Underrepresentation of black and ethnic minority patients in clinical trials must be addressed
3 October 2016
Lack of standardised approaches to assessing language proficiency and limited tailoring to people from different cultural backgrounds may reduce opportunities for ethnic minority patients to participate in randomised controlled trials (RCTs). As a result, researchers call for developing and adopting valid language assessment tools to support RCT recruitment.
Researchers from the Universities of Bristol and Nottingham have advocated for change after the results of a systematic review showed that RCTs may be excluding black and ethnic minority (BAME) patients based on inconsistent assessments of their language skills.
The study, which reviewed ethnic minority representation and the role of language assessment in recruiting type 2 diabetes patients to telehealth trials, found a lack of common procedures to assess language proficiency during recruitment, and inconsistent reporting of the ethnic composition of the recruited patient sample across studies.
“Randomised controlled trials test the effectiveness of new medical treatments. Ideally, the group of patients participating in a trial should reflect the wider population who will be using the treatment.”
Type 2 diabetes is more likely to affect BAME patients, and so underrepresenting these patients in trials makes it difficult to generalise the findings and to determine whether the treatment is beneficial for the wider patient population.
Nearly two-thirds of the RCTs included in the review provided information about the ethnic composition of the sample of patients recruited. However, less than a third of the studies that reported on this recruited a sizeable proportion of ethnic minorities (defined as over 30% of the total number of participants).
The study looked at research from all Western countries where English is an official language, yet of the 58 telehealth trials in the review, the 14 trials which recruited at least 30% BAME patients were all US-based.
Although the USA has a larger proportion of BAME patients compared to the other countries in the study, it also legislates on the make-up of the patient sample in medical research. US National Institutes of Health (NIH) policy states that “women and members of minority groups and their subpopulations must be included in all NIH-funded clinical research”, unless a clear and compelling rationale and justification exists.
Dr Isaacs commented:
“As the NIH say, it is essential to establish whether or not the intervention affects members of minority communities differently than it does patients from white backgrounds. The greater inclusion of ethnic minority patients in trials in the USA likely, at least in part, reflects this policy.”
The researchers also found that, when making decisions about patient eligibility for RCT participation, language proficiency was only mentioned in half of the trials studied and none specified how the stated language criterion was measured.
Dr Isaacs said this could be addressed by equipping trial recruiters with practical solutions:
“A language assessment tool should be developed in future research to minimise the possibility of patients being unfairly excluded from RCTs based on trial recruiters’ arbitrary judgements to do with patients’ language proficiency.”
“As a result of these findings, we recommend clearer guidelines for reporting on recruited patients’ sociodemographic characteristics, including language background and ethnicity. This is particularly pressing for conditions like type 2 diabetes, which ethnic minority communities are more susceptible to. Here, in the UK, we should be looking at global best practice and learning from it.”
Dr Leila Rooshenas, who also co-authored the article, is a member of ConDuCT-II Hub at Bristol’s School of Social and Community Medicine, which conducts leading methodological research on RCTs. She emphasised one of the implications from the study; that minority patient underrepresentation could be partly addressed by updating the most widely-used methods for reporting of trials:
“There is currently no formal requirement to provide detailed information about the processes that occur between the ‘screening’ and ‘randomisation’ stages of the CONSORT [Consolidated Standards of Reporting Trials] diagram. This makes it difficult to always know what happens in the early stages of RCT recruitment. Practices of excluding or not approaching patients, and the reasons underlying this (e.g. language difficulties) could be much more transparent.”
Dr Isaacs added that the interdisciplinary nature of the study, combining applied linguists with trials methodology researchers, adds value to this research area.
“When we consider how to optimise recruitment to trials to find the best treatments and ultimately to provide more equitable access to health services for all patients, we must examine the wider challenges, including non-health-related obstacles. Crossing disciplinary boundaries can be practically challenging, but when research communities are open to it, it can truly lead to new innovation.”
This study, which was part of a Marie Curie grant, was carried out by an interdisciplinary research team at the Universities of Bristol and Nottingham. It was funded through a Marie Curie grant from the European Commission and supported by the Medical Research Council’s (MRC) ConDuCT-II Hub at Bristol's School of Social and Community Medicine. The study was partially motivated by outputs from the Hub’s Theme 2, which focuses on ways of optimising recruitment to RCTs. It also closely links with pioneering work by Bristol’s QuinteT team on integrating qualitative methods in the design and implementation of RCTs.
‘The inclusion of ethnic minority patients and the role of language in telehealth trials for Type 2 diabetes: A systematic review’ by Isaacs, T., Hunt, D., Ward, D., Rooshenas, L., & Edwards, L. in The Journal of Internet Medical Research.
The original press release announcing the study findings, with further details of the teams involved, can be found here.