Assessing non-white ethnic participation in type 2 diabetes mellitus randomized clinical trials: A Meta-Analysis

ImportanceThe prevalence of type 2 diabetes mellitus (T2DM) is increasing globally, and the greatest burden is borne by non-white ethnic groups. Randomized clinical trials (RCTs) provide evidence regarding the optimal medical therapy for the treatment of T2DM patients and inform national and international guidelines. However, there are concerns that the enrollment of ethnically diverse people into these trials is limited, which has resulted in a lack of ethnic diversity in RCTs of T2DM. Furthermore, the extent of underrepresentation may differ according to whether a trial is government-funded or industry-funded. ObjectiveTo systematically review and meta-analyze the proportion of non-white and white participants relative to their disease burden of T2DM included in large, influential government- and industry-funded RCTs of T2DM pharmacotherapies. Data SourcesThe PubMed electronic database was searched from January 2000 through January 2021. Study SelectionReports of RCTs of T2DM medications with a total sample size of at least 100 participants published in the year 2000 onwards, in high impact general medical journals (i.e., impact factor >10), were included. Data Extraction and SynthesisData including the number of participants, proportion of participants by ethnicity, and funding sources, were extracted from trial reports. Main Outcomes and MeasuresThe main outcome was the participation-to-prevalence ratio (PPR), which was calculated for each trial by dividing the percentage of white and non-white participants in the trial by the percentage of white and non-white participants with T2DM for the countries or regions of recruitment represented in each trial. A random-effects meta-analysis was used to generate the pooled PPR and 95% confidence intervals (CI) across study types. A PPR <0.80 indicates underrepresentation and >1.20 indicates overrepresentation. ResultsA total of 82 trials were included involving 296,964 participants: 14 were government-funded trials, and 68 were industry-funded trials. For government trials, the PPR for white participants was 1.11 (95% CI; 1.00-1.23) and for non-white participants was 0.73 (95% CI:0.62-0.86). Among industry trials, the PPR for white participants was 2.19 (95%CI: 1.91-2.50), and the PPR for non-white participants was 0.33 (95%CI: 0.29-0.38). Heterogeneity was high across all PPRs. Conclusions and RelevanceNon-white participants are underrepresented in both government- and industry-funded T2DM trials, compared to white participants. The greatest disparity in ethnic diversity in RCTs is observed for industry-funded trials. Key PointsO_ST_ABSQuestionC_ST_ABSWhat is the representation of non-white participants in type 2 diabetes randomized clinical trials relative to their disease burden? FindingsIn this meta-analysis, non-white participants are underrepresented in both government-and industry-funded type 2 diabetes randomized trials, compared to white participants. The greatest disparity in ethnic diversity in randomized trials was observed for those funded by industry. MeaningDeliberate strategies to improve recruitment and enrolment of diverse participants proportional to the type 2 diabetes disease burden into industry and government-funded randomized controlled trials are needed to enhance the generalizability of research findings.