Intended for healthcare professionals


Representation of South Asian people in randomised clinical trials: analysis of trials' data

BMJ 2003; 326 doi: (Published 05 June 2003) Cite this as: BMJ 2003;326:1244
  1. Su Mason (medsam{at}, principal research fellow1,
  2. Mahvash Hussain-Gambles, research fellow2,
  3. Brenda Leese, reader in primary care research2,
  4. Karl Atkin, senior lecturer2,
  5. Julia Brown, head of unit1
  1. 1 Northern and Yorkshire Clinical Trials and Research Unit, University of Leeds, Leeds LS2 9LN,
  2. 2 Centre for Research in Primary Care, University of Leeds
  1. Correspondence to: S Mason
  • Accepted 5 February 2003


Excluding patients of ethnic minority groups from clinical trials is unethical,1 introduces substantial bias, and means that findings are based on unrepresentative populations.2 The National Institutes of Health Revitalization Act 1993 requires that all minority groups be represented in the sample in research projects supported by the National Institutes of Health, unless there is a clear and compelling justification not to do so. In the United Kingdom no such legislation exists.

Participants, methods, and results

To determine the proportion of South Asian participants (the biggest minority ethnic group in Britain) included in clinical trials we investigated the ethnicity profile of six clinical trials recently conducted by the Northern and Yorkshire Clinical Trials and Research Unit, University of Leeds. All were phase III, multicentre, randomised, controlled trials and had recruited large numbers of participants and centres. Two were national breast cancer trials for which regional recruitment figures were available; two were national gynaecological trials (of different surgical techniques for hysterectomy and for ovarian cancer); one was a national, minimally invasive trial of surgery in colorectal cancer; and the sixth was a regional study investigating the eradication of Helicobacter pylori in general practice. None contained exclusion criteria that were directly related to ethnic origin, but all patients' information sheets and consent forms were in the English language only.

We identified ethnic origin in the baseline data of three trials and identified South Asian patients by an SQL programming query. For the three remaining trials, we analysed the names of the participants by using a computer program (Nam Pehchan) developed by Bradford Health Authority. Researchers experienced in analysing South Asian names manually checked for discrepancies. The Nam Pehchan programme was not infallible. It did not differentiate between different Muslim surnames (for example, South Asian, Far Eastern, or Arabic). Therefore, forenames (which are often more specific to language or region than surnames) were needed to identify some South Asian participants.

South Asian (Indian, Pakistani, and Bangladeshi) people comprised up to 1.7% (mean 0.6%) of total participants in the six trials (table). A community trial in Leeds and Bradford recruited the highest number of South Asian participants, but this number was lower than expected compared with estimates of population figures from the Office for National Statistics.3

Proportion of South Asian participants in six multicentre clinical trials

View this table:


People of South Asian ethnic origin seem to be under-represented in clinical trials. Though not previously supported by data, this has been of concern to researchers. This under-representation might be due to investigator bias, inappropriate strategies for recruitment, or cost issues—for example, for translators or translations of information sheets.4

Such inequality in the ethnic origin of participants in trials has ethical and scientific ramifications because genetic predisposition, dietary intake, and exposure to environmental and occupational hazards lead to ethnic differences in susceptibility to diseases. Furthermore, patients' response to drugs, how they metabolise drugs, and their concurrent diseases, as well as the side effects of drugs, can vary between different ethnic groups.5

Our small survey used only the rather crude comparative data available for the expected South Asian population. We would have liked to compare more precisely the trials' inclusion criteria of age range and sex with those of the expected South Asian population, but this information is not currently available. Trials that recruited older patients would not be expected to comprise a percentage of South Asian people equal to the overall figures of the Office for National Statistics because the number of elderly South Asian people in the UK population is small.

Increased awareness and monitoring of recruitment and retention of ethnic minority groups in clinical trials are needed, and analysis of data by ethnicity of subjects should be done consistently. More rigorous review by the research ethics committee of clinical trial protocols, payment for translation of information supplied to participants, community participation, and education of ethnic minority groups may contribute to attaining proportional representation of ethnic minorities in trials.


  • Contributors SM had the idea to write the report, collated the data, helped to draft the report, and critically revised it. MH-G, JB, BL, and KA helped to formulate and revise the report. MH-G also helped to draft the report. All authors have seen and approved the final version. SM is the guarantor.

  • Funding NHS Health Technology Assessment Programme. The views in the paper are those of the authors and not necessarily those of the Department of Health.

  • Competing interests None declared.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.