Racial Disparities Skew Most Type 2 Diabetes Clinical Trials Racial Disparities Skew Most Type 2 Diabetes Clinical Trials

Researchers published the study covered in this summary on medRxiv.org as a preprint that has not yet been peer reviewed.

Key Takeaways

  • A meta-analysis of 82 published randomized clinical trials (RCTs) of treatments for patients with type 2 diabetes showed that non-White participants appear to be underrepresented relative to their disease burden, while White participants appear over-represented.

  • The greatest disparities in diversity were found in industry-funded trials, which constituted most of the trials examined.

Why This Matters

  • The burden of type 2 diabetes is disproportionately higher in non-White than White racial and ethnic groups.

  • RCTs produce the most reliable evidence on efficacy and safety of various interventions, and they inform society recommendations and public health prevention strategies.

  • Underrepresentation of non-White individuals in RCTs could limit the generalizability and uptake of the trial findings among groups with the highest disease burden.

Study Design

  • The report focused on a systematic review and meta-analysis of 82 RCTs with a total of 296,964 participants. Industry sources sponsored about 83% of the trials and government-funded entities sponsored the remaining 17%.

  • More than 60% of the studies included were run between 2011 and 2020, and all studies dated from after 1999. The RCT reports had primarily been published in top-tier journals: The New England Journal of Medicine, The Lancet, the Journal of the American Medical Association, and The BMJ, which in particular, showcased industry-sponsored trials. To add non-industry funded trials, the analysis also included studies published in Circulation, Diabetes Care, and Annals of Internal Medicine.

  • The authors used participation-to-prevalence ratios (PPR) to estimate trial enrollment relative to local disease burden by race. These ratios were derived from the percentage of White or non-White people in the trial divided by the percentage of White or non-White people with type 2 diabetes for populations that supplied participants for each study. A ratio of 1 corresponds to a trial with a racial composition that matched the population that supplied patients for the study. A ratio of less than 0.80 indicates a study with under-representation of a racial group, and a ratio greater than 1.2 indicates over-representation.

Key Results

  • The proportion of non-White participants in the trials increased significantly from 10.9% in 2000-2005 to 23.8% in 2006-2011, then remained relatively constant in 2006-2020. Non-White participation was 23.4% across all 82 studies.

  • The most proportional recruitment of non-White populations occurred in trials that recruited primarily in North America, which averaged 32% non-White participation, and the United States, which averaged 26% non-White participation.

  • The proportion of non-White participation was higher in government-funded than industry-funded RCTs, at 25.3% compared with 23.1%.

  • The pooled PPR for White participants was 2.19 for industry trials, consistent with overrepresentation, and 1.11 in government trials, consistent with proportional representation.

  • The PPR for non-White participants was 0.33 for industry trials and 0.73 for government trials. Both ratios indicate underrepresentation of non-White study participants.


  • Limited data exist for type 2 diabetes prevalence and demographics in some parts of the world, and so the authors estimated those values.

  • The analysis focused exclusively on reports published after 1999.

  • The pooled estimates in the meta-analysis demonstrated a high level of heterogeneity, but their directionality and magnitude of effect showed cross-trial consistency.


This is a summary of a preprint research study “Assessing non-white ethnic participation in type 2 diabetes mellitus randomized clinical trials: A meta-analysis,” written by researchers at McMaster University, Hamilton, Ontario, Canada on medRxiv provided to you by Medscape. The study has not yet been peer reviewed. The full text of the study can be found on medRxiv.org.