Statistics from Altmetric.com
Ethnic variations in the prevalence, severity, and management of asthma have been reported1,2 but, apart from these recent reports, our understanding of the relationship between ethnicity and asthma is limited. This is a concern as ethnicity may be an important confounder in asthma studies through such varied influences as differences in lung function, socioeconomic disparities, and concordance with treatments.3 Proactive ethnic trial recruitment policies exist in the US but there is no comparable legislation in Europe. We sought to investigate the hypothesis that asthma trials conducted in the US are more likely than European studies to report on the ethnicity of subjects.
Our sampling frame was the Cochrane Airways Group Trials Register which contains records of published and unpublished clinical trials. We manually searched this database for English language reports of randomised controlled trials of asthma reported during the period 2000–2. Clinical trials conducted in either the US or Europe were identified and 35 European and 35 US studies were selected using simple random sampling. Full text hard copies of all selected articles were obtained for detailed assessment of any mention of the ethnicity of study participants. A deliberately broad working definition of “ethnicity” was used that included any reference to race, ethnic origin, language, or nationality. On the basis of this information the studies were categorised as either detailing the ethnicity of subjects or not. One reviewer systematically extracted data on whether the ethnicity of subjects was reported using a pre-piloted data extraction form and a second reviewer independently verified a sample of extracted data. Disagreements were resolved through discussion with provision to refer to a third reviewer if necessary.
Descriptive statistics were used to determine the proportion of studies reporting information on the ethnicity of study subjects and the χ2 test was used to compare reporting of ethnicity in European and US published trial reports. Assuming that 50% of US and 15% of European asthma trials report on the ethnicity of participants, we estimated that we would need to extract data on a total of 64 studies (32 from each area) in order to have 80% power of detecting a difference at the 5% significance level.
Overall, 23 of 70 reports (32.9%) included information on the ethnic profile of participants (table 1). US studies (n = 22) were significantly more likely to report on ethnicity than European studies (n = 1): 62.9% v 2.9%; RR = 22; 95% CI 3.1 to 154.4; p<0.0001. Among the European trials, 13 were UK based and the single European trial reporting ethnicity was from the UK.
Thus, recent asthma trials conducted in the US are 22 times more likely than those conducted in Europe to report information on the ethnicity of study participants. The random selection procedures adopted to identify suitable trials and the standardised extraction of data with independent verification and an “a priori” agreed approach to handle disagreements should, we believe, have minimised the impact of selection and/or information biases affecting these findings.
Our results seem likely to reflect proactive policies in the US. For example, all federally supported programmes with sufficient sample size are required to report statistics according to race/ethnicity.4 Furthermore, the National Institute of Health regularly issues guidelines regarding inclusion of women and ethnic minorities into clinical trials, and less than 4% of grant applications are reported to be in breach of these guidelines.
Although the US fares better than Europe, it is still worrying that only 62.9% of recently published clinical trials from the US report on the ethnicity of participants. This finding may have several possible explanations, including difficulties in identifying, enrolling, communicating with, and following up patients from minority ethnic groups in asthma trials. Mechanisms and standards to ensure inclusion and reporting of ethnic minority communities in asthma studies are needed. These are absent in current trial reporting guidelines (CONSORT) and we suggest that, where appropriate, the merits of insisting on presentation of such data should be debated.5 Although the Standards for Reporting of Diagnostic Accuracy (STARD) statement is a move in the right direction, this still lacks explicit requirements for the reporting of the ethnicity of study participants. In addition, European governments and respiratory bodies should consider the US model for promoting the inclusion of participants from ethnic minorities in asthma research.