Clinical research in the community

A 54-year-old African American man who lives with and cares for his elderly mother presents with generalized lymphadenopathy, weight loss, and night sweats. Diagnostic lymph node biopsy is pathologically consistent with germinal center diffuse large-cell lymphoma (DLBCL) without mutations or gene rearrangements of BCL-2, BCL-6, or MYC. Upon staging, he was determined to be stage IVB. His Eastern Cooperative Oncology Group performance status was 1, and his age-adjusted International Prognostic Index (IPI) was 3, with an estimated poor survival of 30 months.¹  He was offered participation in an ongoing phase 3 randomized clinical trial of polatuzumab, rituximab, cyclophosphamide, doxorubicin, and prednisone (Pola-R-CHP) versus rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisone (R-CHOP) by his community oncologist but declined due to a required 30-mile travel to an academic center, his caregiver responsibilities, and a desire to be treated by his well-trusted local care team. He then received 6 cycles of R-CHOP and achieved complete remission. Nine months later, he relapsed. He is interested in aggressive therapies but wants to remain in the care of his community oncologist.

The most common high-risk hematologic malignancies typically requiring intense and complicated therapies are acute leukemias, multiple myeloma (MM), and lymphomas, with approximately 150 000 patients diagnosed annually in the United States,^2,3  most of whom are treated at centers within 5 miles of their residence through their community medical oncologist.⁴  However, clinical cancer trials have traditionally been conducted at academic medical centers, while 80% to 85% of cancer patients are treated at community-based clinical practices.^5,6  The current workforce of community medical oncology comprises highly trained physicians with substantial experience and expertise in clinical research, but most do not have access to clinical trials of new and emerging therapies for high-risk hematologic malignancies. Therefore, community-based cancer research is critical in advancing cancer care for the large, diverse patient population receiving treatment in various local health care delivery settings. In addition, the participation of community oncologists and primary care physicians in cancer prevention, control, and treatment trials significantly facilitates the translation of research advances into practice.⁷  Most clinical trial enrollees receive their treatment at academic centers, but community oncologists enroll more patients into cooperative trials sponsored by the National Cancer Institute (NCI), with about 65 percent of these patients entering from community-based practices.⁸ 

Challenges to the availability of cancer therapeutic trials in community oncology practices include inappropriateness of trial designs and significant system-, patient-, and physician-level barriers resulting in less than 5% enrollment.⁹ 

Since 1982, NCI has initiated several programs to improve access to NCI-sponsored oncology clinical research, aiming to engage community physicians and academic medical centers and to facilitate the incorporation of research findings into practice. Two such networks, the Community Clinical Oncology Program (CCOP) and the Minority-Based Community Clinical Oncology Program (MB-CCOP),¹⁰  were created to increase participation in clinical trials. MB-CCOP was developed to provide infrastructure for clinical trials in the institutions that serve communities with large minority and underserved populations. In 2007, NCI expanded its efforts by creating the NCI Community Cancer Centers Program (NCCCP). NCCCP was a pilot, public-private program emphasizing new research and care delivery partnerships with organized patient communities, community-based health care providers, and academic researchers.¹¹  Self-reported data from NCCCP sites showed an increase of NCI-supported phase 3 studies by 16% compared with 8% nationally and an improvement of patient accrual by 66% compared with 30% nationally. Additionally, enrollment of racial and ethnic minorities in oncology trials increased by 82% (Figure 1). The accrual of patients aged 65 years or older also rose by 221%.¹²  In 2014, NCI created the NCI Community Oncology Research Program (NCORP) to further community oncology clinical research. The goal of NCORP is aligning the 2 existing programs, CCOP, MBCCOP, their research bases, and NCCCP. NCORP is intended to build on the strengths of the CCOPs/MBCCOPs and NCCCPs and expand the scope of research to include cancer care delivery. NCORP is intended to serve as a network to support clinical trials, cancer care delivery, and cancer disparities research. Core principles of NCORP are “including community- based organizations with a variety of research capacities linked to the NCI's Clinical Trials Network; providing support to oncology practices with varied organizational settings as a collaborative network; engaging patients within and outside of clinical trials, organizations, and clinicians as research subjects; encouraging commitment of management within organizations to support the research agenda; and integrating cancer care disparities, care delivery research, and clinical trials.”¹³ 

Despite lofty efforts by the NCI, however, clinical trial disparities persist, particularly with novel therapeutics through pharmaceutical industry–sponsored trials and not NCI. A study of 358 trials (pharmaceutical company–sponsored trials, 85; Southwest Oncology Group [SWOG] Cancer Research Network trials, 273; comprising 93,825 patients [pharmaceutical company- sponsored trials, 46,313; SWOG trials, 47,512]) for 15 cancer types from 2008-2018 also found significant underrepresentation of Blacks in pharmaceutical company–sponsored trials compared with SWOG trials (2.9% vs 9.0%), which was consistent across individual cancer types.¹⁴  Race reporting is frequently omitted in clinical trials resulting in regulatory approval but is worse in studies outside regulatory purview. Between 2008 and 2018, only 7.8% of 230 trials (recruiting 112,293 patients) documented the 4 major races in the US. The representation of trial participants was 76.3% White, 18.3% Asian, 3.1% Black, and 6.1% Hispanic. In perspective, compared with their proportion of US cancer incidence, this underrepresents the proportion of Blacks and Hispanics (22% and 44% of expected, respectively) compared with Whites and Asians (98% and 43.8% of expected, respectively).^15,16  This gap in representation is worse for specific tumor types, particularly in prevalence-adjusted participation for cancers that are more common in African Americans.¹⁷  Pooled data from 9 large cooperative group clinical trials in newly diagnosed MM over 2 decades showed only 18% of participants were non-White,¹⁸  shocking for a disease with incidence rates in Blacks more than double those seen in Whites (15.9 vs 7.5 cases per 100,000). This trend also extends to mortality (5.6 vs 2.4 MM deaths per 100 000 for African Americans compared with Whites.^19,20  Additionally, in pivotal trials leading to US regulatory approval of immune checkpoint inhibitors, Black patients constituted less than 4% of enrollees. This is particularly problematic because clinical responses to immunotherapeutic agents depend on unique, individual, frequently racially determined, genetically mediated host and tumor biological interactions.²¹ 

NCI has also mandated NCI-designated comprehensive cancer centers to create catchment areas that promote increased community participation in clinical research, but recent data in patients with acute myelogenous leukemia showed that over a 15-year study, there were 3041 trial enrollees at US sites; national incidence adjusted enrollment odds by race-ethnicity showed that non-Hispanic Black, non-Hispanic Asian, and Hispanic persons were enrolled at significantly lower rates than non- Hispanic Whites (NHW). Non-Hispanic Native American enrollment was significantly higher. Enrollment odds were lower for non-Hispanic Black, non-Hispanic Asian, and Hispanic enrollees at comprehensive cancer center sites when adjusted by catchment area incidence (Figure 2). Among trial enrollees, there were no univariable predictors of biobank participation; however, NHW race-ethnicity (OR 1.33; 95% CI 1.12, 1.57; P < 0.001) was associated with correlative study participation. Multivariable models of correlative study participation, with predictors selected based on univariable significance, are for all trial enrollees and when restricted to biobank enrollees; in both cases, the NHW race predicted participation.²² 

The American Cancer Society estimates that 89 380 new cases of lymphomas will be diagnosed in the US in 2023, with 21 080 estimated deaths.²³  Despite improvements in treatments for hematologic malignancies and specifically in lymphomas, disparities in outcomes remain. In non-Hodgkin's lymphomas, recent data showed that compared with White patients at diagnosis, Black patients were younger and more likely to have ≥1 comorbidity, be HIV positive, and have both B-symptoms and stage IV disease (all p < 0.001). Compared with age-matched White patients, Black patients age ≤60 had worse median overall survival (OS) (46 vs 76 months) along with 5- and 10-year OS (65% vs 69%) (all p < 0.001). Comparable results were seen for Black and White patients between the ages of 61 and 79 years, but these differences were not demonstrated for patients ≥80 years. On multivariate analysis, Black race was independently associated with worse OS (HR 1.06; CI 1.01-1.10; p  =  0.02). Interestingly, the propensity-matched analysis demonstrated no significant OS difference between Black and White patients (median 127 vs 117 months; HR 1.0; CI 0.94-1.06; p   =   0.90).²⁴  Comparable results were also seen in patients with chronic lymphocytic leukemia (CLL). Results from the National Cancer Database of CLL patients diagnosed from 2004-2018 showed that White patients compared with Black patients had a shorter median OS of 7.0 years (CI 6.7-7.3 years) versus White patients' (9.14 years [CI 9.0-9.3]); p < 0.001), as well as inferior OS at 5 years (61% vs 69%) and 10 years (36% vs 46%), p < 0.001. On a multivariate analysis adjusted for age and Charlson-Deyo score, Black race was independently associated with shorter OS (HR 1.51 [CI 1.46-1.57]; p < 0.001). Referenced to the White population, Black patients diagnosed between 2004 and 2006 had an HR of 1.64 (CI 1.52-1.76) for mortality, and those diagnosed between 2016 and 2018 had an HR of 1.64 (CI 1.44-1.85).²⁵  This was a surprising finding, given the availability of Bruton's tyrosine kinase and BCL-2 targeted inhibitors in treating CLL over the same period. Pivotal studies of targeted agents currently approved for CLL in the US enrolled a limited number of Black patients.²⁶ 

Racial disparities are due to several factors, chiefly the lack of access to high-quality care across the cancer continuum, including participation in clinical research. Due to emerging complex treatment protocols, this issue is particularly important in high-risk hematologic malignancies. Thus, strategies to decrease these disparities must include an attempt to close the research participation gap. However, increasing access alone is insufficient to close these gaps. For example, even among individuals with a median annual household income of ≥$75,000, 5-year relative cancer survival is lower among Black people (67%) than among White people (72%²⁷ ).

Clinical trial disparities also result in the creation of 3 main problems of health care big data disparities: data absenteeism (lack of representation from underprivileged groups), data chauvinism (faith in the size of data without considering quality and contexts),²⁸  and, in its most extreme form, what we now term data hallucination (the assumption and use of nonexistent data to generalize treatment recommendations, outcomes, and results). For illustration and relevant to our clinical case, examples of these problems are seen in the clinical studies of polatuzumab²⁹  in the initial treatment of DLBCL (7% Black and Hispanic enrollment), chimeric antibody receptor T-cell therapy in relapsed DLBCL (5% Black and Hispanic enrollment),³⁰  and the generation of the widely used IPI³¹  respectively (no racial or ethnicity data reported).

Barriers to participation in clinical trials (Figure 3) typically disproportionally affect minority patients and ultimately results in delayed accrual, delayed generation of clinical data, and the generalizability of such data to all persons, thus promoting outcome disparities. Therefore, addressing such barriers may also improve cancer population outcomes. Improving clinical trial availability in community oncology practices where most cancer patients are treated can uniquely improve and address these disparities.

1. Access to a clinic can be influenced by many structural factors, such as transportation, travel costs, access to insurance, and availability of childcare

2. Availability of a clinical trial for the patient's histology and stage

1. Narrow eligibility criteria

1. Physician decision or preference, a primary reason for nonparticipation in half the patients for whom a protocol was available and the patient was eligible

2. Lack of appropriate incentives to participate in clinical research

3. Time-consuming trial paperwork

4. Obtaining informed consent

1. Patient unease or fear about the prospect of participating in clinical trials, including residual mistrust of medical science due to past abuses such as the infamous Tuskegee Syphilis Study or the history of human experimentation with radiation following World War II

2. Complicated consent forms

3. Patient dislike of randomization

4. Patient's unease about potentially toxic effects of chemotherapy in trials, especially for experimental therapies

5. Patient's ardent desire for a particular treatment they wish to receive after discussion with their physicians

6. More frequent monitoring than nontrial care, for example, traveling to and from the clinic

7. Concern about how to pay for trials

1. Older age

2. Race

3. Low socioeconomic status

Our patient illustrates the challenges of clinical trial participation for minority high-risk lymphoma patients treated in the community. First, the lack of minority enrollment in major clinical trials prevents us from truly estimating his prognosis using the IPI, potentially resulting in us underestimating his clinical risk (data hallucination). Second, significant structural and personal barriers precluded his participation in clinical studies that could have improved his survival. Additionally, upon relapse, his access to more complicated but potentially curative therapy was impacted by the lack of this therapy's immediate availability in the community due to the limited participation of community oncologists in CAR-T trials, thus impacting its clinical availability at community practices and resulting in further outcome disparities. Thus, strategies to overcome and minimize these barriers are desperately needed to improve outcomes for high-risk hematologic malignancies. These strategies include enhancing community oncology access to clinical trials and the participation of diverse populations, including our patient.

We recommend the adoption of our 5-step strategy at research sites and sponsors to overcome these disparities and to promote clinical research in community oncology practices. DRIVE³³  is a practical, immediately actionable 5-step strategy for promoting and improving diversity, equity, inclusion, and access (DEIA) in clinical trials for minority patients. DRIVE aims to use these strategies to correct inequalities and promote the overall health of humankind, as established in the Greenberg Report,³⁴  equating diversity and safety due to the potentially generalizable clinical data generated. Additionally, DRIVE uses proven principles and techniques to create meaningful improvements in cancer care and outcomes.

D: Diversity officer for clinical research studies; to develop an actionable, flexible, and prospective diversity plan for clinical research. Diversity officers should be funded by study sponsors and institutions as part of the clinical trial enterprise, very similarly to funding of data safety and monitoring boards.

R: Ranking of clinical studies for diversity; generating an informational tool for determining clinical research diversity (Figure 4). Ranking should be based on independent audits of self-reported accrual data and generated by DEIA promoting heath care organizations and major medical societies.

I: Individual DEIA plan; created by each investigator to promote enrollment of diverse subjects.

V: Verification of study diversity; creating a process for confirming research diversity. Verification should be based on the same auditing used for evaluating clinical trial data for efficacy and safety.

E: Elevate and enhance training of minority investigators and research team members; to promote minority patient participation by all clinical trial stakeholders

Strategies to enhance the participation of patients with high-risk malignancies in community practices can address health disparities and promote the early adoption of potentially curative therapies in community oncology practices. Outcomes for lymphomas are significantly affected by race, as shown in the clinical case presented, as well as in the other high-risk hematologic malignancies (Figure 5). Thus, efforts to overcome barriers that potentially prevent clinical trial availability and enrollment in community oncology practices are important steps in this direction. Our recommendations include adoption of the DRIVE strategy, as well as a full-throated adoption of the elements promulgated in the US Food and Drug Administration Omnibus Act of 2022 by the pharmaceutical industry, including the creation of a prospective diversity plan for major clinical trials and expanding clinical trials to community oncology practices. Additionally, we suggest promoting the NCORP program's goals. These steps would further enhance clinical research availability in community oncology for high-risk hematologic malignancies.

Maya N. Birhiray is the daughter of Ruemu E. Birhiray.

Ruemu Ejedafeta Birhiray: no competing financial interests to declare.

Maya Nicole Birhiray: no competing financial interests to declare.

Ruemu Ejedafeta Birhiray: Nothing to disclose.

Maya Nicole Birhiray: Nothing to disclose.