Reporting of race,ethnicity,and gender in lower-extremity peripheral artery disease randomized controlled trials

Randomized controlled trials (RCTs), the clinical research gold standard, must represent the disease of interest’s patient population, including those from structurally marginalized communities who disproportionately experience adverse outcomes. Reports have shown for decades that Black, Hispanic, and Native American communities have a higher peripheral artery disease (PAD) prevalence and major lower-extremity amputation rates than their White counterparts. ¹ To ascertain RCT representativeness, we must first elucidate race or ethnicity reporting frequency. We examined demographic reporting and potential drivers of reporting in PAD RCTs.

This systematic review of PAD RCTs was exempted by our Institutional Review Board. PubMed, CINAHL, Cochrane Central, Embase, and Web of Science were searched on June 30, 2023 using both controlled vocabulary (when applicable) and keywords related to ‘peripheral artery disease.’ We limited the search to articles published in English between February 29, 2000 to December 31, 2022. Studies that included children were excluded. Commentaries, editorials, letters, preprints, and reviews were excluded.

A depiction of search criteria is provided in Supplemental Table 1. The initial search yielded 15,061 records, which were imported into EndNote20 (Clarivate). The ‘Find duplicates’ function in EndNote was used to remove 5080 duplicates, and the remaining records (n = 9981) were uploaded to Covidence, which excluded another 2066 duplicates. An additional 140 duplicates were manually removed. The abstracts of 7775 records were manually reviewed, and 411 full-text manuscripts were reviewed by two reviewers (KN, BM) and one adjudicator (OA) per an established set of a priori criteria (Supplemental Table 1). Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) reporting guidelines were followed. ²

Primary and secondary outcomes include race and ethnicity reporting in PAD RCT publications, respectively. Factors of interest were determined a priori by coauthor consensus and included publication before or after ClinicalTrials.gov data mandatory demographic reporting (which was mandated in 2017), primary funding source (industry, US government, academic, other), disease severity (all PAD, claudication, chronic limb-threatening ischemia [CLTI], no option CLTI, other), primary study type (pharmacologic intervention, device-related, basic science, exercise/rehabilitation-related, provider behavioral intervention/care delivery, other), intervention model (parallel assignment, crossover assignment, single group assignment, factorial assignment), study phase, enrollment volume (< or ≥ 500 patients), and number of sites. Race and ethnicity categories were defined using the Office of Management and Budget (OMB) 1997 Standards. ³ We captured race categorization schemes (the sex, race, or ethnicity categories that authors chose) in each RCT included in this analysis.

Descriptive statistics were compared and RCT ‘participation-to-prevalence’ ratios (PPR), or a comparison of a study’s demographic participation to the specific demographic’s prevalence rate in the US population, were calculated. ⁴ The association between covariates and outcomes of interest were examined using logistic regression.

Among 174 RCTs (91,677 enrollees), 50.0% (n = 87) were primarily industry-funded and 26.4% (n = 46) were US Government-funded. The majority of included RCTs enrolled fewer than 500 patients (83.9%, n = 146) and 8.6% (n = 15) of included RCTs were focused on a CLTI study population. Over half (51.1%, n = 89) of the RCTs tested pharmacologic intervention and most studies employed a parallel assignment design (90.8%, n = 158). One-third of all included RCTs were conducted at a single site (35.1%, n = 61). Sex was reported in 98.2% of publications and 97.4% of ClinicalTrials.gov site records.

Race was reported in 64.4% of publications and 69.8% of ClinicalTrials.gov sites. Race reporting remained consistently below 80% between 2000 and 2022 (Figure 1) and was comparable pre- and post-2017 (66.1% vs 61.3%, p = NS). Race was described with 24 categorization schemes (Supplemental Table 2). Fifteen studies (9%) allowed participants to self-report their racial identity. The median PPR for Black participants was 0.15 (IQR 0.08, 0.26). No factors of interest were associated with race reporting (Supplemental Table 3).

Figure 1.

Race and ethnicity reporting by year among peripheral artery disease randomized controlled trials, 2000–2022.

Ethnicity was reported in 27.6% of publications and 26.7% of ClinicalTrials.gov sites. Ethnicity reporting was comparable pre- and post-2017 (26.8% vs 29.0%, p = NS) and was consistently low throughout the study period (Figure 1). Five studies (3%) self-reported ethnicity. The median PPR for Hispanic participants is 0.04 (IQR 0.02, 0.09). No factors of interest were associated with ethnicity reporting.

There is clearly value in RCTs enrolling patients who are representative of the disease of interest. However, equitable representation cannot be measured without first improving the rate of race and ethnicity reporting in the surgical literature as well as in RCTs. ⁵ Despite widely acknowledged disparities in PAD care and outcomes, our study reveals that demographic reporting of both race and ethnicity requires significant improvement.

One explanation could be ignorance of how innocuous RCT design and conduct details can present structural barriers to participation by racial and ethnic minority patients. Even unassuming eligibility criteria can present barriers to RCT participation for subjects from underrepresented communities. ⁶ Funding agencies have asked for more representative RCTs; however, one cannot determine the representativeness of an RCT if few of these studies actually report race or ethnicity. ⁷

Another explanation could be a lack of knowledge and comfort in the subject area. Improving the diversity of primary investigators who may be more comfortable or sensitive to this issue is one small step in the right direction. ⁸ Sweeping efforts to clarify language by funding agencies and journals would have a high impact. The 1997 OMB Standard promoted uniform race/ethnicity reporting and mandated federal programs to adopt these standards. The intent was to provide guidance for inclusive language that decreases bias and protects structurally marginalized populations. ³ Some journals have published or refer to guidelines to standardize race/ethnicity reporting. ⁹ Defined reporting standards from all journals and funding agencies would help investigators become more comfortable with using clear, consistent, and inclusive language while reporting their work. Additional recommendations are provided in Supplemental Table 4.

We examined PAD RCTs with US sites limiting generalizability to international RCTs. We did not qualitatively assess barriers/facilitators to demographic reporting, potentially a future avenue of inquiry.

In conclusion, accurate and consistent reporting of self-identified sociodemographic data, including utilizing proper inclusive terminology, is essential for achieving equity and moving towards disparity elimination in PAD care. This becomes particularly important in the generalizability of pragmatic trials, real-world application of tested interventions, and the dissemination of results to various communities.

Supplemental Material

sj-pdf-1-vmj-10.1177_1358863X251322177 – Supplemental material for Reporting of race, ethnicity, and gender in lower-extremity peripheral artery disease randomized controlled trials

Supplemental material, sj-pdf-1-vmj-10.1177_1358863X251322177 for Reporting of race, ethnicity, and gender in lower-extremity peripheral artery disease randomized controlled trials by Kasthuri Nair, Brandi M Mize, Mia S White, Tabia Henry Akintobi, Rachel E Patzer, Shipra Arya and Olamide Alabi in Vascular Medicine