interviews). Fieldworkers were included as a semi-lay group of interviewees. They are the members of the research team that are normally tasked with study participant recruitment. All fieldworkers were originally from the communities and populations from which MalariaGEN research participants were recruited. In Kenya, most fieldworkers were trained up to and including secondary school level. In Gambia, the fieldworkers were trained nurses.

People donating samples to the MalariaGEN project – the original research participants – were not included in this study for a number of reasons (see Limitations below). Limited guidance in literature about what the issues might be complicated the development of research tools appropriate for conducting research with the original MalariaGEN participants. In addition, we considered it necessary that interviewees had some prior knowledge of concepts like ‘genomics’, ‘ethics’ and the reasons for collecting ethnic data.

All but one of the interviews were conducted in person by the same interviewer. One interview was conducted over the phone. Detailed fieldnotes were recorded throughout the project.

In the discussion of our results, interview quotes are identified by the following acronyms: R indicates MalariaGEN researchers, FW indicates MalariaGEN Fieldworkers and REC indicates a member of an ethics committee that reviewed the original MalariaGEN proposal.

Interview guides

The interview guides were developed in consultation with colleagues and with two informants at the research sites in Kenya and The Gambia. Although the themes covered in the interviews were similar across all stakeholder categories, the interview guides were adapted to suit the experience of the participants in the four categories. For instance, questions about the actual measurement of ethnicity were suitable for researchers and fieldworkers, but were not appropriate for ethics committee members. The interviews covered issues in the current practice of using, defining and measuring ethnicity; awareness of particular ethical issues in using ethnicity for genomics research; issues in identifying ethnic groups and genomic data in research and publications; implications of labelling ethnic groups; issues in the sharing and re-use of ethnic data in genomics; and possible solutions to the challenges identified. Although interviews gave extensive opportunity for adverse effects of genomics to be raised or discussed, the terms ‘stigma’ or ‘stigmatisation’ were not introduced by the interviewer. Where the interviewee used the term, interviews explored the meaning of the term and its relation to genomics research.

Data analysis

The processes of data collection and analysis were conducted iteratively during this study; transcripts of interviews were analysed throughout the project in tandem with data collection and new insights were integrated into subsequent interviews. This process continued until no new issues, themes or insights were generated during the interviews or coding. At this point, the study reached so-called ‘saturation’ (Mason, 2010), and interviewing was stopped subsequently.

Interviews were recorded and transcribed verbatim. Data was analysed inductively using specialized software (QSR, 2009). The first stage of open coding was followed by hierarchical coding where emerging patterns and themes in the data were established (Quinn-Patton, 2002). Interpretations of the data were discussed amongst the research team. Early insights were explored critically in subsequent rounds of coding and analysis to explore their authenticity and appropriateness. The use of detailed fieldnotes was essential in this process, to trace the development of insights and understandings, and to offer a means for critical reflection.

Limitations of the study

The aim of this project was to develop a better understanding of the ethical issues raised by the use of ethnic data in a particular genomic project in Africa, and not to investigate these issues in the specific research contexts of two research centres in Africa. For this reason, results are not reported for the two research sites specifically. No major systematic differences were found between the views of the various stakeholders. Where we found some divergence of opinion, we have indicated this in the text.

A second important limitation of this study is that MalariaGEN research participants were not included in this study. Our work therefore does not elucidate the relation between genomics research and stigma as a lived experience for research participants and their communities (Yang et al., 2007). It will be essential to draw on the findings reported in this paper in the design of a research project aimed at examining research participants' views on and experiences of stigma in genomics research in Africa.

Lastly, our study examined the perspectives of a variety of stakeholders in genomics research in Africa on the potential for genomics research to contribute to ethnic stigmatisation. Although we have generated a better understanding of whether, when and how ethnic group stigma could arise in relation to genomics research in Africa, this relation remains largely hypothetical. What will be required next is to monitor in practice whether stigma or harm actually emerge in relation to genomics research on the African continent, now and in the future, and how it affects the lives of those participating in research, or affected with particular conditions.

Ethics approval and consent

This study was reviewed and approved by the Oxford Tropical Research Ethics Committee in the UK (OX 22-08), the KEMRI/National Ethical Review Committee (SCC4547) and The Gambia Government/MRC Laboratories Joint Ethics Committee (SCC1137v2). All interviewees gave informed consent prior to the interview. Consent was given for participation in the study, for recording of the interview, and for use of anonymised quotes in research materials.

Results

Benefits of using ethnicity in genomics research in Africa

Interviewees described a number of advantages of using ethnicity in genomics research. Although most interviewees (34 out of 48) identified benefits to the use of ethnic categories in the MalariaGEN project, a small number of interviewees (9 total; 1 fieldworker, 4 each of ethics committee members and researchers) considered the focus on genetic diversity between ethnic groups in genomics studies as particularly positive. One reason is that the inclusion of ethnicity in genomics research enables the use of this methodology for resource-restrained sites and projects. Namely, although alternatives exist for the use of population categories (described for instance by Fujimura & Rajagopalan, 2011), such alternatives are more costly as they involve the screening of larger numbers of genetic variants.

“we are only genotyping the top 200 SNPs in new samples and of course in those new samples we don't have principal components […] I mean these big studies are really expensive and […] even though from my point of view since most of the analysis I do is on the big datasets I don't technically need [ethnicity], all [smaller MalariaGEN] sites to do just the analysis on candidate genes and things like that of they will…”(R).

When hundreds of thousands of genetic variants – also called single nucleotide polymorphisms or SNPs – are typed per individual, then one can use a statistical tool called ‘principal component analysis’ to stratify project data into genetic ancestry clusters. This obviates the need for using ethnic groups. But when the resources only allow for the typing of a far smaller number of variants in different populations then the use of population categories is considered essential to account for population stratification. In a severely resource-restrained setting, which is the case for many research centres in Africa, interviewees considered this to be an important benefit to using ethnic categories.

Interviewees also described an additional motivation for using ethnic categories in genomics research in Africa. Namely, they considered that neglecting population differences in research is a harmful response to historical abuses in the use of population categories in research.

“Because of fascism and the Nazis, the experiments that were done on certain ethnic groups, people were very reluctant to talk about human difference you know. I think that we have enjoyed a good 50 years of pretending that human differences do not exist. So this is an exaggerated political correctness that can backfire”(R).

Some interviewees also expressed the hope that genomics would offer a more value-neutral catalogue of human history that is based on genetic evidence. For these interviewees, such a catalogue may be less biased than other ‘catalogues of variation’, such as ones based on skin colour.

“So rather than having something relatively unimportant like skin colour […], you could actually have something that gives you a bit more of a catalogue of variation across the genome to show you relationships between groups. [.It's] rather, having a less biased catalogue of variation for ethnic groups”(R).

The hope that genomics research would offer a relatively value-neutral catalogue of variation was shared not only by some researchers, but also by one member of an ethics committee based in an African country. This perspective seems to rest on the belief that little social or normative value can and will be attributed to genomic differences between populations or individuals. Processes of stigmatisation normally rely on the identification of difference that can be relatively easily perceived (Link & Phelan, 2001), and where value is attached to the difference. The hope that genomics research could offer a catalogue of apparently insignificant differences that have not historically been stigmatised or considered of normative value, was for some interviewees a reason to expand these methods to all diseases and populations.

Potential for group harm in genomics research

At the same time, all research participants identified some form of potential group harm relating to the use of ethnic categories in genomics research. The most frequently defined challenge was the possibility that genomics research could lead to the clear linking of (susceptibility to) diseases to particular ethnic groups. The fear is that such information could be used as evidence of difference that is rumoured to exist, to fuel animosity, or to affirm the social and cultural superiority of one ethnic group over another, for instance when one group is found to have some genetic resistance against diseases.

One important concern that the interviewees identified was that the genomic focus on ethnic difference could be seen to affirm the social and biological importance of these categories.

“one of the things I can see as potentially a problem is that saying that groups are different genetically or saying that this group is more similar to this group than this group perhaps legitimises antagonism that already exists. Even if we try to stress that by far and away most variation is shared by everyone, if we put out our plots and say something like ‘this group is more related to this group or this group is distinct from these other groups’ that somewhat legitimises the fact that one group thinks of themselves as different”(R).

For fieldworkers, the potential harm is in ‘shaming people’, which is a risk especially where genomics research describes differences between ethnic groups and those differences have normative implications. One possible harmful implication of using ethnic groups in genomics research is that it could influence opportunities for marriage for affected individuals or their family members.

“I don't think it would be [good] if we bring in the aspect of the tribe […] when people know that a particular tribe they have this disease, […] if it comes to marriage people will be shying away from that particular [tribe] saying that okay they have this disease and we know very well it's not curable so it will be like a disadvantage”(FW).

Overall, the nature of the harm that is described by interviewees seems to relate to damage to the reputation, integrity or social status of the population groups involved in the studies. Such damage can be harmful in itself, but can also subsequently give rise to discriminatory practices.

Many interviewees introduced the term ‘stigma’ to describe such harm. Nearly all researchers (15 out of 19), ethics committee members (10 out of 12) and funding body representatives (2 out of 2) used this term. Only a small number of fieldworkers (5 out of 14) also used this term to describe possible harm of using ethnicity in genomics research. Ethics committee members for instance, described that concerns over stigma were discussed in their review of genomics research proposals.

“Our concern has always been relating findings to tribes or to ethnic groups because we felt that that could lead to people being stigmatised or at the risk of being stigmatized”(REC).

When prompted to describe what they meant by the term, interviewees defined stigma mostly in terms of its consequences, and then predominantly in terms of shame and social isolation. Stigma was seen as a cause of low self-esteem, social rejection, ridicule and solitude.

Interviewees tended to describe stigma as a trait that can be mapped onto people.

“I can imagine that if there was a genetic trait that would be understood by the people as being detrimental […] I could understand that people could be stigmatised if that population was associated with that particular trait”(R).

The fear is that genomics research could come up with results that “label” (REC) a particular group, that give “certain attributes” (REC) to groups, and that allow others to “say evil things” (FW) about members of the ethnic group. These ideas about how genomics research could lead to stigma seem to be quite different from the notion that genomics research could offer a value-neutral ‘catalogue of variation’ that we described before. For these interviewees, the risk of harm arises in the translation of these findings into socially meaningful terms and concepts. Whereas on the one hand, these interviewees consider genomic information to be innocuous – for instance, when it concerns statistical information about group prevalence of a particular genetic variant that reduces disease susceptibility by a few percent – they consider that the potential for harm arises when such information is translated into socially meaningful terms – for instance, when genomic study results are seen as evidence that a particular group is genetically inferior to another.

Examples of genomics research that could lead to group harm

When describing possible harm for ethnic groups, interviewees give important clues about the types of research that could have such consequences. Genomic research results relating to malaria were not expected to be particularly sensitive.

“if we do research concerning AIDS and we give the findings along the tribal lines, it is like you are watching those tribes. but if you are doing research on malaria and you mention that these people are susceptible to malaria there is no shame there” (FW).

Where diseases are already stigmatised research results are more likely to harm the participating population groups.

“it's more the association of known stigmatising diseases. For example if […] all of the sickle [cell disease variants] were in one group and not in another group that would be […] something which could get sort of publicised and blown up and mishandled and communities might get feedback that all the bad genes are in one ethnic group”(R).

In our sample, what interviewees considered important was whether the disease under investigation was already stigmatised. Examples of already stigmatised conditions mentioned by the interviewees were HIV/AIDS and epilepsy. Interviewees mentioned examples of diseases or conditions that relate to behaviour as being more likely to lead to harm for the population groups involved, such as for instance research on alcoholism or intelligence.

“people go looking through the genome for positive selection […]. So there certainly have been people in the past […] trying to claim that there are brain-specific alleles that have swept up and have been positively selected that [one finds] for example only in Europeans”(R).

The concern is that genomics data could be used to make normative statements about people of different ethnicities, or different races. A notorious example mentioned by interviewees is that of a pioneer geneticist and co-discoverer of the structure of DNA, who expressed a view that genetic research would lead to the discovery of genetic factors explaining differences in ‘the power to reason’ between Africans and Westerners (Milmo, 2007).

Another set of research findings that may be more sensitive are findings that have the potential to be claimed as evidence of what is considered immoral behaviour. A particular example is data on extramarital paternity frequencies between ethnic groups, where these frequencies differ.

“for instance if […] you write an article and say in this particular ethnic group we've found that 20% of people had a different father […] and most other ethnic groups only have 5% then you start to get into trouble”(REC).

In some cases, samples were collected from so-called family trios, comprising a child that was infected with severe malaria, and its biological parents. In a number of instances, the reported father was found not to be the genetic father. The frequency with which this occurs tends to differ between population groups. Where this is thought to be a proxy for sexual promiscuity, then interviewees considered this a possible example of harm to the reputation or integrity of ethnic groups.

A final challenge that interviewees identified arises where genomics research results are in conflict with traditional narratives of descent. One example is where a population group is composed of people that were originally enslaved by an ethnic group. These people now consider themselves fully part of that ethnic group, whereas their former capturers still consider them as different. Although the collection of information remains important to account for population stratification, if the results of such studies were published they could upset the precarious relations between these ethnic groups.

“I don't think that to disclose the information in an open way would be a good idea. We can call them something else. They're like three ethnic groups and they're one, two, three and not [ethnic group A, ethnic group B]”(R).

Other interviewees also considered that harm could arise where biological (genetic) accounts of ancestry were in conflict with social narratives.

“But I think that the issue of ancestry and common knowledge or common beliefs of descent can be sacred, it's something we really need to handle with care”(R)

There is some evidence to suggest that a conflict between biological and social or traditional narratives of descent could be harmful. For instance, in the recent case of the Havasupai in North America, biological evidence was perceived – by tribal members or possibly by commentators on the case – to be in conflict with the tribe's traditional narrative of origin (Harmon, 2010). At least one author (McGregor, 2010) argued that the group had been harmed as a result. By contrast, the Lemba – a black Southern African ethnic group – saw their claims to an ancient Jewish ancestry confirmed when genetic studies found similarity between the Y chromosome of some Lemba males, and the Y chromosome carried by contemporary members of the Jewish priesthood in Israel (Parfitt & Egorova, 2005). This seems to suggest that genomics research has a potential to authenticate, challenge or destabilize communal narratives of origin, and that this may have beneficial or harmful consequences to the groups involved.

Harm and benefit of repeated inclusion of specific ethnic groups in genomics research

Interviewees described both challenges and benefits relating to the repeated inclusion of some population groups in genomics research. In some cases, there may be public health value in studying particular ethnic groups, for instance, when groups are genetically predisposed to certain diseases, or when a group's genomic material indicates a difficulty in metabolising particular drugs.

“Say for example there is a particular allele of a gene that is found in a sub-group and we know that that allele predisposes the population to some level of severity of the disease […] then you know if you have this [disease] in this group then you need to watch out because you know that they harbour a particular gene which will predispose them to a more severe form of the disease […] it would be useful from a medical point of view”(R).

At the same time, interviewees – and especially ethics committee members – felt that there are also good reasons why ethnicity should not become part of the rationale for participant selection.

“you see, we tie it in as a matter of justice. Because you don't want to imagine you're only studying the Kikuyu because they are Kikuyu or you're only studying the Luhya because they're Luhya. You should be able to study all”(REC).

More importantly, the fear is that particular ethnic groups would become the target for genomics research. Where there is an expectation that genomics research could have benefits for the ethnic group involved – for instance, they become the focus of further medical research which may improve treatment for that group both short and long term – repeated exclusion of particular groups from research may disadvantage those groups. Where only majority groups are included in research, and minority groups excluded, the fear is that this could reinforce social inequality between those groups.

“I can imagine it being a potential problem if there was either a big focus on a particular group and they might be seen as either getting favouritism by getting input from the foreign doctors coming and doing stuff, […] or it might be this group has been singled out because they've got some dreadful disease, susceptibility and therefore they're weaker or there's some other problem that makes them different in a bad way”(REC).

There is already some anecdotal evidence that some ethnic groups in Africa are repeatedly included in medical genomics research in Africa. For instance, to date the MalariaGEN project focussed sequencing efforts on ethnic groups from The Gambia (Jallow et al., 2009), Ghana and Malawi. The same populations from The Gambia were also included in the first GWA study to be reported in literature, the Wellcome Trust Case Control Consortium (2007). Connections between genomics researchers in MalariaGEN and other projects has also meant that ethnic groups from The Gambia, Ghana and Malawi were included in another large genomics research initiative, the 1000 Genomes project (http://www.1000genomes.org). It will be important to understand what the ethical implications of the repeated inclusion of certain population groups in genomics research will be, and whether such inclusion is likely to generate any harm for those population groups.

Genomics research as a cause of stigma?

Interviewees for this project believed that the risk of genomics research causing harm to population groups could be elevated where research is carried out on already stigmatised conditions, on marginalized population groups, or when research results relate to what can be considered immoral behaviour. In most other circumstances, interviewees seemed to consider that the risk of ethnic stigmatisation is not acute in genomics research, and arises largely as a side-effect to the original study.

“I think that exacerbation [of ethnic stigma] is a more probable problem than causing [it] because if there isn't ethnic tension then the results of scientific studies are not going to be hitting a fertile ground in terms of causing tension, I suppose they'd say ‘oh that's odd, a proportion of us are different to the others’. […] The main risk is where […] there is already competition between tribes for power, for position, for land, for influence. I think when that happens research results can be used in a way that the researchers don't anticipate”(REC).

What this suggests is that genomics research should be considered not so much as a cause of stigma than as a factor that could feed into stigma where it is already present. This does not mean, however, that the use of ethnic categories in genomics research cannot cause some other form of harm.

Means of minimizing the potential for harm

Interviewees argued that a very important component of understanding the potential harm of using ethnic groups in genomics research relates to understanding the way in which research results are reported or interpreted.

“if you say one particular group of people is more susceptible to diabetes […] I wouldn't regard that as stigmatisation, I'd regard it as a fact of life. If on the other hand you said these people have got, to start getting emotive, they've got rotten genes of one sort or another and they over-eat because they're lazy and they're alcoholics then you start to get into a description of stigmatisation”(REC).

The issue is that genomics results can be interpreted in ways that are positive or negative. The onus seems to be on genomics researchers to commit to presenting research results in ways that emphasize positive and not negative implications for population groups.

“It's two-sided you know because [genomics] can show that people have different origins, but it can also show that people have common origins. So [these two tribes] are of the same origin but they are fighting for many years. You can show them that they have really no major differences, that they have a recent common origin. But if you tell [them] that they are different […] it might have a blow back”(R).

A related challenge is that genomics research identifies statistical differences between categories of people, but these statistical differences may not translate into meaningful differences for medical practice. How to engage with communities in a way that aptly explains the nature of this difference is a significant challenge for genomics researchers working in Africa.

“We are talking here about statistical differences. With big numbers even a small difference can be found in statistics. How we can translate this in ethnic differences that's the big issue. […] We really need to think deeper to see how we can present such statistical data to a community who don't know anything about statistics”(R).

One solution, interviewees suggested, would be to specify in papers how ethnicity was measured, and how research results were published.

“if you're going to write about it you have to make sure you've qualified it you know. You can only talk about ethnicity of that particular group in that particular area”(R).

This is in line with recommendations made by numerous authors (see for instance Martin et al., 2007; Shanawani, Dame, Schwartz, & Cook-Deegan, 2006). However, for reasons that merit further investigation, biomedical scientists have resisted calls for greater care in reporting on race and ethnicity, even when coming from editors of high-impact scientific journals (Rose, 2006). Whether the possibility of harm to ethnic groups will provide more compelling reasons to qualify the meaning of ethnic group specific research results, remains to be seen.

Discussion

Our interviews demonstrate that the relation between genomics research and stigma is not straightforward, and that claims about ethnic stigmatisation as one of the outcomes of genomics research need further substantiation. In line with suggestions made in literature, our interviewees also tended to describe as problematic the potential for genomics research to highlight differences in group susceptibility to disease – in other words, that genomics research can ‘mark’ particular populations as being more susceptible to disease, and that this would be sufficient to stigmatise those populations.

But the focus on stigma as a ‘mark’, ‘attribute’ or ‘characteristic’ has been challenged as being reductive and as treating stigma as something that is ‘mapped onto people, who in turn by virtue of their difference, are understood to be negatively valued in society” (Parker & Aggleton, 2003, p. 14). Link and Phelan, amongst others, have worked to define stigma as a social process, comprising the identification of targets of stigma, the attribution of normative beliefs to the targets, and the subsequent labelling of people (Link & Phelan, 2001). As a social process, stigmatisation builds upon existing social and power relations and dominant cultural beliefs about similarity and difference. Simply ascribing difference to a population group – through genomic research or other means – would in this understanding of stigma not be sufficient to cause stigmatisation of that group. Drawing on our findings, we would like to suggest that it seems unlikely that genomics research could be the cause of stigma, but rather that it could be a factor contributing to the process of stigmatisation. Stigma also entails a process of embodiment – where possible targets of stigma can internalise or resist stigma (Parker & Aggleton, 2003; Yang et al., 2007). It remains unclear how genomic research results that pertain to population groups, will be internalised and resisted by individual members of those groups, and how such results will relate to traditional understandings of disease and inheritance (Fullwiley, 2010). Tekola et al. (2009) found that genomics research findings were integrated with traditional understandings of family inheritance and blood relations. Many communities have been found to have notions of inheritance which could encapsulate novel genomic knowledge and findings (Lock & Nguyen, 2010; Simpson, 2000). In the MalariaGEN experience, such notions were sometimes used to explain the genomics research project to research participants, as we have described elsewhere (De Vries et al., 2011). Future research should seek to further investigate the relation between traditional understandings of biological inheritance, genomics and stigma.

As an alternative, the concept of ‘harm to ethnic groups’ can take into account other relevant outcomes that are disadvantageous for ethnic groups but that do not constitute ‘stigma’ in the way that it is currently understood in literature. Such harm may include for instance negative publicity for an ethnic group because of particular genomics research findings or negative consequences of being excluded from research based on ethnic affiliation.

We have outlined a number of circumstances in which the likelihood of harm arising as a consequence of genomics research, is increased. We believe harm is more likely when the ethnic groups included in research are already stigmatised or discriminated, for instance in the case of minority groups in a population. We also believe that we can identify three categories of genomic research projects or findings that could signal greater potential for group harm. These are:

  • Disease-related comparisons between ethnic groups when investigating diseases or conditions with normative implications or stigmatised conditions. Examples of the first are diseases or conditions that relate to behaviour, for instance sexual promiscuity, alcoholism or intelligence. Already stigmatised conditions could be HIV/AIDS, epilepsy, or podoconiosis (Tekola et al., 2009);

  • Findings that are evidence of what is considered to be immoral behaviour. For instance, data on false paternity frequencies between ethnic groups. Another example is evidence of admixture between ethnic groups, where reproduction between members of two groups is taboo. Evidence of taboo family relationships such as incest would also fall into this category;

  • Projects looking for evidence of group origin or composition, where these are in conflict with social and traditional narratives. A recent example is the case of the Havasupai, who argued, amongst other things, that their inclusion in a genomic study caused them harm because it provided a genetic narrative of descent that was in conflict with their own beliefs. Where genomic data confirms traditional narratives, such as was the case for the Lemba of Southern Africa, such harmful effects do not seem to arise.

The potential for research projects and results to lead to group harm corresponds with existing relations between ethnic groups and with existing stigma. Where relations are not problematic, and where the disease under question is not stigmatised, there is limited ground to believe that research results can be used in ways that constitute harm for participants. But where relations are strained, or where diseases are stigmatized, there is a much greater cause for concern.

What our findings suggest is that there is a clear need for genomics researchers to take account of the local social context of the work that they are proposing to do. This may involve careful preliminary consultations with research participants and their communities to understand the local realities and relations between ethnic groups, as well as whether the diseases they are hoping to investigate are already stigmatised. Our research also raises a question about whether the potential for group harm arising out of genomics research should be discussed with communities and population groups prior to their enrolment in such research. In addition, many genomics research projects are funded internationally, and are led by researchers based in Europe or the United States (Newport & Rotimi, 2009). How the increasing geographical spacing of research across the globe influences researchers' ability to understand or recognise the potential for harm needs to be further explored. In addition, there is also a need to critically examine the implications of repeated inclusion of specific ethnic groups in subsequent genomics research projects.

In this research project, we did not include the perspectives of research participants in genomics research in Africa for methodological reasons. The implication is that our work does not address deeper questions about the relation between genomics research and stigma as a lived experience for research participants and their communities (Yang et al., 2007). It is pertinent that future research examines how the lives of the people participating in genomics research can be affected by stigma that arises out of or is aggravated by genomics research.

Acknowledgements

We thank Drs Vicki Marsh and Susan J. Bull and three anonymous reviewers for comments received on earlier versions of this manuscript. Our thanks to all research participants in this project, and to colleagues of the MalariaGEN Consortium, at the Ethox Centre and the Wellcome Trust Centre for Human Genetics.

JdV gratefully acknowledges the support of a Wellcome Trust Research Studentship for this research (WT083326). DPK receives support from the UK Medical Research Council (G19/9). The Wellcome Trust also provides core awards to Wellcome Trust Centre for Human Genetics (075491/Z/04) and the Wellcome Trust Sanger Institute (077012/Z/05/Z). TNW is funded by the Wellcome Trust (WT076934) and by the European Union Network 7 EVIMalR Consortium. MP is supported by a Wellcome Trust Biomedical Ethics Enhancement Award (WT087285) and Strategic Award (WT096527). The MalariaGEN Project was supported by the Wellcome Trust (WT077383/Z/05/Z) and the Bill and Melinda Gates Foundation through the Foundations of the National Institutes of Health (566) as part of the Grand Challenges in Global Health Initiative. The Medical Research Council (G0600230) provided additional support for genotyping, bioinformatics and analysis.

Footnotes

1

The research team for this study consisted of JdV, MP and RF. JdV and MP had previously been working with MalariaGEN as bioethicists. RF was not involved in the MalariaGEN project. The study received separate funding and was financially and academically independent of the MalariaGEN project.

References

  1. Aultman JM. Eugenomics: eugenics and ethics in the 21st century. Genomics, Society and Policy. 2006;2:28–49. [Google Scholar]
  2. Caulfield T, McGuire AL, Cho MK, Buchanan JA, Burgess MM, Danilczyk U, et al. Research ethics recommendations for whole-genome research: consensus statement. PloS Biology. 2008;6:e73. doi: 10.1371/journal.pbio.0060073. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Coloma J, Harris E. Molecular genomic approaches to infectious diseases in resource-limited settings. PLoS Med. 2009;6:e1000142. doi: 10.1371/journal.pmed.1000142. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. De Vries J, Bull S, Doumbo OK, Ibrahim M, Mercerau-Puijalon O, Kwiatkowski DP, et al. Ethical issues in human genomics research in developing countries. BMC Medical Ethics. 2011:12. doi: 10.1186/1472-6939-12-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Fairhead J, Leach M, Small M. Where techno-science meets poverty: medical research and the economy of blood in The Gambia, West Africa. Social Science & Medicine. 2006;63:1109–1120. doi: 10.1016/j.socscimed.2006.02.018. [DOI] [PubMed] [Google Scholar]
  6. Foster MW, Sharp RR. Ethical issues in medical-sequencing research: implications of genotype-phenotype studies for individuals and populations. Human Molecular Genetics. 2006;15:R45–R49. doi: 10.1093/hmg/ddl049. [DOI] [PubMed] [Google Scholar]
  7. Frank R. What to make of it? The (re)emergence of a biological conceptualization of race in health disparities research. Social Science & Medicine. 2007;64:1977–1983. doi: 10.1016/j.socscimed.2007.01.010. [DOI] [PubMed] [Google Scholar]
  8. Fujimura JH, Rajagopalan R. Different differences: the use of ‘genetic ancestry’ versus race in biomedical human genetic research. Social Studies of Science. 2011;41:5–30. doi: 10.1177/0306312710379170. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Fullwiley D. Revaluating genetic causation: biology, economy, and kinship in Dakar, Senegal. American Ethnologist. 2010;37:638–661. [Google Scholar]
  10. Goodman KW. Ethics, genomics, and information retrieval. Computers in Biology and Medicine. 1996;26:223–229. doi: 10.1016/0010-4825(95)00059-3. [DOI] [PubMed] [Google Scholar]
  11. Greely HT. To the Barricades! The American Journal of Bioethics. 2010;10:1–2. doi: 10.1080/15265161.2010.494229. [DOI] [PubMed] [Google Scholar]
  12. Harmon A. Tribe wins fight to limit research of its DNA. New York Times; New York: 2010. [Google Scholar]
  13. Hunt LM, Megyesi MS. The ambiguous meanings of the racial/ethnic categories routinely used in human genetics research. Social Science & Medicine. 2008;66:349–361. doi: 10.1016/j.socscimed.2007.08.034. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Jallow M, Teo YY, Small KS, Rockett KA, Deloukas P, Clark TG, et al. Genome-wide and fine-resolution association analysis of malaria in West Africa. Nature Genetics. 2009;41:657–665. doi: 10.1038/ng.388. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Kaye J, Boddington P, de Vries J, Hawkins N, Melham K. Ethical implications of the use of whole genome methods in medical research. European Journal of Human Genetics. 2010:18. doi: 10.1038/ejhg.2009.191. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Koenig BA, Soo-Jin Lee S, Richardson SS, editors. Revisiting race in a genomic age. Rutgers University Press; New Brunswick, New Jersey and London: 2008. [Google Scholar]
  17. Lee C. “Race” and “ethnicity” in biomedical research: how do scientists construct and explain differences in health? Social Science & Medicine. 2009;68:1183–1190. doi: 10.1016/j.socscimed.2008.12.036. [DOI] [PubMed] [Google Scholar]
  18. Link BG, Phelan JC. Conceptualizing stigma. Annual Review of Sociology. 2001;27:363–385. [Google Scholar]
  19. Lock M, Nguyen VK. An anthropology of biomedicine. Blackwell Publishers; Chicester: 2010. [Google Scholar]
  20. McGregor J. Racial, ethnic, and tribal classifications in biomedical research with biological and group harm. The American Journal of Bioethics. 2010;10:23–24. doi: 10.1080/15265161.2010.492888. [DOI] [PubMed] [Google Scholar]
  21. Marsh V, Kamuya D, Rowa Y, Gikonyo C, Molyneux S. Beginning community engagement at a busy biomedical research programme: experiences from the KEMRI CGMRC-Wellcome Trust research programme, Kilifi, Kenya. Social Science & Medicine. 2008 doi: 10.1016/j.socscimed.2008.02.007. http://dx.doi.org/10.1016/j.socscimed.2008.02.007. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Martin P, Ashcroft R, Ellison GTH, Smart A, Tutton R. Reviving ‘racial medicine’? The use of race/ethnicity in genetics and biomedical research, and the implications for science and healthcare. St. George's, University of London; London: 2007. [Google Scholar]
  23. Mason M. Sample size and saturation in PhD studies using qualitative interviews. Forum: Qualitative Social Research. 2010:11. [Google Scholar]
  24. Meiser B, Mitchell PB, McGirr H, Van Herten M, Schofield PR. Implications of genetic risk information in families with a high density of bipolar disorder: an exploratory study. Social Science & Medicine. 2005;60:109–118. doi: 10.1016/j.socscimed.2004.04.016. [DOI] [PubMed] [Google Scholar]
  25. Milmo C. Fury at DNA pioneer's theory: Africans are less intelligent than Westerners. The Independent. 2007 [Google Scholar]
  26. Need AC, Goldstein DB. Next generation disparities in human genomics: concerns and remedies. Trends in Genetics. 2009;25:489–494. doi: 10.1016/j.tig.2009.09.012. [DOI] [PubMed] [Google Scholar]
  27. Newport MJ, Rotimi CN. Reducing the global genomic inequity gap: development of an African genome project. Public Health Genomics. 2009;12:251–252. doi: 10.1159/000197973. [DOI] [PubMed] [Google Scholar]
  28. Parfitt T, Egorova Y. Genetics, mass media, and identity: A case study of the genetic research on the Lemba and Bene Israel. Routledge; London: 2005. [DOI] [PubMed] [Google Scholar]
  29. Parker R, Aggleton P. HIV and AIDS-related stigma and discrimination: a conceptual framework and implications for action. Social Science & Medicine. 2003;57:13–24. doi: 10.1016/s0277-9536(02)00304-0. [DOI] [PubMed] [Google Scholar]
  30. Phelan JC. Geneticization of deviant behavior and consequences for stigma: the case of mental illness. Journal of Health & Social Behavior. 2005;46:307–322. doi: 10.1177/002214650504600401. [DOI] [PubMed] [Google Scholar]
  31. QSR . NVivo 8. Melbourne, Australia: 2009. [Google Scholar]
  32. Quinn-Patton M. Qualitative research and evaluation methods. Sage Publications; Thousand Oaks, CA: 2002. [Google Scholar]
  33. Reardon J. Race to the finish. Identity and governance in an age of genomics. Princeton University Press; Princeton and Oxford: 2005. [Google Scholar]
  34. Rose N. Introduction to the discussion of race and ethnicity in nature genetics. BioSocieties. 2006;1:307–311. [Google Scholar]
  35. Rosenberg NA, Huang L, Jewett EM, Szpiech ZA, Jankovic I, Boehnke M. Genome-wide association studies in diverse populations. Nature Reviews Genetics. 2010;11:356–366. doi: 10.1038/nrg2760. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Sankar P, Cho M, Root Wolpe P, Schairer C. What is in a cause? Exploring the relationship between genetic cause and felt stigma. Genetics in Medicine. 2006;8:33–42. doi: 10.1097/01.gim.0000195894.67756.8b. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Shanawani H, Dame L, Schwartz DA, Cook-Deegan R. Non-reporting and inconsistent reporting of race and ethnicity in articles that claim associations among genotype, outcome, and race or ethnicity. Journal of Medical Ethics. 2006;32:724–728. doi: 10.1136/jme.2005.014456. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Simpson B. Imagined genetic communities: ethnicity and essentialism in the twenty-first century. Anthropology Today. 2000;16:3–6. [Google Scholar]
  39. Tekola F, Bull S, Farsides B, Newport M, Adeyemo A, Rotimi C, et al. Impact of social stigma on the process of obtaining informed consent for genetic research on podoconiosis: a qualitative study. BMC Medical Ethics. 2009;10:13. doi: 10.1186/1472-6939-10-13. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Teo YY, Small KS, Kwiatkowski DP. Methodological challenges of genome-wide association analysis in Africa. Nature Reviews Genetics. 2010;11:149–160. doi: 10.1038/nrg2731. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. The MalariaGEN Consortium A global network for investigating the genomic epidemiology of malaria. Nature. 2008;456:732–737. doi: 10.1038/nature07632. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Tindana PO, Singh JA, Tracy CS, Upshur REG, Daar AS, Singer PA, et al. Grand challenges in global health: community engagement in research in developing countries. PLoS Med. 2007;4:e273. doi: 10.1371/journal.pmed.0040273. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Wellcome Trust Case Control Consortium Genome-wide association study of 14,000 cases of seven common diseases and 3,000 shared controls. Nature. 2007;447:661–678. doi: 10.1038/nature05911. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Wensley D, King M. Scientific responsibility for the dissemination and interpretation of genetic research: lessons from the ‘warrior gene’ controversy. Journal of Medical Ethics. 2008;34:507–509. doi: 10.1136/jme.2006.019596. [DOI] [PubMed] [Google Scholar]
  45. WHO . Report of the Advisory Committee on health research. World Health Organisation; Geneva: 2002. Genomics and world health. [Google Scholar]
  46. Yang LH, Kleinman A, Link BG, Phelan JC, Lee S, Good B. Culture and stigma: adding moral experience to stigma theory. Social Science & Medicine. 2007;64:1524–1535. doi: 10.1016/j.socscimed.2006.11.013. [DOI] [PubMed] [Google Scholar]

RESOURCES