Racial and Ethnic Inequities in US Oncology Clinical Trial Participation From 2017 to 2022
By Harlan Pittell, PhD; Gregory S. Calip, PharmD, MPH, PhD,; Amy Pierre, MSN, ANP-BC1,3; et al
On July 21, 2023
Key Points
Question How did racial and ethnic inequities in oncology clinical trial participation persist between 2017 and 2022?
Findings In this cohort study of 50 411 participants and 5 cancer types, Latinx and non-Latinx Black patients had lower participation in oncology trials than non-Latinx White patients. Furthermore, during the COVID-19 pandemic (2020-2022), there were greater inequities in participation between Black and White patients compared with before the pandemic (2017-2019).
Meaning These findings reinforce the need for guidance from regulatory agencies to improve enrollment of participants from historically excluded racial and ethnic populations in oncology trials.
Abstract
Importance There is increasing recognition from regulatory agencies that racial and ethnic representation in clinical trials is inadequate and linked to health inequities. The extent of racial inequities in clinical trial participation is unclear because prior studies have synthesized enrollment data from published trials, which often do not report participant race and ethnicity.
Objective To evaluate racial and ethnic inequities in oncology clinical trial participation in a contemporary cohort of patients with cancer before and during the COVID-19 pandemic.
Design, Setting, and Participants This cohort study used a nationwide electronic health record–derived deidentified database, which includes data for approximately 280 US cancer clinics (approximately 800 sites of care). The study included Latinx, non-Latinx Black (hereinafter, Black), and non-Latinx White (reference; hereinafter, White) patients aged 18 years or older who had been diagnosed with advanced non–small cell lung cancer, metastatic colorectal cancer, metastatic breast cancer, multiple myeloma, or metastatic pancreatic cancer between January 1, 2017, and June 30, 2022 (follow-up through December 31, 2022). Data analysis was performed between August 1, 2022, and February 7, 2023.
Exposures Electronic health record–documented race and ethnicity.
Main Outcomes and Measures The main outcome was oncology trial participation (ie, receipt of a clinical study drug). Stratified cause-specific hazard models were used to estimate adjusted hazard ratios (HRs) and 95% CIs for likelihood of participation. Participation was assessed overall, by cancer type, and by period of diagnosis (2017-2019 vs 2020-2022).
Results Of the 50 411 patients in this study, 28 878 (57.3%) were women and 21 533 (42.7%) were men. Black and Latinx patients were younger than White patients, with a median age of 65 (IQR, 57-72), 64 (IQR, 54-73), and 68 (IQR, 60-76) years, respectively. Oncology trial participation was lower among Black patients (307 of 6912 [4.4%]) and Latinx patients (166 of 3973 [4.2%]) relative to White patients (2858 of 39 526 [7.2%]) over the entire study period. Inequities in participation were observed across the 5 cancer types studied, with notably large inequities observed among Black patients (HR, 0.54 [95% CI, 0.36-0.81]) and Latinx patients (HR, 0.46 [95% CI, 0.27-0.77]) with metastatic pancreatic cancer. Moreover, inequities between Black and White patients in terms of participation widened among those diagnosed in the COVID-19 era (2020-2022: HR, 0.49 [95% CI, 0.40-0.60] vs 1.00 [95% CI, 0.93-1.09]) relative to those diagnosed before the pandemic (2017-2019: HR, 0.61 [95% CI, 0.53-0.70] vs 1 [reference]).
Conclusions and Relevance The findings of this cohort study suggest that oncology trial participation was lower among Black and Latinx patients relative to White patients across all 5 cancer types examined. These findings, including potentially widening inequities in the COVID-19 era, support the need for regulatory guidance to improve enrollment of participants from historically excluded racial and ethnic populations in clinical trials.
Introduction
US Food and Drug Administration (FDA) draft guidance from April 2022 calls for the improvement of clinical trial enrollment of participants from historically excluded racial and ethnic populations, and key organizations have made similar calls for more diverse representation among oncology trial participants.1,2 Earlier research offered evidence of racial and ethnic inequities in clinical trial participation, but such studies analyzed enrollment data from published trials, which often lack information on participant race and ethnicity.3 Given evidence of worsening racial and ethnic inequities in access to health care and health outcomes, especially since the start of the COVID-19 pandemic, this study aimed to assess the extent of racial and ethnic inequities in oncology clinical trial participation and how they may have changed in recent years.4
This retrospective cohort study used the nationwide electronic health record (EHR)–derived deidentified database from Flatiron Health Inc. This database is longitudinal and comprises deidentified patient-level structured and unstructured data curated via technology-enabled abstraction.5,6 During the study period, the deidentified data originated from approximately 280 US cancer clinics (approximately 800 sites of care). This study included 50 411 patients, with diagnoses of advanced non–small cell lung cancer (n = 20 961), metastatic colorectal cancer (n = 10 596), metastatic breast cancer (n = 9912), multiple myeloma (n = 4348), or metastatic pancreatic cancer (n = 4594). These are among the most common cancers in the US and cancers with known inequities in incidence and survival.7 The cohort was limited to patients aged 18 years or older with a confirmed diagnosis between January 1, 2017, and June 30, 2022 (follow-up through December 31, 2022). We excluded patients with potentially incomplete therapy data (ie, discrepancies between unstructured and structured treatment records) and patients without a documented race and ethnicity (25.3% of otherwise eligible patients). The attrition table is presented in the eTable in Supplement 1.
Institutional review board approval of the study was obtained by Flatiron Health and included a waiver of informed consent because deidentified data were used. The study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline.
We examined oncology trial participation by EHR-documented race and ethnicity and focused on Latinx, non-Latinx Black (hereinafter, Black), and non-Latinx White (hereinafter, White) patients. We excluded non-Latinx Asian patients from the analysis because the relatively small sample size may offer imprecise estimates of participation rates. We also excluded patients with a documented race of “other” due to the unclear interpretation of such results (ie, multiple racial groups are masked as “other” for privacy reasons; additional details are provided in the eTable in Supplement 1). Race and ethnicity values vary by practices at individual clinical sites and the limitations of the EHR software used. Although they are likely to reflect self-reported assignments by patients, it is unknown whether these values are self-reported or clinician or staff reported. Trial participation was defined as EHR-documented receipt of a clinical study drug during an oncologist-defined, rule-based line of therapy. We reported participation rates and used Cox proportional hazard models to estimate hazard ratios (HRs) of participation, which accounted for differences in follow-up time after adjusting for clinical factors such as age, gender (detailed in Table 1), Eastern Cooperative Oncology Group performance status, year of diagnosis, and cancer type. In an exploratory analysis, we estimated models stratified by cancer and year of diagnosis, with the latter comparing trial participation between patients diagnosed in the pre–COVID-19 and COVID-19 eras (2017-2019 vs 2020-2022). We conducted all statistical analyses using R, version 4.0.2, from August 1, 2022, to February 7, 2023. Estimates are presented with 95% CIs.
Of the 50 411 patients in this study, 28 878 (57.3%) were women and 21 533 (42.7%) were men. Table 1 summarizes the cohort characteristics. With regard to race and ethnicity, 6912 patients (13.7%) were Black, 3973 (7.9%) were Latinx, and 39 526 (78.4%) were White. Black and Latinx patients were younger than White patients, with a median age of 65 (IQR, 57-72), 64 (IQR, 54-73), and 68 (IQR, 60-76) years, respectively. Metastatic colorectal cancer was the most common type among Latinx patients, whereas advanced non–small cell lung cancer was the most common type among Black and White patients. Notably, 938 of 4348 patients (21.6%) with multiple myeloma were Black, denoting a greater prevalence of this cancer among Black patients in the cohort.
Table 2 lists oncology trial participation by race and ethnicity. Overall participation was 4.4% (307 of 6912) among Black patients, 4.2% (166 of 3973) among Latinx patients, and 7.2% (2858 of 39 526) among White patients. Absolute differences in participation were greatest among patients with metastatic pancreatic cancer (25 of 487 Black patients [5.1%] and 16 of 332 Latinx patients [4.8%] vs 370 of 3775 White patients [9.8%]). The lowest absolute rates of participation were observed among Latinx patients, with rates of 3.5% (43 of 1222) for metastatic colorectal cancer trials and 3.2% (34 of 1047) for metastatic breast cancer trials; this contrasted with metastatic colorectal cancer as the most common cancer among Latinx patients in the cohort. In adjusted models, Black and Latinx patients were less likely to participate in clinical trials than White patients in all disease states with all 95% CIs below 1 (with the exception of Latinx patients with multiple myeloma).
Table 3 lists oncology trial participation in our exploratory analysis by year of diagnosis, comparing participation between patients diagnosed in the pre–COVID-19 and COVID-19 eras. Overall participation was lower among Black and Latinx patients in both periods relative to White patients. Specifically, the likelihood of trial participation among Black patients in the COVID-19 era was lower relative to the pre–COVID-19 era (2017-2019: HR, 0.61 [95% CI, 0.53-0.70] vs 2020-2022: HR, 0.49 [95% CI, 0.40-0.60]), whereas the likelihood of participation among White patients did not meaningfully change over time (2017-2019: HR, 1 [reference] vs 2020-2022: HR, 1.00 [95% CI, 0.93-1.09]).
Discussion
These findings suggest that racial and ethnic inequities in oncology clinical trial participation persisted from 2017 to 2022 and may have widened since the start of the COVID-19 pandemic. We observed lower participation among Black and Latinx patients (relative to White patients) across the cancers studied, which underscores the importance of the recent FDA draft guidance for diversity plans to improve enrollment of participants from underrepresented racial and ethnic populations in clinical trials.1,2 Many published clinical trials have not reported participant race and ethnicity.3 Therefore, EHR-derived data can be used to elucidate persistent inequities in cancer trial participation.
The historical underrepresentation of Black and Latinx patients in oncology clinical trials is in stark contrast with the higher cancer incidence and known inequities in cancer outcomes, particularly in pancreatic cancer and multiple myeloma, which can exacerbate cancer inequities through the exclusion of such groups from investigational drug access.8,9 Broadly, the COVID-19 pandemic has deepened the extent of health inequities between racial and ethnic groups, and our study offers evidence that inequities between Black and White patients with cancer may have worsened following the pandemic.4 Additional work is needed to confirm whether these trends have persisted since the start of the pandemic. Beyond the FDA draft guidance intended for industry sponsors of clinical trials, there is further opportunity to potentially address these and other trial inequities through advances in technology. Digital tools that automate patient screening and trial matching have the ability to more easily evaluate larger patient populations for eligibility in an unbiased manner, potentially reducing the burden on research teams while also expanding enrollment opportunities.10 Similarly, equitable decentralization of clinical trials may help to reduce financial or transportation barriers to clinical trial participation, as could broadening eligibility criteria to improve trial access for all—although these hypotheses are yet to be proven.11,12 The outcomes of such strategies on narrowing clinical trial inequities should be a focus of future research.13 By bridging the divide between routine care and clinical research, integrated evidence has the opportunity to make clinical trial participation more available to patients, increasing the potential depth of data, representativeness, and efficiency of clinical research.14
Our findings on greater inequities in cancer clinical trial participation among Black patients during the COVID-19 pandemic add to the growing evidence on pandemic-related decreases in participation in clinical trials. In an interrupted time-series analysis of trials conducted between 2016 and 2021 in the SWOG Cancer Research Network,15 the total number of patients enrolled in clinical trials decreased during the first year of the pandemic. However, enrollment reductions were primarily to cancer control and prevention trials, whereas evidence of enrollment reductions to treatment trials was lacking. Distinct from the SWOG analysis, an exploratory objective of our study was to examine clinical trial participation in the COVID-19 era; although there was little evidence of statistical interaction (post hoc analysis) between year of diagnosis and racial and ethnic group, we estimated race- and time period–specific estimates of clinical trial participation. Also, we defined participation differently than the analysis of SWOG trials, which may have implications for how we interpret and compare results from the postpandemic time period, such as the different settings for industry-sponsored trials and our inclusion of Latinx patients in our analysis. These differences highlight the importance of carefully considering the context and methodology of each study when interpreting and comparing rates of cancer clinical trial participation.
Historically marginalized or minoritized groups, including Black and Latinx patients, women, and patients aged 65 years or older, are consistently underrepresented in clinical trials, considering their disproportionate burden of cancer incidence.16–18 In a study of landmark trials that led to FDA oncology drug approvals, Loree et al19 found that compared with rates of US cancer incidence, Black and Latinx patients with cancer were underrepresented in oncology trials (22% and 44% of expected, respectively) compared with White patients with cancer (98% of expected). This lack of representation of patients with cancer from historically marginalized groups may be explained, in part, by the contributions (or lack thereof) of industry-sponsored trials relative to federally sponsored trials. In an analysis of 358 oncology clinical trials of 93 825 patients across 15 cancer types between 2008 and 2018, the overall proportion of Black patients represented was 2.9% for industry-sponsored trials and 9.0% for National Cancer Institute–sponsored trials.20
The barriers to recruitment and participation of historically marginalized, minoritized, and excluded patients in cancer clinical trials are considerable. In a systematic review of 65 oncology therapeutic and cancer prevention trials, multiple barriers to the awareness (eg, education on clinical trials), opportunity (eg, eligibility, lack of communication), and ultimately participation (eg, mistrust, costs, transportation) were consistently reported.21 The effects of language barriers and lack of culturally relevant education about clinical trials were also frequently reported. Although there are often multiple proposed ways to improve opportunities for enrollment and participation, including using less rigid study designs and developing systems that facilitate the participation of health care professionals, the steps taken by investigators and drug sponsors to increase enrollment too often will prove inadequate in the absence of an organizational and technology infrastructure to support contemporary trials.22,23 Opportunities exist for digital tools and empirical data from clinical practice to potentially reduce the diversity gap between the underlying patient population and those that are eligible and subsequently enrolled in clinical trials. In a study using administrative claims data, Riner et al24 determined how inclusion and exclusion criteria can result in disparities in pancreatic ductal adenocarcinoma clinical trial eligibility. These empirical data demonstrated that revising certain eligibility criteria (eg, comorbidities and measures of kidney function) resulted in similar rates of eligibility between racial groups.
This study has some limitations. First, our use of clinical study drug administration as a surrogate for oncology trial participation likely underrepresents clinical trial accrual, since it may not capture participation in surgery, radiation, observational, or nontreatment interventional trials and is reliant on EHR documentation. Second, we were unable to examine inequities in clinical trial participation among other historically excluded groups due to data deidentification processes and small cohort sizes; further work is needed to investigate the potential for such inequities. Third, we excluded patients without a documented race or ethnicity from our analysis; thus, our results may be an underestimation of the inequities in participation among Black and Latinx patients. Finally, the generalizability of these results to the broader US oncology population may be limited, as we focused on a select group of cancers and the composition of the Flatiron Health population can differ in certain aspects from the US oncology population (eg, distribution of race and ethnicity not similar to that observed in other national databases).5
Conclusions
In this cohort study of racial and ethnic inequities in oncology clinical trial participation, we observed that Latinx and Black patients had lower participation in oncology trials than White patients across the 5 cancer types studied over the 2017 to 2022 study period. Our findings support the FDA recommendation for diversity plans that aim to improve enrollment of participants from historically excluded racial and ethnic populations in clinical trials, particularly given evidence of worsening inequities observed during the COVID-19 era.
This piece was republished from the Jama Network.