Clinical Trial Diversity, Equity & Inclusion: Roadmap of the Cardiothoracic Surgical Trials Network

Abstract

BACKGROUND:

There is a recognized lack of diversity among patients enrolled in cardiovascular interventional and surgical trials. Diverse patient representation in clinical trials is necessary to enhance generalizability of findings, which may lead to better outcomes across broader populations. The Cardiothoracic Surgical Trials Network (CTSN) recently developed a plan-of-action to increase diversity among participating investigators and trial participants -- the focus of this paper.

METHODS:

Review of literature and enrollment data from CTSN trials.

RESULTS:

More than a dozen major clinical trials were completed by CTSN (2008–2022), enrolling over 4,000 patients, of whom 30% were women 11% non-White and 5.6% were Hispanic. The Network also completed trials of hospitalized COVID-19 patients, wherein enrollment was more diverse, with 42% women and 58% either Asian, Black, Hispanic or from another underrepresented racial group. The discrepancy in diversity of enrollment between cardiac surgery trials and COVID trials highlights the need for a more comprehensive understanding of: 1) the prevalence of underlying disease requiring cardiac interventions across broad populations, 2) differences in access to care and referral for cardiac surgery, and 3) barriers to enrollment in cardiac surgery trials.

CONCLUSIONS:

Committed to diversity, CTSN’s multi-faceted action plan includes developing sitespecific enrollment targets, collecting social determinants of health data, understanding reasons for nonparticipation, recruiting sites that serve diverse populations, emphasizing greater diversity among clinical trial teams, and implicit bias training. The CTSN will prospectively assess how these interventions influence enrollment as we work to ensure trial participants are more representative of the communities we serve.

Graphical Abstract

It has become increasingly clear that cardiovascular trials in general, and cardiothoracic surgical trials in particular, are characterized by underrepresentation of women and under-represented racial and ethnic groups (UREGs).¹ A review of 35 pivotal FDA trials evaluating novel cardiometabolic drugs internationally, for example, showed that only 4% of randomized patients were Black, 12% Asian and 11% Hispanic/Latino, compared with 81% non-Hispanic White adults, with no significant change in representation over a 10-year period.² Similar racial and ethnic distributions have been observed in other studies.^3,4. These observations need to be placed in context of the demographic make-up of the United States, where nearly half of the population is non-White, with at least ~33% of the population identifying as non-Hispanic Black, Hispanic, or American Indian or Alaska Native, and with almost an equal split between men and women.^3,4 Addressing the appropriate distribution of specific groups in clinical trials requires insight into barriers to recruitment and enrollment, sociocultural factors and system-related biases in the provision of care as well as the prevalence of cardiovascular diseases by race, ethnicity, gender and other factors. In some of these areas, the necessary population-based studies are lacking.

Clinical trials are the foundation of a rigorous approach to establishing what is safe and effective in patient care. The lack of diverse representation in cardiovascular clinical trials means that trial findings, many of which are life-saving and/or quality of life enhancing, may not be generalizable to broad patient populations. The importance of enrolling diverse populations is further underscored by the differential outcomes of treatment that are often observed in underrepresented groups. As such, it is crucial to increase the diversity of clinical trial enrollment.

This paper focuses on strategies for improving diversity in trials of cardiac surgical interventions, such as coronary artery bypass surgery (CABG). Through the lens of the Cardiothoracic Surgical Trials Network (CTSN), we review the barriers to enrollment of diverse patient populations, which shaped our plan of action. Although diversity encompasses many constructs such as age, sexual orientation, disability, area deprivation index, income, educational level and other social determinants of health, here we limit the discussion to sex, race and ethnicity, aspects for which CTSN data are available. In this paper, we review the representativeness of enrollment in cardiac surgery trials, including the experience of CTSN, and offer a comparison with enrollment demographics in CTSN trials outside of the cardiac surgery domain (COVID-19). We then outline a roadmap for enhancing diversity in our Network, which we believe has broader applicability for investigators designing and conducting cardiovascular clinical trials.

DIVERSITY AND INCLUSION IN CARDIAC SURGERY TRIALS

There is accumulating evidence that cardiac surgical trials are insufficiently diverse. A study including 178 cardiac surgery trials, conducted between 1998 and 2017 and identified through Clinicaltrials.gov, found that 28.3% of nearly 240,000 participants were women. Only 43% of the trials reported race; with 4% of enrollees identified as African American and 10.4% as Asian. Ethnicity was reported only in 29% of trials, with 11.2% of these enrollees being Hispanic. This study also found that in CABG trials the participation-to-prevalence ratio (PPR) for female participants was 0.64, while for African American participants the PPR was 0.28, for Hispanic participants 0.72 and, in contrast, 3.20 for Asian participants.⁵

Since its inception, CTSN has completed nine major cardiovascular randomized trials. Concurrently, during the peak of the COVID-19 pandemic, the CTSN conducted a COVID-related ARDS trial and participated as an international coordinating center for the ACTIV-3 COVID platform, which included multiple trials evaluating novel monoclonal antibodies and anti-viral agents in hospitalized patients with COVID-19 (Supplemental Table 1). When focusing on cardiac surgery trials, 30% of patients were women, 11% were non-White, and, in terms of ethnicity, 5.6% were Hispanic. However, across COVID trials, which combined race and ethnicity in terms of data collection, 42% were women and 58% were either Asian, Black, Hispanic or from another underrepresented racial group. Figure 1a, b, and c depicts enrollment stratified by sex, race, and ethnicity for each trial. Across the spectrum of CTSN trials, which were conducted in the US and Canada, as well as Germany for the tricuspid trial, considerable variation was observed in the demographic profiles of randomized patients. The reasons for these variations, as discussed below, are multi-factorial.

Figure 1.

Cardiothoracic Surgical Trials Network (CTSN) enrollment by a) sex b) race, and c) ethnicity*. Abbrv: MVS- Surgical Ablation of Atrial Fibrillation during Mitral Valve Surgery; MMR- Surgical treatment of moderate ischemic mitral regurgitation; SMR- Mitral-Valve Repair versus Replacement for Severe Ischemic Mitral Regurgitation; POAF- Rate Control versus Rhythm Control for Atrial Fibrillation after Cardiac Surgery; NP- Effect of Cerebral Embolic Protection Devices on CNS Infarction in Surgical Aortic Valve Replacement: A Randomized Clinical Trial; TR- Concomitant Tricuspid Repair in Patients with Degenerative Mitral Regurgitation: VEST- External Support for Saphenous Vein Grafts in Coronary Artery Bypass Surgery: A Randomized Clinical Trial; MPC-LVADII- Intramyocardial Injection of Mesenchymal Precursor Cells and Successful Temporary Weaning From Left Ventricular Assist Device Support in Patients With Advanced Heart Failure; MSC-ARDS- A Randomized Trial of Mesenchymal Stromal Cells for Moderate to Severe Acute Respiratory Distress Syndrome from COVID-19; TESICO-Intravenous aviptadil and remdesivir for treatment of COVID-19-associated hypoxemic respiratory failure (ACTIV 3 Network); As part of the ACTIV 3 Network, A master platform trial (TICO) was conducted, evaluating several alternative monoclonal antibodies and antiviral treatments for hospitalized COVID-19. *In the cardiac surgery trials, we collected data on race and ethnicity separately. In the ACTIV 3 trials, data were collected on sex and the combined category of race and ethnicity (Asian, Black, White and other).

Differences in Prevalence and Presentation of Disease among Groups

Clearly, some of the variation seen in CTSN trials was driven by differences in disease prevalence across different race-ethnic and socio-demographic groups. The trials of hospitalized COVID patients had broad representation of UREGs, owing, in part, to the disproportionate burden of the disease and higher incidence of adverse outcomes in such individuals.^6,7 Focusing on the CTSN cardiac surgery trials, there were major sex-based differences in enrollment across trials. In a trial evaluating surgical ablation in patients with persistent or long-standing persistent atrial fibrillation requiring mitral-valve surgery, 46% were female compared to just 11% in the trial evaluating intramyocardial injections of allogeneic mesenchymal precursor cells during LVAD implantation.⁸ The latter is consistent with data from the InterMACS registry^9–11, indicating that only 25% of LVAD recipients are women, reflecting the clinical fact that women tend to present with heart failure with preserved ejection fraction more frequently while dilated cardiomyopathy with reduced ejection fraction is more commonly encountered among men.¹² In the same CTSN LVAD trial, the percentage of Black patients randomized (14.5%) was greater than in other CTSN trials, which is consistent with the comparatively higher prevalence of dilated cardiomyopathy among Black compared to White patients.^13–15 Disease prevalence should be taken into account in setting enrollment targets by demographic group.

The Effect of Disparities in the Utilization of Cardiac Surgical Procedures

Notable differences exist in the utilization of cardiac surgical procedures by race, ethnicity and gender that may have a major effect on “downstream” enrollment in surgery cardiac clinical trials. As early as the 1990s, landmark studies identified disparities in the utilization of coronary revascularization procedures in women and racial minorities, with higher risk-adjusted mortality in these groups.^16,17 Whereas some studies showed a narrowing of racial and gender differences in utilization rates over time, mortality differences continue to persist.^18,19 In an analysis of the Society of Thoracic Surgery (STS) Adult Cardiac Surgery Database, which captures 95% of cardiac surgery in the US, women comprised 25% of all patients undergoing CABG between 2011 and 2018, with the majority (84.9%) of patients being white and 7.7% Black; both women and Black patients had higher risks of 30-day mortality compared with White men.²⁰ In the area of valvular heart disease, surgical and transcatheter procedures have been underutilized in racial and ethnic minorities.²¹ In an STS analysis of all mitral valve surgery performed between 2014 and 2018, 45.2% of the cohort of 165,405 patients were women, 77.2% were white, 9.9% Black, and 5.9% Hispanic²², relative to 12.6% Blacks and 18.9% Hispanics in the overall 2018 US population.²³ There are few population-based studies that stratify outcomes by race, ethnicity and gender after mitral valve surgery; however, worse outcomes have been reported for UREGs compared to other valvular procedures, such as surgical aortic valve replacement (SAVR) and transcatheter aortic valve replacement (TAVR).²¹

Such differences are partially driven by variations in rates and timing of referrals, whereby women and individuals from UREGs are less likely to be referred for cardiovascular evaluations and interventions such as coronary angiography²⁴, surgery²⁵, and cardiac rehabilitation.²⁶ Specifically, women have been observed to present at later stages of disease compared to men, potentially accounting for reports of worse outcomes following cardiac surgical procedures.^20,27–29 In the CTSN trial evaluating surgical approaches for severe ischemic MR, for example, women were found to have different echocardiographic features, such as smaller LV volumes and effective regurgitant orifice areas (EROA) but greater EROA / LV end-diastolic volume ratios, and experienced higher mortality and worse quality of life (QOL) after mitral valve surgery compared with men.³⁰ In addition to differences in the timing of referral, UREGs are less likely to be seen at high-volume centers or operated on by high-volume surgeons.^31–33 As lower-volume centers may have a less developed clinical research infrastructure and, as such, may be less likely to participate in large multi-center surgical trials and gain early access to newer technologies, a deleterious effect on trial diversity will likely ensue. This is very different situation than what was experienced with COVID-19 research, were there were few restrictions to hospitalization and treatment for COVID-19 of any patient group and greater efforts were made to include study sites where COVID was prevalent, which is reflected in the higher numbers of women and UREGs randomized in the CTSN COVID trials.

Socio-Cultural and Socio-Economic Factors

The reasons for under-enrollment of women and UREGs are complex and multi-faceted (at the patient, clinician, health center, hospital, health system and community-level), and partially rooted in systemic racism (Figure 2).³⁴ Factors cited as barriers include patients’ mistrust of the healthcare system, practitioners’ potential implicit bias or unwillingness to enroll or refer patients into new studies, practitioner inertia, trial sponsor and investigator uncertainty regarding best practices to incentivize enrollment by different groups, and under-representation of women and UREGs among clinical site PIs and investigators. ^5,35

Figure 2.

Potential barriers to optimal underrepresented racial and ethnic groups (UREG) enrollment in clinical trials

Patient trust has been widely considered as a factor influencing a patient’s willingness to participate in research,³⁶ particularly amongst those who have been historically marginalized by healthcare. This mistrust can make it difficult to engage specific communities in the research process, especially in clinical trials that have an invasive arm. In a study by Peterson et al., conducted to ascertain reasons as to why patients would “probably or definitely not consider trial enrollment,” the vast majority reported a dislike of a “randomization process.”³⁷ In another study of 13 different primary care and cardiology clinics, higher levels of distrust were observed among Black patients relative to White patients even after accounting for differences in socioeconomic status and access to care.³⁸ Distrust among Black patients in particular is, in part, related to the legacy of the U.S. Public Health Service Tuskegee Syphilis Study as well as other historical research related transgressions.³⁹ Interestingly, however, once exposed to clinical trials, racial differences in trust may be overcome.⁴⁰

Many minority communities have unique cultural beliefs and practices as well as language barriers that may influence their willingness to participate in clinical trials. Without an understanding of these cultural factors, researchers may inadvertently exclude certain patient groups or fail to effectively engage them in the research enterprise. Similarly, patients may be less inclined to hear about clinical trial opportunities if they are not able to relate to or speak the same language as research coordinators and investigators, which can significantly impact patient-investigator communication, leading to possible mistrust of scientific investigators and of academic institutions. Socioeconomically disadvantaged communities may be disproportionately affected by low health literacy and lack of access to care. Time commitment in the setting of job inflexibility, caregiver issues, loss of income, or transportation can also be significant barriers to clinical trial recruitment, participation and follow-up.

Clinician or investigator unconscious bias may result in the perception that UREGs are less likely to be eligible for enrollment in trials, due to possible non-compliance or time constraint. There may also be the perception that UREGs are less interested in clinical trials. A similar concern relates to the frequent underrepresentation of elderly patients across studies. Many randomized trials by design exclude patients above a certain age cut-off, even though elderly patients comprise the fastest growing segment of the population, are at higher absolute risk of adverse outcomes, may benefit from new treatments, and may represent the population for which new therapies may most aptly pertain. Language is another major barrier to enrollment pertains to non-English-speaking patients, who often require interpreter services or bilingual research staff. Many Hispanic patients have a Spanish-language preference. Clearly, having diverse clinical site investigative teams can help address these issues and favorably affect diversity of enrollment, as several studies have underscored.^41,42 In 2023, CTSN conducted an anonymous survey of its investigative teams to understand better the demographic make-up of those personnel participating in its international Network. Based on feedback from close to 250 respondents, 58% were female, 7.3% were Hispanic, 9% were Black, 17% were Asian, and 66% were White, and 41% were below 45 years of age; these data are informing efforts to increase diversity across the Network’s proceedings.

Over the past three decades, sponsors have started focusing more heavily on the inclusion of UREGs in clinical trials. The National Institutes of Health (NIH), and its family of institutes, have historically set targets for enrollment of women and UREGs since the NIH Revitalization Act mandated minority inclusion in 1993 with recommended benchmarks.⁴³ More recently, industry has started to invest in programs to recruit diverse patient populations, stimulated in part by FDA initiatives, such as the requirement to submit a race and ethnicity diversity plan early in development of new technology.⁴⁴ At the same time, the importance of reporting representativeness of study populations was recognized by major medical journals when they announced that clinical trial papers should provide background information on the disease, problem or condition studied, and the representativeness of the randomized sample. The efforts by public and private sponsors, as well as by investigators, to improve diversity have yet to be fully realized and may not yet be reflected in actual trial data. A recent report by the National Academy of Medicine argues for increased investment in this area by all stakeholders, including sponsors, regulatory agencies, and all parties involved in conducting clinical trials.⁴⁵

CTSN ROADMAP FOR IMPROVING TRIAL DIVERSITY

The Network recognizes its responsibility to enroll diverse populations in its clinical trials, and upon reviewing its track-record in the cardiac surgery domain, believes that a more concerted effort is needed to improve trial enrollment. Given the complex nature of barriers to enrollment of representative groups in clinical trials, especially in an international Network, CTSN developed a multi-pronged action plan for improvement.

Creation of a DEI Taskforce and DEI Statement

A first step was to create an enduring structure within the Network to address these issues, with recruitment of a Vice Chair for Diversity and appointment of a lead investigator in the coordinating center. The newly formed Diversity, Equity and Inclusion (DEI) Taskforce, includes surgeons, cardiologists, coordinators, and clinical trialists, and meets regularly to propose and implement new diversity initiatives and monitor impact on the Network’s performance. The Task Force has created the following DEI statement, since approved and disseminated by the Network:

The CTSN recognizes the benefit of and its responsibility to achieving Diversity, Equity, and Inclusion (DEI) in its scientific inquiry:

• The CTSN is committed to health equity in carrying out its mission to conduct research that improves the health status of all adult patients with cardiovascular diseases

• The CTSN is committed to inclusion by enrolling and retaining diverse & representative patient populations in its clinical trials

• The CTSN is committed to diversity among the investigators and coordinators participating in the Network

In the context of clinical trials, equity is defined as representation of particular demographic sub-groups proportional to the disease prevalence across a given population. Unfortunately, for many disease categories, such as valvular heart disease, there are few U.S. population-based epidemiological studies.^46–49 Moreover, there are even fewer studies that characterize the epidemiology of valvular heart disease among non-white U.S. populations (Hispanics, Blacks, Asians, Native Americans).^50,51 Recognizing this knowledge gap, NHLBI has expanded its investment in several large population-based cohort studies, such as the RURAL cohort study that ascertains the prevalence of heart and lung disorders in rural counties in the South-Eastern US.⁵² In the absence of such rigorous population-based studies, establishing enrollment benchmarks that reflect the diversity of disease burden is more challenging. However, it would take years to conduct much needed population-based studies, and, yet there is an immediate need to ensure that our clinical trials are generalizable to broader populations. Thus, we should establish enrollment benchmarks based on the best available evidence, recognizing that such evidence is less than perfect, such as national treatment utilization data, when available, or the demographics of the population at large.

Design Considerations to Enhance Diversity

Trial design characteristics can assist in enhancing the diversity of enrolled patients. Typically, early phase trials or mechanistic trials have more limited sample sizes, and despite adhering to enrollment targets, sufficient representation of demographic sub-groups is usually unachievable. By comparison, larger-scale confirmatory trials or comparative effectiveness trials should be able to enroll sufficient proportions of women and racial and ethnic minorities to allow for meaningful sub-group analyses. An important consideration in ensuring this diversity is to make sure that eligibility criteria do not systematically exclude specific demographic groups. In recent years, trials that focus exclusively on specific target populations, such as women, minorities or particular age groups, have been funded to overcome the failure of diversity in large-scale trials. A case in point is the recent funding by the Patient Centered Outcomes Research Institute (PCORI) of two companion trials comparing CABG to PCI in women, and in African American and Hispanic patients (RECHARGE program).⁵³

Site-Specific Enrollment Targets and Site Selection Issues

Historically, the Network had established enrollment targets by sex, race and ethnicity for each trial. Ideally, prevalence data from population-based studies should be the basis for these targets. If epidemiological studies are not available, the study population should reflect the composition of the overall U.S. population (51% women, 30% minorities). Data on the utilization of particular procedures from national registries, such as the STS Adult Cardiac Surgery Database, are considered as well with the caveat that they may be clouded by the disparities in utilization that have been already described. Recognizing that clinical sites may serve populations with different demographic profiles, CTSN is starting, in key sites, to implement site-specific UREG benchmarks. This requires sites to collect local demographic data to inform enrollment targets. This focus on location-specific targets has underscored the need to include sites that serve under-represented populations. As part of the Network renewal, NHLBI developed funding opportunities for five linked clinical research centers, which pair experienced clinical sites with sites that serve high-disease burden areas, such as “the Mississippi Delta, Appalachia, and the Stroke Belt (Southeast US), Stroke Buckle (coastal plains areas of the Carolinas and Georgia), Newfoundland/Labrador (Canada), rural or inner-city locales, or involve predominantly minority or low socioeconomic status populations.”⁵⁴ The intent of the NHLBI funding strategy was to provide clinical research resource and mentorship grants to sites and communities that serve underrepresented patient populations, but had few federally-funded clinical trial grants. In addition to enhancing research infrastructure the Network is expanding its collaborations with a broader number of institutions that predominantly serve UREG communities.

Capturing Data on Diversity and Reasons for non-Participation

Recognizing that diversity is broader than sex, race and ethnicity, CTSN has designed its case report forms (CRFs) to incorporate social determinants of health, such as sex at birth, gender identity, race, ethnicity, geographic location, education level, insurance and socioeconomic status. When designing trials, the Network pre-specifies key secondary analyses based on sex, race, ethnicity, and other factors relevant to diversity. While such analyses may not be robustly powered to assess for differential outcomes of therapies, they serve as hypothesis generating and may influence the design and conduct of future trials. In addition, decliner surveys are an important instrument that CTSN uses to understand the systematic loss to enrollment of specific groups of patients. We added such a survey to an ongoing trial evaluating the addition of anticoagulation to background anti-platelet therapy in patients with new onset post-operative atrial fibrillation (PACeS trial)( NCT04045665). These surveys will help elucidate the specific reasons for non-participation, such as transportation issues, time burden, issues related to insurance coverage, language barriers, and provider preference.

Strategies to Enhance Diverse Enrollment

These survey data and discussions with site coordinators and investigators, give insight into specific strategies that the Network may adopt to enhance diverse enrollment, including addressing logistical issues, providing more flexibility for study visits or allowing remote follow-up. One important factor, we discuss above, is variation in rates and timing of referrals, whereby women and individuals from UREGs are less likely to be referred for diagnostic imaging and surgical interventions. The Network is pursuing grant funding to create a consortium of referring primary care physicians and cardiologists who serve underrepresented populations, to increase awareness of clinical trial opportunities and provide direct connections to trial centers. Meanwhile, the CTSN is also forging relationships with other organizations devoted to the mission of improving diverse enrollment, such as the Association of Black Cardiologists and the American Heart Association’s Strategically Focused Research Networks program to improve diversity, equity, and inclusion in cardiovascular clinical trials.⁵⁵

Increasing the Diversity of Investigative Teams and Implicit Bias Training

Having diverse investigators and research staff is vital to improve the overall science and recruitment of diverse patient populations. We plan periodic re-surveying of the demographic make-up of Network participants, so that we can identify the need to improve recruitment of investigators and coordinators from UREGs. These data will also be important to guide inclusion of more women and UREGs in CTSN committees and leadership teams.

In addition to improving the diversity of our research teams, formal training on recruitment of underrepresented patients and implicit bias is needed to support enrollment of UREGs. As mentioned, UREG trial participants often face significant barriers to enrollment and follow-up, many of which are affected by systemic racism and bias. Importantly, research and clinical professionals have not routinely received interactive in-person training on how to address these barriers, beyond routine modules intended to increase awareness on this issue required at most institutions to maintain clinical and research privileges. Additional personalized educational and training programs focused on interventional trials are essential to equip research staff with the knowledge and skills to address these socio-economic and cultural barriers that may influence UREG enrollment. The Network is working to establish these training sessions for its members across a variety of platforms.

CONCLUSIONS

The importance of enrolling more diverse populations into randomized trials cannot be overstated. The need for measures to improve the representativeness of study participants is widely acknowledged. Yet, implementation has lagged due to the complex and multifaceted nature of the problem. Comparing the Network’s experience with CTSN COVID-19 trials suggests that the difficulties encountered with randomizing diverse patients in cardiac surgery trials are heavily shaped by inequities in the health care delivery system. Recognizing these challenges, the CTSN has developed a plan of action to implement specific strategies to improve the diversity of enrollment. The CTSN will prospectively assess how these strategies influence trial enrollment over time as we work to recruit trial participants who are more representative of the community populations we serve. It is our hope that the CTSN experience and plan will also help inform other efforts to move our field forward.