The role of social, economic, and medical marginalization in cancer clinical trial participation inequities: A systematic review

Abstract

Extant literature reveals how patients of marginalized social identities, socioeconomic status (SES), and medical experiences – especially patients of color and older adults – are underrepresented in cancer clinical trials (CCTs). Emerging evidence increasingly indicates CCT underrepresentation among patients of lower SES or rural origin, sexual and gender minorities, and patients with comorbid disability. This review applies an intersectional perspective to characterizing CCT representativeness across race and ethnicity, age, sexual and gender identity, SES, and disability. Four databases were systematically queried for articles addressing CCT participation inequities across these marginalizing indicators, using the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. One hundred one articles were included in a qualitative evaluation of CCT representativeness within each target population in the context of their intersectional impacts on participation. Findings corroborate strong evidence of CCT underrepresentation among patients of color, older age, lower SES, rural origin, and comorbid disabling conditions while highlighting systemic limitations in data available to characterize representativeness. Results emphasize how observed inequities interactively manifest through the compounding effects of minoritized social identity, inequitable economic conditions, and marginalizing medical experiences. Recommendations are discussed to more accurately quantify CCT participation inequities across underserved cancer populations and understand their underpinning mechanisms.

Introduction

Despite the necessity of representative cancer clinical trials (CCTs) to optimize equitable progress in cancer outcomes, overall CCT participation is remarkably low [1,2]. Strong evidence reveals that these low rates of CCT participation are still further compromised among individuals experiencing social, economic, and/or medical marginalization, particularly among patients of color [3–7] and older adults [1,8,9]. Further research increasingly suggests potential underrepresentation in CCTs among other marginalized populations, including sexual and gender minority (SGM) patients [10,11], patients of lower socioeconomic status (SES) [12,13], and patients experiencing greater disability in disease severity, comorbidity, or performance status [1,14,15]. However, systematic limitations in data collection and target variables addressed in prior literature render CCT inequities impacting these underserved groups more elusive [11,16–18], necessitating further research regarding the effects of these marginalizing characteristics on CCT participation. Considering the disproportionate cancer burden imparted upon groups enduring these forms of marginalization [9,10,15,16,19–23], representative CCTs that are generalizable to these populations are of the utmost importance for achieving equitable cancer care outcomes and associated progress across sociodemographic divides.

While prior investigations of CCT participation inequities have explored their effects on various underserved communities in cancer as previously described, these studies have primarily applied a singular perspective to marginalizing characteristics. Specifically, most existing CCT participation literature is limited by inadequate regard for the interactivity of overlapping forms of disadvantage, which serves a critical role in understanding CCT representativeness among the underserved. Intersectionality as a theoretical framework – in its focus on interlocking oppressive systems at the social-structural level and their manifestation in individual experiences [24,25] – is an apt scaffold through which these inequities may be interactively explained. However, despite increasing emphasis on the criticality of an intersectional approach to contextualizing public health outcomes [26], this framework has yet to be directly applied to CCT participation inequities.

This systematic review aims to provide a more comprehensive, ecologically valid characterization of CCT participation inequities to date across social, economic, and medical vectors of marginalization. In examining their independent and multiplicative influences through an intersectional lens, the authors seek to illustrate how race and ethnicity, age, sex, SGM status, SES, and diverse ability indicators have contributed to inequities in CCT participation across time.

Materials and methods

Search methods

This review adheres to Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [27]. The first author developed and tailored a systematic search strategy to four databases, following general consultation with library sciences (Supplementary Table 1): (1) PubMed, (2) Web of Science, (3) PsycINFO, and (4) CINAHL. An initial search using this strategy was conducted on February 15, 2021, and then replicated on April 8, 2024, for newly published articles given significant time passage since the initial search. Covidence, a web-based collaboration software platform that streamlines the production of systematic and other literature reviews, was utilized to facilitate methodology [28].

Eligibility criteria

Inclusion criteria for qualitative synthesis of results were (1) original research; (2) focus on CCT participation inequities regarding age, race and/or ethnicity, SES or one of its specific indicators (e.g., income, education, insurance, employment status), sexual identity, sex and/or gender identity, or ability status or relevant indicators (e.g., performance status, comorbidities); (3) peer-reviewed; and (4) full-text availability in English. Exclusion criteria required the removal of original protocols or reviews that (1) address trial participation disparities unrelated or nonspecific to individuals with cancer; (2) do not explicitly comment on CCT participation inequities; (3) encompass scope beyond the focus of this review, including papers exploring underlying barriers to identified inequities, developing solutions, and pediatric populations; (4) are case studies or reviews without quantitative analysis; or (5) are not peer-reviewed, full-text publications. Excluded papers per criterion four were scanned for eligible references unidentified by the search strategy.

Data extraction procedures

Data extraction was standardized across three domains: (1) study characteristics, (2) methodology, and (3) sociodemographic reporting. The first domain specified the cancer population addressed, aims, sample size, intervention type(s), and target marginalizing indicators (Table 1). The second domain extracted information regarding study design, recruitment methods and databases, measures, and statistical procedures (Table 2). The third domain reported on available study information on sociodemographic characteristics relevant to the modes of marginalization addressed in this review (Table 3).

Table 1. Basic study characteristics

Note: Abbreviations included in this table are utilized as follows, listed alphabetically: ACOSOG = Alliance for Clinical Trials in Oncology; CALGB = Cancer and Leukemia Group B; CAR-T = chimeric antigen receptor T-cell; CCC = comprehensive cancer center; CCT = cancer clinical trial; EDI = Equity, Diversity, and Inclusion; FDA = Food and Drug Administration; HBRT = hypofractionated breast radiotherapy trials; HPV = human papillomavirus; JH-SKCCC = Johns Hopkins Sidney Kimmel Comprehensive Cancer Center; NCDB = National Cancer Database; NCI = National Cancer Institute; NIH = National Institutes of Health; NR = not reported; NSCLC = non-small cell lung carcinoma; OPSCC = oropharyngeal squamous cell carcinoma; SES = socioeconomic status.

Table 2. Methodology of studies

Note: Abbreviations included in this table are utilized as follows, listed alphabetically: AA = African American; AAPI = Asian American or Pacific Islander; ACOSOG = American College of Surgeons Oncology Group; AI/AN = American Indian/Alaska Native; ANOVA = analysis of variances; API = Asian or Pacific Islander; B = Black; BMC = Boston Medical Center; CBTRUS = Central Brain Tumor Registry of the United States; CCC = comprehensive cancer center; CCT = cancer clinical trial; CCR = California Cancer Registry; CCSG = Cancer Center Support Grant; CDC = Centers for Disease Control and Prevention; CDUS = Clinical Data Update Service; CI = confidence interval; CINAHL = Cumulative Index of Nursing and Allied Health Literature; CSPOC = Cancer Surveillance Program of Orange County; CTED = Clinical Trials on Chronic Thromboembolic Disease; CTEP = Clinical Trial Evaluation Program; CTMS = Clinical Trials Management System; DF/HCC = Dana-Farber/Harvard Cancer Center; ECOG = Eastern Cooperative Oncology Group; EMPacT = Enhancing Minority Participation in Cancer Clinical Trials; EH/MR = electronic health/medical record; FDA = Food and Drug Administration; FFS = fee-for-service; GLOBOCAN = Global Cancer Observatory; GOG = Gynecologic Oncology Group; H = Hispanic; HS = high school; JH-SKCCC = Johns Hopkins Sidney Kimmel Comprehensive Cancer Center; ISS = International Staging System; KCI = Karmanos Cancer Institute; L = Latine; LACCSP = Los Angeles County Cancer Surveillance Program; MDCSS = Metropolitan Detroit Cancer Surveillance System; MDS = myelodysplastic syndromes; MHLW = Ministry of Health and Labor, Welfare; NCDB = National Cancer Database; NCI = National Cancer Institute; NDA = new drug application; NH = non-Hispanic; NIH = National Institutes of Health; O:E = observed:expected; NR = not reported; PDQ = Physician Data Query; PS = performance status; SEER = Surveillance, Epidemiology, and End Results; SE = standard error; SES = socioeconomic status; SWOG = Southwest Oncology Group; TCR = Thames Cancer Registry; TNM = Tumor Nodes Metastases; UCSF = University of California – San Francisco; W = White.

Table 3. Social, economic, and medical indicators of marginalization

Note: Abbreviations included in this table are utilized as follows, listed alphabetically: AA = African American; AAPI = Asian American or Pacific Islander; ACOSOG = American College of Surgeons Oncology Group; AFAB = assigned female at birth; AI/AN = American Indian/Alaska Native; AMAB = assigned male at birth; AML = acute myeloid leukemia; ALL = acute lymphocytic leukemia; API = Asian or Pacific Islander; B = Black; CCC = comprehensive cancer center; CCT = cancer clinical trial; COPD = chronic obstructive pulmonary disease; CV = cardiovascular; DLBCL = diffuse large B-cell lymphoma; ECOG = Eastern Cooperative Oncology Group; edu = education; F = female; IQR = interquartile range; M = male; NCI = National Cancer Institute; NH = non-Hispanic; NIH = National Institutes of Health; NR = not reported; O = Other; PCa = prostate cancer; PL = poverty line; PS = performance status; SES = socioeconomic status; SGM = sexual and/or gender minority; W = White; WTP = willingness to participate; y.o. = years old.

Quality and bias assessment

The Mixed-Methods Appraisal Tool (MMAT) [29] was applied across studies to ensure uniform quality ratings across while affording flexibility appropriate to specific article type. The MMAT includes five sets of five-item criteria, with one set applied to each article reviewed contingent on its specific study design. Fulfillment of each of the five criteria for a given study design yields one point. As such, scores range from “0” to “5,” with higher ratings indicating stronger evidence quality. Case-control, cohort, cross-sectional, and meta-analytic studies were assessed using MMAT criteria directed toward nonintervention, descriptive analyses. While this iteration of the MMAT has not been applied to reviews specific to CCT representation, multiple versions of the MMAT have been utilized in recent reviews addressing cancer health disparities [30–32].

Results

Article selection

The combined search strategies yielded an initial 1,812 articles. Nine hundred ninety-three duplicates were removed, including articles from the 2021 search identified in the 2024 search. Eight hundred nineteen titles and abstracts were reviewed for relevance. Two hundred ninety-five articles underwent full-text screening, with 194 studies excluded as detailed in the PRISMA flow diagram (Figure 1). A resulting total of 101 articles were included in the review.

Figure 1. PRISMA flow diagram of selected articles.

Study information

Approximately 66% of included studies (n = 67) primarily addressed tumor-directed, therapeutic trials, with 39.6% providing specific information on intervention types studied (n = 40). Only 12.8% of studies (n = 13) explicitly addressed trials with at least one supportive care, psychosocial, or behavioral component, with 5.0% of all studies (n = 5) exclusively focused on such CCTs. Approximately 18.8% of studies provided minimal detail on trial type eligibility criteria (n = 19). This information is summarized in Table 1.

Case-control studies – that is, studies examining differences between participants and non-participants using patient data repositories – constituted 44.6% (n = 45) of articles, with most utilizing population data to compare characteristics between CCT participants and corresponding oncologic populations. Nearly 29% of studies were conceptualized as cohort designs (n = 29, i.e., using patient data to evaluate predictors of CCT enrollment over time), and 7.5% (n = 7) were cross-sectional studies. Twenty-one studies were meta-analyses of aggregate demographic, socioeconomic, or medical characteristics across published trials (20.8%). Where classification according to these descriptions was ambiguous, our authors relied on self-identification of study design by the cited authors. This information is detailed in Table 2.

Approximately 81% (n = 82) of studies included race and/or ethnicity as a primary focus. In comparison, 54.4% (n = 55) addressed age, 35.6% (n = 36) addressed at least one socioeconomic indicator, 30.7% addressed sex or gender (n = 31), and 26.7% (n = 27) addressed at least one indicator of disability as CCT inequity targets. Only one eligible study addressed SGM status (0.99%) even following several modifications of advanced search strategies (Supplementary Table 1). The availability of social, economic, and medical characteristics across studies is detailed in Table 3.

Quality assessment

Across all 101 studies, the quality mean MMAT score calculated was 4.59, with a median of 5. Score distribution was as follows: 5: 65.3% (n = 66), 4: 28.7% (n = 29), 3: 5.9% (n = 6), 2: 0.0% (n = 0), 1: 0.0% (n = 0). Quality ratings are summarized in Table 1.

Synthesis of findings

Race and ethnicity

Extant literature reflects robust evidence of CCT underrepresentation among patients of color, with mixed findings on representativeness across specific racial and ethnic minority groups. Early studies reflect lower enrollment among patients of color across multiple diagnostic sites, including in therapeutic lung, breast, colorectal, lymphoma, leukemia, and reproductive system CCTs [33], with some revealing decreased participation among patients of color across time (e.g., Baquet: 8.9% annual decrease among Black patients) [34]. Similar early trends are documented in surgical breast, colorectal, and thoracic CCTs [35]. Across the 50 largest National Cancer Institute (NCI) CCTs from 1996 to 2002, Black, Hispanic, and Asian American or Pacific Islander (AAPI) patients were all less likely to enroll in lung and colorectal CCTs, with Black and Hispanic women also less likely to enroll in breast trials [36]. These data indicated the poorest representation among Hispanic patients overall, and lower prostate CCT enrollment – a disparity not observed among other minority groups. This case-control study also showed a proportional decline in CCT enrollment among patients of color despite increased overall CCT participation from 1998 to 2002 [36]. While other evidence supports attenuation of some of these inequities with clinical cooperative group efforts (e.g., Newman: % Black CCT participants vs. cancer population: 10.5% and 9.4%) [37], early research consistently reflects national CCT underrepresentation among patients of color across various cancers.

Studies from the following decade demonstrate continued CCT underrepresentation among patients of color, adjusting for other relevant characteristics [38,39]. Longitudinal data emphasize stagnant therapeutic lung CCT enrollment among Black, Hispanic, and AAPI patients from 1990 to 2012, even with NCI cooperative group support [40]. Similar studies indicate worsening representation of Black women in gynecologic CCTs up to 2013, demonstrating 4.5–15 times lower enrollment than expected [41]. A meta-analysis from this period suggests still poorer trends, reporting 6.5 and 18.5 times lower enrollment than expected for Black women in cervical and ovarian trials, respectively, with representation worse from 2015 to 2018 compared to the late 1990s [4]. Other studies corroborate underrepresentation among patients of color in prostate, breast, colorectal, pancreatic gastric, hematologic, myelodysplastic, and varied sample CCTs at the catchment area level [42,43], in multi-site and multi-trial pharmacologic studies [44,45], Food and Drug Administration (FDA) CCTs with pharmaceutical sponsors [46], and in NIH CCTs from 1999 to 2019 [5]. Meta-analytic studies corroborate persistence of these inequities, reflecting poorest representation among Black and Hispanic patients in various therapeutic breast, colorectal, lung, prostate, pancreatic, renal, melanoma, and multiple myeloma CCTs, with such data collectively spanning 1981 to 2017 (e.g., Guerrero et al.: Not Reported, White, Black, and Hispanic CCT enrollment fractions [EF], respectively: 66.95%, 25.94%, 1.08%, 0.16%).[47,48]. Other national gynecologic CCT data not only accentuate Black and Hispanic underrepresentation but also larger disproportionate effects on Hispanic women with uterine and cervical cancers [49]. State-level studies reveal similar trends, with women of color less likely to enroll in early-stage breast radiotherapy CCTs overall, with Hispanic, then AAPI, then Black women, respectively, showing the lowest representation [50]. However, other findings during this period indicate the highest relative underrepresentation among AAPI, then Hispanic, then Black patients across breast, prostate, colorectal, and lung CCTs in national databases [6].

A few studies prior to 2021 suggest minimal inequities in CCT representation among patients of color with certain diagnoses, locations, and trial types. One national study reports no racial or ethnic differences in opportunities to participate in breast CCTs from 2013 to 2014 [51], with similar findings regarding prostate CCTs in earlier years [33,35]. A case-control analysis of FDA-approved therapies for breast, lung, colorectal, and prostate cancers showed persisting underrepresentation among patients of color relative to non-Hispanic White patients, though with recent improvements (% participants of color, 2008–2013: 20%; 2014–2017: 29%) [45]. Other evidence emphasizes representative accrual to surgical breast, thoracic, and sarcoma CCTs among Black and Hispanic patients [52], as is observed in NCI Community Cancer Centers Program CCTs overall [53]. Some findings during this period suggest equitable representation among patients of color in therapeutic lung CCTs, despite participation inequities in five other diagnostic sites [5]. Nonetheless, nonsurgical breast CCTs reliably demonstrate worsening representation among patients of color despite progress in other cancers (e.g., Zullig: 1996: < 1% vs. 2009: 3.5% enrollment difference between White and minority women, p < .001) [54]. Other studies reporting nonsignificant participation differences in some areas emphasize persisting trends toward underrepresentation among patients of color where typically observed [33].

Studies published within the past three years corroborate the intractability of CCT underrepresentation among patients of color while providing further nuance surrounding these inequities. State and national cohort, case-control, and meta-analytic studies of overall CCT representation evidence participation inequities that disproportionately impact Black [55–59] and/or Hispanic [47,48,56,57,59,60] patients in phase I [57], II [61], and III [59]; radiation [56,61]; drug [55,58,59]; brachytherapy [62]; and general CCTs [48,60] utilizing updated datasets and study repositories (e.g., Bero: Black proton participants vs. population 6.0% vs. 12.7%; Choradia: Hispanic participants vs. population: 7.1% vs. 13%; ). While some studies note mild representative improvement among Black [60,63] and Hispanic patients (e.g., Saphner: no significant inequities in White, Black, and Hispanic accrual: 90.4%, 6.6%, 1.9%; p = .078) [63], others demonstrate worsened representation in common cancers over time (e.g., 2009 vs. 2011–2015) [47]. Other case-control and meta-analyses emphasize underrepresentation among AAPI, Native Hawaiian, and American Indian/Alaska Native patients, in recent, CCTs for various prevalent cancers [48,56,64]. Still other findings evidence minimal underrepresentation among patients of color [63] and demonstrate even higher CCT participation among Asian patients, though such results have been primarily limited to singular institutions diagnostic sites, or trial types (e.g., Wagar: polymerase inhibitor CCT EF: White: 1.5%, Black: 0.47%, Hispanic: 0.33%, AAPI: 2.38%) [47,57,59,63].

Contemporary, cancer-specific studies reveal the importance of tumor site in dictating such inequities. Recent cohort, case-control, and meta-analytic studies of the most prevalent cancers continue to demonstrate underrepresentation among Black [62,65–72], Hispanic [62,65–70,72], Asian [62,72,73], and American Indian and Alaskan Native [62,67,69] individuals in breast [62,66–68,71,74], lung [65,67–71], and prostate CCTs (e.g., Ajewole: FDA oral chemotherapy CCT enrollment composition, 2009–2019: White: 71.5%, Black: 2.5%, Hispanic: 2.3%; Ladbury.: brachytherapy CCTs, enrollment incident disparity: Asian: −2.65%) [62,65,67,71,72,75–77]. Similar patterns are observed in understudied cancers, with CCT underrepresentation among Black [68,72,78–87], Hispanic [68,72,78–80,82,83,86], and Asian [72,80,86] patients with gastrointestinal [65,67,68,73,88], hepato-biliary [78,79], pancreatic [68,78,89,90], gynecologic [62,80,91], HPV-associated oropharyngeal [92], renal and urologic [72,87,93], hematologic [68,81–85], and neurologic [86] malignancies. Some evidence suggests potential mitigation of such inequities for certain cancers in recent years [65,80,86,94–97], particularly among Asian individuals (e.g., Javier-DesLoges: breast, colorectal, lung, and prostate participation odds ratios [OR], 2000–2004 vs. 2015–2019: Black: 2.19, 1.15, 1.54, 1.14; Hispanic: 3.32, 2.46, 2.21, 1.70; Asian: 1.94, 2.48, 3.88, 1.64) [47,65,67,85]. Nonetheless, such findings are primarily limited to studies with strong infrastructural support or smaller, singular institutional studies, while other contemporary studies reveal stagnation or worsened disparities over time (e.g., Owens-Walton: unchanging representation quotients from 2000 to 2017) [72,74].

Limited data reflect mixed findings regarding representation among patients of color in psychosocial CCTs. Some such evidence suggests minimal enrollment inequities between non-Hispanic and Hispanic women; however, even these data reflect higher attrition risk for Hispanic and immigrant women [98]. An institutional study of all cancers identified no racial/ethnic participation inequities among patients who were eligible for two pain and symptom-focused CCTs; however, patients of color were more likely to be ineligible [99]. Recent analyses suggest that psychosocial CCT representation among patients of color may be particularly contingent on intervention type, target population, and funding. For example, national evidence indicates Black underrepresentation in prostate exercise, advanced disease, and nongovernmental CCTs; adequate representation in dietary and multi-component trials; and disproportionately higher participation in pelvic floor muscle training, localized disease, and government-funded trials relative to their non-Black counterparts [100]. While observing poor representation among patients of color overall, a meta-analysis of integrated palliative CCTs suggests mitigated underrepresentation among Black patients compared to therapeutic CCTs (EF: 5.7% vs. 3.0%) [101], with similar, state-level results observed among Black men in behavioral CCTs [94]. Analyses of two psychosocial CCTs among women with gynecologic, gastrointestinal, and thoracic cancers demonstrate parallel trends, with even higher enrollment among Hispanic patients [102]. Nonetheless, other evidence investigating breast, lung, prostate, and colorectal CCTs reveals trends toward poorer representation among Black patients in supportive care trials compared to tumor-directed studies [6].

Age

Strong evidence demonstrates CCT underrepresentation among older adults across time. Early such inequities are observed in NCI Cooperative Group trials in prevalent cancers, with patients 65 to 74 and patients older than 75 years old, respectively, exhibiting progressively lower accrual to nonsurgical, therapeutic trials compared to younger patients [36]. Such findings are replicated in later general CCT samples [13,34], surgical CCTs [35], and NCI, state-specific data further classifying older age subgroups [33,34]. Early large studies corroborate these trends across cancer types and within drug-specific trials, with underrepresentation among older adults relative to their incident populations[45,103–105] and lower likelihood of CCT enrollment with age [53], with progressively greater underrepresentation (e.g., Talarico: participants vs. population, respectively: 65+: 25% vs. 60%; 75+: 4% vs. 31%) [105].

Contemporary literature corroborates these findings, reflecting continued CCT underrepresentation among older adults over time. Recent institutional, state, and national cohort, case-control, and meta-analytic studies reveal persisting underrepresentation among older adults in surgical [39], drug [42,66,106], brachytherapy [62], and other trial types [65,94,51]; phase I [57], II [107], and III[108–110] trials; and multimodal [107] CCTs in general (e.g., Baldini: CCT referral vs. population 70 +: 17.7% vs. 50%) [42,57,58,63,106,107,111,112]. Such findings are replicated in specific cancers, including breast [42,65,66,39,108,113], gynecologic [49,51], lung [65,96,108], prostate [65,76,94,108], pancreatic [89,90], hepatic [79], gastroesophageal [114], gastrointestinal [42,65,73,108,115], renal [87], skin [109], head and neck [116], other solid organ [5,107], and hematologic cancers [5,110,117], with further evidence of greater inequities among the oldest groups [115].

Nonetheless, other recent studies report no age differences in CCT enrollment, especially controlling for relevant covariates (e.g., Dudipala: OR: 1.023) [44,50,92,97,118,119]. However, these findings have primarily been exclusive to one institution, diagnosis, or state. Further, evidence of more equitable age representation overall is qualified in persistent inequities relative to the incident population for that specific cancer[40], or among patients initially referred to [40], eligible for [102,113], or discussed for such trials [97,112]. Furthermore, other longitudinal and population-based studies demonstrate stagnated or worsened age inequities in CCT participation over time, particularly among the oldest patients (e.g., Zhao et al. median age difference [DMA] between participant and incident disease median age: −8.15; US DMA before 2017 vs. after 2017: −5.90, −8.00) [49,108,111].

Socioeconomic status (SES)

Early national case-control analyses reflect breast CCT underrepresentation among low-SES patients by multiple indicators, including area poverty, unemployment, income, education, and individual government insurance [120]. Other site-specific studies document similar findings in various cancers, with lower CCT participation associated with higher material deprivation and lower social class (e.g., Mohd Noor.: Index of Multiple Deprivation [IMD] = 5 enrollment OR: 0.53, relative to least deprived IMD = 1) [34,119]. Another national, cross-sectional study revealed strongly prohibitive effects of low income on breast, colorectal, prostate, and lung CCT participation, controlling for other variables (< $50,000 income vs. $50,000+: OR: 0.73) [13], with progressively larger disparities among patients with the lowest incomes [13,118]. State analyses extend similar findings to area income in breast, genitourinary, gastrointestinal, and myelodysplastic CCT enrollment (e.g., Brierley et al.: average median income, participants vs. non-participants: $68,896 vs. $61,241) [43,121]. Other earlier studies reveal how unemployment, lower educational attainment [13,122], and governmental insurance [42,49] predict CCT underrepresentation in breast and other common cancers.

Contemporary studies within the past three years have increasingly focused on and further substantiated CCT underrepresentation among lower SES patients. Multilevel cohort, case-control, and meta-analytic studies evidence the effects of lower area income (e.g., Hue: stage IV participants vs. non-participants < $40,227: 11.4% vs. 19.1%) [63,87,89,112,123], education (e.g., Eskander: CCT participation, higher vs. lower high school attainment OR: 2.0) [73,78,80,87,89,90,94,112], insurance (e.g., Shinder: CCT participation, uninsured, Medicaid, or Medicare vs. private insurance ORs, respectively: 0.57, 0.43, 0.59) [70,73,78,87,90,96], or overall SES (e.g., Kwak: CCT participation, lowest [1] vs. median [4] SES group OR: 0.60) [63,70,82,96] on breast [66], prostate [77,94], lung,[70,96] gastrointestinal [73], pancreatic [78,89,90], hepatic [73,78], gynecologic [80], renal [87], brain [123], hematologic [83], and mixed CCT underrepresentation [57]. Other meta-analyses emphasize how limited SES reporting in CCTs significantly compromises research regarding its effects on representation [77,88].

Nonetheless, other studies present contrasting findings. data have shown higher breast CCT enrollment among Medicaid-eligible and lower-education patients [38]. Other studies have reflected higher surgical breast CCT participation with higher area education, but lower income [39], with similar income findings in gynecologic trials [80]. Some contemporary studies have observed no SES impact on CCT enrollment (e.g., Perni et al. participation OR, $100,000 median income vs. < $50,000: 1.28) [57] or attenuated effects in multivariate models [63,87,89], though these studies only examined socioeconomic factors as covariates. While the most equivocal evidence appears in the relationship between income and CCT participation, recent authors conceptualize such findings in reliance on area, rather than patient, indicators due to systemic data deficiencies [80]. Despite the nuances observed in these mixed findings, the literature provides growing evidence of socioeconomic CCT inequities by various indicators.

Sex

Mixed literature on sex-related CCT inequities suggests contingency of representation on cancer and trial type. For example, early national data suggest higher therapeutic CCT enrollment among men with colorectal and lung cancers (participation, men vs. women OR: colorectal, lung, respectively: 1.30, 1.23) [36] with similar results replicated in center-specific analyses [124]. However, other early evidence regarding surgical CCTs reflects the reversal of this trend, with women five times more likely than men to enroll overall in a combined, national breast, colorectal, lung, and prostate sample [35]. Other data provide further insight into contrasting results, suggesting greater overall CCT participation among men, though lower enrollment compared to women when examining sex-specific cancers [34].

More recent studies have observed more equitable CCT representativeness across sexes. National cohort and meta-analytic studies addressing various cancers, including sex-specific [44] and rare diagnoses [117], reflect minimal sex differences in representation (e.g., Costa: observed-expected ratio, % male participants: 1.03). Similar evidence has emerged in psychosocial CCTs, revealing no sex differences in participation (e.g., Huang: % eligible enrolled in symptom CCT, within each sex: women: 75%, men: 78%) [99]. Longitudinal analyses reveal improvement in lung CCT representation over time among women younger than 65 years old (overall enrollment disparity difference between sexes reduced 0.07 to 0.03, 1994–2012) [40], as is consistent with equitable sex representation among younger patients in earlier lung and other CCTs [36]. Some state-specific evidence reflects even higher therapeutic lung, colorectal, and sex-specific CCT enrollment among women relative to men [54], as with the aforementioned surgical CCT findings [35]. Nonetheless, women’s underrepresentation persists in certain rarer cancers, such as myelodysplastic syndrome [43] or HPV-associated oropharyngeal CCTs (e.g., Gordis: % total female enrollees: 11.8%) [92]. Conversely, other data reflect disadvantages for men for certain CCT types across cancers, such as eligibility for chemoradiation trials [125] and participation in sex-related CCTs [34]. FDA approvals between 2008 and 2017 similarly demonstrate attenuated inequities when including sex-related CCTs, while simultaneously revealing worsened women’s representation over time when exclusively examining trials for cancers affecting all sexes (% women: 2008 to 2013: 47%, 2014 to 2017: 37%) [45].

Studies within the past three years continue to reveal minimal sex-related CCT inequities. Multiple institutional, state, and national cohort and case-control studies suggest equitable CCT participation across sexes in colorectal [68], lung [68–70], pancreatic [68,89], neurologic [123], hematologic [68,83,85], and mixed samples [63], with some analyses suggesting higher representation among women (e.g., Saphner: participation OR, men vs. women: 0.70) [63,70]. However, some of these findings are restricted to specific institutions, with their results challenged by more nationally representative analyses suggesting persistent underrepresentation among women in colorectal [65,88], lung [65], neurologic [86], and hematologic CCTs [84]. Additional studies document lower participation among women in hepatic [79], head and neck [95], and renal CCTs [87], in addition to women’s underrepresentation in overall therapeutic [60], radiation [56], phase II and III [57], non-sex- [63], and sex-specific diagnostic CCTs [60]. Though recent evidence of improved representation among women is qualified by these contrasting findings, contemporary results suggest partial mitigation of such inequities over time for certain diagnostic sites (e.g., Javier-DesLoges: women’s participation OR, 2015–2019 vs. 2000–2014: 1.38, with remaining inequities relative to men [OR: 0.89]) [65,86].

Ability, staging, and functional status

Until the past three years, few studies had examined indicators of ability status as direct contributors to CCT participation, typically focusing on staging (i.e., measured by tumor size, lymph node presence, and/or metastases) [126], comorbidity, and more rarely, performance status ratings. Earlier findings evidence higher participation in breast [50,51,113,127], colorectal, lung, prostate [13], and multiple myeloma CCTs [117] among patients with lower staging or fewer comorbidities (e.g., Unger: participation OR, comorbidity score: 0.81), though primarily examine such indicators as covariates. Some institutional analyses characterize exclusionary comorbidities as restrictive to CCT participation across multiple cancers [122], while other data document positive relationships between symptom risk and therapeutic CCT enrollment in rarer cancers (e.g., Brierley: participation OR, high vs. very low risk: 1.88) [43]. Still, other investigators report no association between disease characteristics, comorbidities, and CCT participation [33], though these early studies still conceptualize such ability proxies as covariates, rather than key predictors.

While evidence remains scarce compared to other marginalizing indicators, contemporary studies have increasingly documented relationships among comorbidity, functional impairment, and CCT participation. National cohort, case-control, and meta-analytic studies reveal the potentially restrictive impact of comorbidity burden or associated lower performance status on pancreatic,[78,89,90], breast [66], lung [128], hepatic [78], gynecologic [80], other gastrointestinal and genitourinary [128], renal [87], head and neck [116], solid organ [129], and overall CCT representativeness (e.g., Green: % comorbidity score = 0, participants vs. non-participants: 69.2% vs. 51.6%) [58,112]. Other analyses, while not directly centering ability proxies as enrollment predictors, evidence the covarying impact of performance status on CCT participation (e.g., Bruno: lung participation OR, Eastern Cooperative Oncology Group [ECOG] score, 2 vs. 0: 0.27) [130]. Still other recent results evidence positive relationships between comorbidity burden and CCT participation, though these findings are exclusive to one state and disease site [95].

Intersectionality in CCT participation inequities

The above-summarized data provide robust evidence of persistent CCT underrepresentation among patients of color and older adults, with mixed evidence of changing representativeness over time across diagnostic sites and trial types. Recent evidence reveals similar relationships between SES and CCT participation, demonstrating how lower education, inadequate insurance, and to a smaller extent, lower income, may further stifle CCT representativeness. While sex disparities have negatively impacted CCT participation depending upon cancer type, some contemporary studies evidence more equitable CCT representation in common cancers. While the singular impacts of such factors quantitatively vary, the interactivity among these social, economic, and medical marginalizing indicators further complexifies CCT representativeness.

This review characterizes the nexus among race, ethnicity, and SES as one of the most intricate intersections in determining CCT representation. Early breast CCTs have revealed diminishing underrepresentation among Black patients after considering area poverty, unemployment, and Medicaid coverage (participation OR, Black vs. White: 0.99) [120]. Later population data corroborate such findings, illustrating partial attenuation of Black and Hispanic underrepresentation in surgical breast CCTs when accounting for income and education [39], as well as insurance [131]. Institutional analyses of multiple cancers have demonstrated resolution in CCT underrepresentation among patients of color after accounting for age, sex, and deprivation index [119]. However, other evidence reveals underrepresentation among higher income and privately insured Black and Hispanic women compared to their less affluent counterparts in gynecologic [49,98] and breast CCTs (e.g., Fayanju: participation OR, Black and Hispanic, respectively, median income $63,000+ vs. < $38,000: 0.45, 0.19) [39]. These findings constitute a reversal of typically observed relationships, wherein racial, ethnic, and socioeconomic marginalization multiplicatively serve to restrict CCT participation with concurrent marginalization, rather than poorer participation among higher SES women of color. A meta-analysis of FDA approvals, regardless of SES, emphasizes the intersection among sex and minoritized identity, with the greatest underrepresentation observed among women of color in prevalent cancers (i.e., % Black participants breast sample: 2%) [45].

Studies within the last three years have increased explicit efforts to explore the interactive influences of racial, ethnic, and socioeconomic marginalizing indicators on CCT participation, while similarly indicating nuanced results across diagnostic sites. Multilevel cohort, case-control, and meta-analytic studies demonstrate the simultaneous impact of minoritized race/ethnicity, lower area SES, and inadequate insurance coverage in limiting breast [66], gynecologic [80], pancreatic [90], and renal [87] CCT participation (e.g., Khadraoui: participation ORs, racial/ethnic minority vs. White: Black: 0.70, Hispanic: 0.53, Asian: 0.44, Other: 0.48; education, 15.3%+ vs. < 5.0% without high school education: 0.41). Similar studies demonstrate partial contingency of hematologic CCT underrepresentation among people of color on lower area income or insurance coverage [82]. Still, other recent studies corroborate persistent CCT inequities that disproportionately affect among women of color regardless of income, in gastrointestinal trials [73], as is consistent with earlier breast and gynecologic CCTs [49,39].

Studies investigating relationships among race, ethnicity, and SES in determining CCT representation have increasingly revealed potential contributions of disease characteristics, comorbidity burden, and performance status. For instance, early analyses demonstrate how controlling for advanced disease diminishes otherwise observed racial CCT inequities [37,50]. Similar interactive relationships have been observed in early case-control studies regarding lung CCTs, interpreting underrepresentation among Black and other patients of color within the intersections among race, SES, insurance, comorbidity, and performance status [124]. Other national data corroborate higher comorbidity among Black patients considered for CCTs (medical comorbidity presence, OR: 1.53) [53]. More recent analyses directly explore how ability indicators color the intersectional effects of race, ethnicity, and SES on CCT representation [66,80,87,94,129]. While some such studies reveal how higher staging and comorbidity may further limit CCT participation among minoritized or lower SES patients (e.g., Yekeduz: % Black participants vs. population: 2.1% vs. 9.8%, with 82% total sample with ECOG 0–1) [66,72,80,87,129], others offer opposing evidence among certain underserved populations. Specifically, some studies indicate increased CCT participation among patients of color with higher comorbidity burden and staging, such as Hispanic men with prostate cancer [94]. Still others indirectly examine complex, intersectional influences of comorbidity, illness characteristics, and ability on CCT representativeness, suggesting poorer overall CCT participation due to the COVID-19 pandemic, though with unexpected impacts on participation inequities (e.g., Choradia et al.: participation ORs, 2005–2020, each vs. White patients: Hispanic: 0.52, American Indian: 0.41, AAPI: 0.81; peak participation among these underserved in 2020, despite lowest year of enrollment across population) [60,68].

Such patterns are further influenced by age and sex, especially among older adults of color with an increased comorbidity burden. Early analyses demonstrate how older age compromises breast, colorectal, thoracic, and prostate CCT participation across racial and ethnic groups, though drives underrepresentation otherwise unobserved in younger patients among women of color [36]. Other investigators demonstrate how older age heightens gynecologic CCT attrition risk for Hispanic, but not for non-Hispanic, women [98]. Recent national cohort, case-control, and meta-analytic studies strengthen evidence of simultaneous underrepresentation regarding older age, comorbidity, performance status, and other marginalizing factors underpinning CCT underrepresentation (e.g., Kaanders: % participants with World Health Organization [WHO] 0–1 or Karnofsky performance score 90–100: 70%; median age, participant vs. population: 57, 64 [58,107,109,112,116], with some evidence emphasizing how trial characteristics themselves may limit participation among older adults with higher disease burden [107].

Regarding intersecting sex influences, some early state studies indicate elevated racial disparities among men relative to women in therapeutic trials for common cancers [34,54], with recent studies similarly accentuating how cancer sex-specificity may underpin racial and ethnic representativeness in radiation CCTs (i.e., Black underrepresentation observed in all CCT types except sex-specific female [13.1% sample] and male [18.4% sample] US trials) [56]. While quantitatively unexamined to date, contemporary studies have begun to comment on how relationships among these marginalizing factors may be furthermore impacted by sexual minoritization, through its influence on preexisting health and CCT eligibility [66]. Overall, relationships among social, economic, and medical marginalizing indicators in underpinning CCT inequities have gained increasing attention in recent years, with more investigators explicitly exploring the structural, intersectional context of such factors when interpreting their findings regarding CCT representativeness [72,39].

Discussion

This review sought to describe CCT participation inequities via multiple modes of social, economic, and medical marginalization, including race, ethnicity, age, sex, SGM identity, SES, and ability. Its findings contribute novel insights regarding the impact of such factors on CCT inequities, including strengthened evidence for national CCT underrepresentation among racial and ethnic minority groups and older adults across various cancers and trial types. To a lesser, albeit increasing extent, these results reveal compromised CCT participation among lower SES patients across various metrics, especially education and insurance; however, these findings are dependent on aggregate, rather than individual, SES indicators. This review further offers insights into the effects of ability status on CCT participation, with a growing focus on comorbidity burden in recent years.

These findings reflect minimal to modest evidence of improvement in representativeness across the past several decades. While exhibiting some progress in racial, ethnic, and sex representativeness in certain intervention types, CCT inequities are observed across most cancers and study designs in recent large-scale analyses. Studies focused on CCT representation among the underserved have more than doubled within the past three years, while accentuating a persisting absence of data investigating such inequities among SGM patients. Nonetheless, while bolstering evidence of intractable CCT inequities across various other marginalizing indicators and cancers, contemporary investigations have increasingly provided more nuanced insights into their complex interplay in determining CCT representativeness.

More important than enduring inequities observed in a singular examination of each marginalizing indicator, however, is the intersection among these social, economic, and medical characteristics and their effects on CCT inequities. These results demonstrate the partial underpinning of CCT underrepresentation among patients of color by parallel preexisting socioeconomic and health disparities. Further, the literature illustrates how the intersection among racial/ethnic minority status, SES, and other marginalizing indicators may interactively predispose individuals with cancer to more aggressive disease, higher comorbidity, or poorer performance status, thus compromising CCT participation among the underserved. These conditions are further influenced by the strong relationship between age and higher ineligibility risk due to similar preexisting health inequities, and their heightened impact on CCT representation in their intersectional context with other marginalizing indicators. Despite an increased focus on these relationships in recent CCT literature, few articles explicitly allude to their intersectional, structural nature, with most studies addressing multiple marginalizing indicators as potential confounding covariates at best.

This review is the first to conceptualize existing CCT inequities across several modes of social, economic, and medical marginalization through an intersectional perspective. These findings accentuate how numerous marginalizing indicators limit CCT representativeness with multiplicative implications, further preventing equitable participation among those with overlapping experiences of social, economic, and medical oppression. Further, this review is uniquely underpinned by a central recognition of social inequality, context, power, and justice using intersectionality as a theoretical scaffold for understanding public health [25,26].

Limitations

This review is limited in its absence of articles addressing CCT participation among SGM individuals, yielding only one study that transiently mentioned SGM identity as one factor affecting CCT participation while interpreting its results. While this may indicate limitations in the search strategies applied to this review, this absence of SGM studies persisted with extensive adjustments, thus likely reflecting large deficits in the literature itself. Another limitation is a lack of explicit investigation regarding the impacts of rurality on CCT participation – a crescent area of research important to understanding CCT representation through an intersectional perspective.

Other limitations consist in a low number of articles that specifically address (1) supportive care, psychosocial, behavioral, or quality of life interventions and (2) longitudinal retention in studies. Further, few included articles directly investigate relationships between social, economic, and medical marginalization through an explicitly intersectional perspective, primarily examining such interactive influences through reductive, additive models that merely control for covarying factors. These results are also qualified by the limitations in article quality evaluation. While the use of the MMAT for quality assessment accommodated the diversity of articles included, this flexibility inversely limits the standardization of ratings across various article types. Further, while intersectionality constitutes a necessary lens through which investigators must accurately view health inequities, optimal practices for quantification of such outcomes through this theoretical paradigm remain tenuous.

Implications and future directions

This review characterizes the current state of the literature quantifying CCT participation inequities that disproportionately impact the underserved in cancer care. Its description of such inequities reveals little ambiguity in CCT underrepresentation among certain marginalized groups, especially among older adults, racial/ethnic minorities, and by some indicators, patients originating in lower SES areas or with greater disability. This review thus constitutes a strong foundation to further investigate underpinning barriers that sustain these inequities and potential solutions to dismantle them. Its findings accentuate the necessity of future research focused on (1) mixed evidence regarding specific social, economic, and medical indicators in determining CCT participation, (2) the role of intersectionality and underlying mechanisms in explaining such inequities, and (3) persistently understudied marginalized populations in the investigation of CCT representation, especially patients who are SGMs, of lower SES or rural origin, or live with comorbid disabilities. Additional research is necessary to understand the generalizability of such findings to CCTs beyond those that are tumor-directed and longitudinal participation patterns.

This review accentuates the persistence of CCT participation inequities across various vectors of social, economic, and medical marginalization through an intersectional perspective across the past four decades. As such, these findings emphasize the urgency of identifying and dismantling barriers that sustain these inequities. Through such efforts, investigators and clinicians may strive toward the eradication of inequities in cancer outcomes and equitable benefits from advancements in cancer care among the underserved.