Medical Law International - Social Inclusivity Vs Analytical Acuity? A Qualitative Study Of UK Researchers Regarding The Inclusion Of Minority Ethnic Groups In Biobanks

SOCIAL INCLUSIVITY VS ANALYTICAL ACUITY?
A QUALITATIVE STUDY OF UK RESEARCHERS REGARDING THE INCLUSION OF MINORITY ETHNIC GROUPS IN BIOBANKS
ANDREW SMART
Bath Spa University

RICHARD TUTTON
Lancaster University

RICHARD ASHCROFT
Queen Mary, University of London

PAUL MARTIN, ANDREW BALMER, RICHARD ELLIOT
University of Nottingham

GEORGE T.H. ELLISON
St George’s, University of London

E DISCUSSION

   The researchers we interviewed were involved in studies exploring the aetiology of diseases which affect all sectors of the population, and they all felt that potential biological differences relating to the ancestry of their study participants could be relevant to the collection, analysis and interpretation of data for their research. While they were therefore concerned about how best to include some measure of ethnicity in their study designs, they perceived themselves as caught between the twin imperatives of social inclusivity and analytical acuity. Indeed, some felt compelled to meet related demands on the inclusion of minority ethnic groups imposed by research funders or regulators. These were demands that some thought could involve methodological directives with questionable scientific value, so that if they complied, they might find themselves generating evidence from studies which included sampling choices over which they would have less than total control, and which they themselves might not consider valid, reliable or generalisable. Before discussing the implications of these findings, and without wishing to discount the opinions expressed by the researchers we interviewed, there are (at least) three challenges to their viewpoint which require consideration.
   First it should be noted that there is no consensus on the potential impact of sampling strategies that include participants from different ethnic groups (i.e., ‘ethnically diverse study samples’) on the analytical acuity of biomedical research. This is not least because the nature, extent and biomedical relevance of genetic and phenotypic differences between ethnic groups are a matter of intense, ongoing debate (for example, see Smart, et al. (2006) for an analysis of the divergences in perspective evident in the 2004 Nature Genetics Special Issue ‘Genetics for the Human Race’). Despite a substantial body of published research describing a range of genotypic and phenotypic differences amongst ethnic groups from studies conducted throughout the world (e.g., Cruikshank and Beevers, 1989; Macbeth and Shetty, 2001), the internal and external validity of this research is undermined by the fluid nature of ethnicity and the crude ethnic categories used to operationalise this (see Ellison and Jones, 2002). This said, genetic and phenotypic differences have been observed in studies examining the least fluid ethnic groups, and those using the most ‘robust’ ethnic categories available. However, these observed differences are predominantly differences in the frequency of a small number of genetic traits and phenotypic conditions (i.e., traits and conditions that occur in all ethnic groups but to differing degrees) and include a number that are likely to be biomedically ‘neutral’ (i.e., have no known or likely positive or negative impact on mechanisms that determine health or disease; Graves, 2006). At the same time, it is known that many of the social processes involved in the formation of ethnic groups (such as colonisation and migration) and in the production of socio-economic inequalities amongst ethnic groups (such as racism and ethnocentrism) can create profound phenotypic differences between ethnic groups (some of which are likely to be biomedically important; Karlsen and Nazroo, 2002).Under these circumstances, it will be difficult to know whether any of the differences that actually exist between ethnic groups will mean that studies that include participants from different ethnic groups will ordinarily involve a reduction in analytical acuity.
   Second, the researchers we interviewed conceptualised ethnicity as having both socio-cultural and genetic components, but their concerns about analytical acuity primarily revolved around the potential genetic differences between ethnic groups. For example, none of our respondents emphasised the potential impact of socio-cultural or socio-economic differences between ethnic groups (or their phenotypic consequences) on the acuity of their research, despite the fact that many were involved in studies focusing on diseases in which such differences are believed to play an important (if not a predominant) role in generating ethnicised inequalities in health (Hayward et al. 2000). This pre-occupation with genetic differences may indicate a conceptual bias that could, itself, have substantial implications on the conduct and interpretation of biomedical research. On the one hand the underestimation of the potential role of socio-cultural and socio-economic factors (and their phenotypic consequences) in ethnicised health inequalities could, itself, compromise analytical acuity. On the other hand, focusing on (presumed) genetic differences between ethnic groups only adds fuel to concerns about the revival of discredited ideologies which propose that different human groupings have distinct biological (and even socio-cultural) characteristics that are innate, essential and immutable.
   Third, the ethnic group categories that researchers adopted are unlikely to be wholly accurate markers of the variations associated with ethnicity that may be of relevance in biomedical science (including genotypic, phenotypic, socio-cultural and socio-economic characteristics). Certainly, many people, including the researchers we interviewed, recognise that there are difficulties with operationalising and measuring ethnicity in a reliable and valid manner (e.g. Aspinall, 1997, 2001; Bradby, 2003). In light of these uncertainties there are growing calls for biomedical researchers to declare the reasoning behind the ethnic categories they use in their research, and to carefully consider the impact of the limited reliability and validity of such categories when interpreting the results of their analyses (Ellison et al. 2007; Smart et al. 2008). With this in mind, the concerns about analytical acuity expressed by many of the researchers we interviewed are far from being substantiated and doing so would require careful consideration of the specifics of the studies, the ethnic groups involved and the quality of the measures used for capturing any variations in characteristics that might be potentially relevant to the analysis.
   Despite these three challenges to the researchers’ perception that the inclusion of participants from different ethnic groups could affect analytical acuity, their viewpoint has significant implications for the governance of some biobanks. When faced with what they viewed as conflicting imperatives, our interviewees were (perhaps unsurprisingly) more comfortable discussing genetic variation and the (apparent) certainties of how legitimate sample sizes are best determined using power calculations. Indeed, Epstein (2004a) notes a similar response in his study, in which the researchers he interviewed recognised the importance of social factors in gender classification, but primarily talked about gender in biological terms and were only comfortable generalising on the basis of biological differences. These ‘leanings’ towards statistical and biological ‘certainties’ are important because they may influence how researchers seek to resolve any conflict they perceive to exist between the competing imperatives of inclusivity and acuity. The resolution of such perceived conflicts is likely to have a significant impact about what they consider ‘appropriate’ research designs and practices. In this way, the logic of either imperative could differentially influence questions about: what research should be undertaken; how the research question/problem should be conceptualised; how the research should be designed; which ethnic groups should be included; how ethnicity should be operationalised; and how any differences observed between ethnic groups should be interpreted and explained. Thus, on the one hand, a belief in the preeminence of the imperative for social inclusivity might emphasise the importance of including all ethnic groups in such research, to ensure that its potential benefits (and risks) are shared equally by all sectors of the wider population. On the other hand, a belief in the pre-eminence of the imperative for analytical acuity might lead to calls for ethnic groups to be analysed separately with studies that are specifically designed to generate evidence of equal statistical and analytical strength for each ethnic group.
   Our findings show that there is a potential for biomedical researchers to be pre-occupied with genetic rather than social factors, and to be more concerned with analytical acuity than social inclusivity. As such, it seems likely that interaction and debate with policymakers, lawyers, social scientists and other interested parties will be necessary to address the conflict biomedical researchers perceive to exist between the imperatives of inclusivity and acuity. This is likely to require debate over the evidence available (and the evidence required) to substantiate concerns about the impact of including participants from different ethnic groups on the analytical acuity of biomedical studies. Without assuming that it will be possible to reach a consensus on these issues, there will also be a need to explore the practical, economic, ethical and scientific circumstances under which inclusivity should drive sampling strategies; and when substantiated concerns over analytical acuity might override the moral, political and ethical benefits of inclusivity. Such discussions will also inevitably raise additional questions concerning regulation and governance, such as: whether decisions about the pre-eminence of inclusivity or acuity will need to be guided, controlled or monitored; if so, how and by whom; and how changing national demographics and socio-political imperatives might affect such decisions. At the same time, the debate is also likely to raise a number of questions specific to the conduct of certain biobanks with respect to choices about study design and the recruitment of participants, as was evident in the discussions that occurred within UK Biobank (Tutton, 2008). Finally, it is likely that this debate will have to confront challenging (and longstanding) questions about the legitimacy and broader social consequences of research that focuses on differences, rather than commonalities, between ethnic groups, in particular its potential to contribute to scientific racism (Ellison, 2006).
   Clearly, there is still considerable debate as to whether it is scientifically or socially ‘sound’ to use ethnic categories to design, analyse or interpret the results of genetic and biomedical research. In relation to the legal and ethical governance of biobanks and biobanking research in the UK, our analyses indicate that at least some researchers feel obliged to collect data on ethnicity, and/or recruit research participants from different ethnic groups. Other researchers, however, seem likely to exclude participants from some ethnic groups (particularly minority ethnic groups), justifying this on the basis of concerns about analytical acuity. Under the RRA (as amended in 2000), publicly-funded biobanks are obliged to collect information on ethnicity to demonstrate that there is no unintentional or unjustifiable discrimination in their activities. However, there appears to be no formal imperative for UK researchers to use this information, either for designing their research or for analysing their findings. Indeed, it could be argued that there are good scientific arguments for not using data on ethnicity in biomedical research, although only if it could be proven that data on ethnicity are unable to sustain reliable or valid analyses and this is likely to vary from one study to the next. Similarly, it could be argued that there are good scientific arguments for not obliging researchers to include participants from different ethnic groups in their studies’ sampling frames although, again, only if there were well-founded reasons for suspecting that such sampling frames might compromise the validity of the research and again, this is likely to vary from one study to the next.
   We have reported that some researchers working on biobanks studies of common complex disease in the UK felt that they faced a dilemma: should they conduct research that they perceive to be scientifically unsound because of the principle of justice favouring equality of opportunity and a prohibition against ‘racial’ discrimination? We have argued that, given the current levels of scientific uncertainty and the significant implications of this issue, this matter should urgently become a topic for debate and discussion between relevant professional and stakeholder groups. In the event that genuine differences in biomedical risk or need amongst different ethnic groups are substantiated, it seems likely that biomedical scientists would then demand to be allowed to design research capable of meaningfully examining ethnic groups separately. Such segregated research designs would have important consequences for the governance, organisation and conduct of some biobank studies. Furthermore, regardless of whether the group differences that would be used to justify such research designs are genetic or social in origin, biomedical science which segregates by ethnicity runs the risk of contributing to a popular perception of ethnic groups as socially ‘real’, ‘biologically meaningful’ entities.

<< D. Findings | Acknowledgements >>