Reporting and representation of race/ethnicity in 1683 RCTs of pharmacotherapy for mental disorders: a meta-analysis

Abstract

Importance: representation of race and ethnicity in randomised controlled trials (RCTs) is critical for understanding treatment efficacy across populations with different racial/ethnic backgrounds.

Objective: we conducted the first comprehensive overview of reviews of race and ethnicity representation and reporting across RCTs of pharmacotherapies for mental disorders.
Data sources: PubMed/Medline/Ovid-Embase/APA-PsycINFO/Web-of-Science were searched (March-01-2024), to retrieve the most updated and largest network meta-analyses (NMA) of RCTs of pharmacotherapies for ICD-10 mental disorders.
Study selection: From each NMA, we identified single-blind or double blind RCTs, quasi-RCTs, open label RCTs, and discontinuation/withdrawal design RCTs, recruiting people of any age with a diagnosis of mental disorder and testing the efficacy of any pharmacological intervention compared to any control arm.
Data Extraction and Synthesis: We used random-effects logit-transformed-proportion meta-analyses to: 1) estimate prevalence rates of race and ethnicity groups and their temporal trends across RCTs, and 2) compare US-RCT prevalence rates with US census data.

Main outcome measures: reporting of data on race and ethnicity, and percentage of Asian, Black, Hispanic, Other/Multi-racial/Multi-ethnic, and White. Hispanic and White were operationalised as “White-including-Hispanic" and "Hispanic-among-White" to fit the categories used in the included RCTs. We also considered: year of publication, type of RCT, continent, age group, and sample size.

Results: we included 1683 RCTs (375,120 participants, 91.7% adults; USA=680, Europe=404, multiple continents=216). Race and ethnicity were reported in 39% of RCTs, reporting was the highest in USA-based RCTs (58.68%) and the lowest in Central and South America (8.70%) and Asia/Middle East (12.43%). 80.2% of participants [95% confidence interval=78.8-81.5] self-reported as White-including-Hispanic (of which 11.0% [9.1-13.3] as Hispanic-among-White), 9.0% [8.1-10.0] as Black, 5.8% [5.2-6.4] as Other/Mixed Ethnicity, and 2.7% [2.1-3.5] as Asian. We found better reporting of race and ethnicity in US-RCTs over time (log odds increase=0.066); and smaller improvement in non-US-RCTs (log odds increase=0.023). Larger RCTs were more likely to report race and ethnicity, albeit not in all continents. “Other/Mixed Ethnicity” was historically over-represented in US-RCTs, while “Black”, “Hispanic-among-White” and “Asian” were under-represented, with the latter two still currently under-represented.

Conclusions and relevance: differences in reporting race and ethnicity across countries and under-representation of certain racial/ethnic groups in US-based RCTs highlight the need for international guidelines to ensure equitable recruitment in clinical trials.

More information

Identifiers

ORCID for Alessio Bellato: orcid.org/0000-0001-5330-6773

ORCID for Valeria Parlatini: orcid.org/0000-0002-4754-2494

ORCID for Samuele Cortese: orcid.org/0000-0001-5877-8075

Catalogue record

Export record

Altmetrics