Statistics from Altmetric.com
The research letter by Sheikh and co-authors1 addresses an important area but, regrettably, does not hit the nail on the head. Information on ethnicity of study participants is often missing in asthma studies, as the authors clearly show. However, presenting this information—for example, in the table of baseline characteristics when reporting a clinical trial—is only a first step. Much more important is the inclusion of this characteristic in the analysis of effect modification or subgroup analysis. Only then will we know whether ethnicity really matters when applying a certain intervention.
I would welcome a further report from the authors giving details about the actual incorporation of ethnicity in data analysis.
We are grateful to Dr van der Wouden for raising the important point of ensuring that ethnicity data are used when analysing trial results. Of the 23/70 studies (32.8%) reporting information on ethnicity in the trial report, all confined mention of ethnicity only to baseline characteristics of included participants. Thus, none of these 23 studies factored ethnicity into the analysis of results. A possible explanation for this rather disappointing observation is that the number of subjects recruited from minority ethnic groups tended to be small, thereby precluding any meaningful ethnic-specific subgroup analysis. Unfortunately, this is a problem not solely confined to asthma studies; for example, less than 30% of clinical trials on epilepsy which reported ethnicity used the information in the analysis.1
In designing clinical trials, pragmatic and cost considerations often force investigators to focus on the main objectives of the study, forcing considerations concerning subgroup analyses onto the back burner during trial planning. Most clinical trials therefore lack sufficient power for subgroup analyses;2 however, this appears to be a problem that is particularly common in relation to subgroup analysis by ethnicity.3
Our motivation for undertaking this study was to highlight the disparity in recruitment between ethnic minorities and the majority into trials in a disease where ethnic considerations may be relevant. Asthma is one such area, but there are also others.4 van der Wouden is entirely correct in asserting that improving recruitment of minority ethnic groups is only half the battle; at least equally important—if not more so—is ensuring that information on ethnicity is meaningfully and competently used when analysing data and presenting results.5 Both issues should now be prioritised by the respiratory research community.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.