Increasing diversity in clinical trials: demographic trends at the National Cancer Institute, 2005-2020

Abstract

Background

We described participant demographics for National Cancer Institute (NCI) clinical trials at the clinical center (NCI-CC participants) of the National Institutes of Health to identify enrollment disparities.

Methods

We analyzed NCI-CC data from 2005 to 2020, calculated enrollment fractions, compared with the US cancer population represented by the Surveillance, Epidemiology, and End Results cancer incidence data (2018) and the Cancer in North America database (2018), and compared further with clinical trial disparities data from the NCI Community Oncology Research Program and National Clinical Trials Network (2005-2019), and from ClinicalTrials.gov (2003-2016).

Results

NCI-CC (38 531 participants) had higher enrollment fractions for older adults (8.5%), male (5.6%), non-Hispanic (5.1%), and Black or African American (5.3%) participants; lower women proportion across race and ethnicity; and fewer female sex-specific cancer (6.8%) than male sex-specific cancer (11.7%) participants. NCI-CC had lower median age than Surveillance, Epidemiology, and End Results (54.0 vs 65.4); more Black or African American participants (12.0% vs 11.1%); and fewer women (41.7% vs 49.5%), White (76.1% vs 80.5%), Asian or Pacific Islander (4.6% vs 6.0%), American Indian or Alaska Native (0.3% vs 0.5%), and Hispanic participants (7.1% vs 13%). NCI-CC had more Black or African American and Asian or Pacific Islander participants; fewer Hispanic participants than the NCI Community Oncology Research Program and National Clinical Trials Network; more Black or African American and Hispanic participants; fewer Asian or Pacific Islander participants than ClinicalTrials.gov data. Improvement was noted for NCI-CC (older adults, Black or African American, Asian or Pacific Islander, Hispanic participants).

Conclusion

We found lower representation of older adults, women, Asian or Pacific Islander, American Indian or Alaska Native, and Hispanic participants vs the US cancer population and higher representation of Black or African American vs US cancer population and oncology clinical trials. Multifaceted efforts are underway to reduce disparities in cancer clinical trials at the NCI-CC.

Efforts to address treatment disparities among underserved racial, ethnic, sex, and elderly population groups have increased over the past 2 decades (1). However, despite these efforts, disparities persist, including poorer survival outcomes and decreased representation in cancer clinical trials for patients from racial and ethnic minority groups. Studies consistently demonstrate disparities in clinical outcomes across cancer types and geographic regions (2,3). Additionally, disparities in clinical trial participation remain pervasive, with barriers for minority and low-income patients limiting their access to clinical research opportunities and inhibiting scientific discovery and growth (4). These disparities not only impact survival outcomes but also diminish the generalizability of trial results across racial, ethnic, sex, and age groups (4). Thus, greater efforts to understand and address disparities in cancer clinical trials are needed.

Multiple factors account for persistent disparities in cancer clinical trial enrollment, including the ability to obtain information about available trials, fear of adverse effects, mistrust of science and the medical field, and the time and financial resources required for travel and treatment (time and financial toxicities) (5-7). Understanding and addressing these factors may enhance clinical trial participation and representation (8). Inequities in cancer outcomes have received much attention in recent years and have led to coordinated efforts to increase diversity in cancer clinical trials (7). Notably, the US Food and Drug Administration has released draft guidance aimed at increasing participation from underrepresented racial and ethnic groups and has recently updated recommendations on collecting race and ethnicity data from trial participants (9,10). Despite these efforts, barriers to trial participation remain, and resulting disparities for certain patient groups persist (11,12).

Numerous divisions of the National Cancer Institute (NCI), part of the National Institutes of Health (NIH), conduct clinical research at the NIH Clinical Center (CC) in Bethesda, Maryland, known as America’s Research Hospital (13). The clinical research involves primarily interventional early phase (first-in-human, phase 1, phase 2) clinical trials that are mainly investigator initiated, NCI-sponsored and funded, single-center or multicenter, addressing all stages of disease but mainly recurrent and metastatic cancer. Participants are either referred by their primary oncology teams outside of the NCI or self-referred. Furthermore, clinical trials conducted by the NCI at the CC provide clinical care at no cost to participants or billing of third-party payers and assist with participants’ travel expenses (14). Although there has been research describing the demographics of participants in NCI-approved cancer clinical trials across the nation, such as those conducted by the NCI Community Oncology Research Program (NCORP) and National Clinical Trials Network (NCTN) (15,16), until now, the demographics of clinical trial participants at the NCI-CC have been little known.

Here we sought to describe the demographic characteristics of the cohort of participants in NCI-CC clinical trials during 2005-2020 and to compare them with the NCI’s Surveillance, Epidemiology, and End Results (SEER) Cancer Registry as well as the North American Association of Central Cancer Registries’ Cancer in North America (CiNA) database to identify potential disparities with the US cancer population. In addition, to identify potential differences in the representation of demographic groups between different trial processes, we compared the race and ethnicity characteristics of our cohort with previously reported results about oncology clinical trials diversity from the NCI’s NCORP and NCTN and from ClinicalTrials.gov.

Methods

NCI-CC population data

We obtained deidentified data on enrollment in clinical research protocols at the NCI-CC for the period of 2005-2020 from the Biomedical Translational Research Information System (BTRIS) database of the NIH Clinical Center, a data repository that contains patient-level information including clinical data in free text form and discrete data elements from the electronic health record for all participants who receive care in the CC (17). Demographic data included age of participants at the time of enrollment, sex (female, male, unknown), race (White, Black or African American, Asian, Hawaiian or Pacific Islander, American Indian or Alaska Native, multiple race, unknown), ethnic group (Hispanic or Latino, non-Hispanic or Latino, unknown), state, and country. For age at enrollment, we created 2 categories: aged younger than 65 years and 65 years and older. For race, we merged Asian with Hawaiian or Pacific Islander to reconcile with the SEER and CiNA databases, and we merged multiple race with unknown under the label unknown. Further, for participants in this dataset we retrieved cancer site parsed, where available, from data abstracted from EHR text notes by NIH BTRIS.

Population cancer data

Data on cancer incidence and patient demographics for 2018 were obtained from the SEER cancer registry (18). The SEER Research Limited-Field Database includes data from approximately 14 million patients with malignant tumors (invasive cancers only) obtained from 18 cancer registries across the country representing approximately 48% of the US population (19). Additional data were obtained from the North American Association of Central Cancer Registries, Inc. (NAACCR) Cancer in North America (CiNA) database, which includes data from registries across all 50 states (20).

Data on diversity in oncology trials

We obtained enrollment data on the diversity of participants in oncology clinical trials from a report on the representation of minorities in all therapeutic trials reported in ClinicalTrials.gov from 2003 to 2016 (21) and a report on the accrual of minorities in clinical trials from NCORP and the NCTN from 2005 to 2019 (15).

Statistical analysis

We tabulated age, sex, race, and ethnicity for the NCI-CC, SEER, and CiNA populations. We used the SEER database for primary comparison, as it is a well-established disparities research resource, includes a diverse demographic population, and has been reported as comparable to the US population (22). We calculated the enrollment fraction for age, sex, race, and ethnicity by dividing the number for each group in the NCI-CC population by the estimated cancer incidence for each demographic subgroup in the 2018 SEER database (23). Enrollment fraction is defined as the number of trial enrollees divided by the estimated number of cancer patients in the United States through SEER and provides a basic understanding of enrollment in trials vs cancer occurrence in the general population. We performed Pearson χ² test and odds ratios (ORs) to evaluate the relationship between enrollment fractions within the demographic subgroups for the review period. To analyze the temporal trends in enrollment, we examined the overall enrollment over the period studied and the proportions of participants by age, sex, race, and ethnicity and performed Pearson correlation trend tests. Two-sided tests were used; a P value less than .05 was considered statistically significant. Statistical analyses were conducted with R version 4.1.2.

Results

During the study period, 40 007 participants were enrolled on clinical studies at the NCI-CC of whom 38 527 (96.3%) came from the United States, which is the population we analyzed further. We first described the main demographic determinants (age, sex, race, ethnicity) for the NCI-CC population and then compared them with the 2018 population from the SEER registry and the CiNA registry (Table 1 also provides the reference distribution of age, sex, race, and ethnicity in the general population from the US Census, 2018). At the NCI-CC, there were more participants aged younger than 65 years at enrollment than those aged 65 years and older (77.1% vs 22.9%, respectively) and more male than female participants (58.3% vs 41.7%, respectively). Most participants identified as White (76.1%), followed by Black or African American (12.0%), Asian or Pacific Islander (4.6%), and American Indian or Alaska Native (0.3%); 6.9% of participants had race documented as Unknown. Furthermore, more participants identified as non-Hispanic or Latino than Hispanic or Latino and Unknown (90.6% vs 7.1% vs 2.3%, respectively).

Table 1.

Demographic characteristics of participants in National Cancer Institute (NCI) protocols, 2005-2020 (n = 38 531), cancer population (NCI Surveillance, Epidemiology, and End Results cancer registry, 2018, and Cancer in North American registry, 2018), and US general population (US Census, 2018; one race^a)

Characteristics	NCI protocol participants	Surveillance, Epidemiology, and End Results registry population	Cancer in North American registry population	US Census
	No. (%)	No. (%)	No. (%)	No. (%)
Age, y
Younger than 65	29 702 (77.1)	348 900 (44.0)	783 715 (42.6)	274 318 162 (84.0)
65 and older	8829 (22.9)	444 855 (56.0)	1 055 983 (57.4)	52 369 339 (16.0)
Sex
Female	16 074 (41.7)	393 279 (49.5)	903 642 (49.1)	165 801 767 (50.8)
Male	22 453 (58.3)	400 476 (50.5)	936 056 (50.9)	160 885 734 (49.2)
Unknown	4
Race
American Indian or Alaska Native	119 (0.3)	3988 (0.5)	7830 (0.4)	4 145 811 (1.3)
Asian or Pacific Islander	1778 (4.6)	47 827 (6.0)	61 648 (3.4)	19 927 892 (6.1)
Pacific Islander	45 (0.1)	—	—	793 787 (0.2)
Black or African American	4634 (12.0)	87 998 (11.1)	195 951 (10.7)	43 732 024 (13.4)
Unknown	2661 (6.9)	15 085 (1.9)	121 161 (6.6)	—
White	29 339 (76.1)	638 857 (80.5)	1 453 108 (79.0)	249 961 025 (76.5)
Ethnicity
Hispanic or Latino	2719 (7.1)	103 268 (13.0)	166 531 (9.1)	59 639 869 (18.3)
Non-Hispanic or Latino	34 909 (90.6)	690 487 (87.0)	1 673 167 (90.9)	267 047 632 (81.7)
Unknown	903 (2.3)	—	—	—

a

Source: Annual Estimates of the Resident Population by Sex, Race, and Hispanic Origin for the United States: April 1, 2010, to July 1, 2019 (NC-EST2019-SR11H); Annual Estimates of the Resident Population for Selected Age Groups by Sex for the United States: April 1, 2010, to July 1, 2019 (NC-EST2019-AGESEX); Source: US Census Bureau, Population Division; Release Date: June 2020. — = not available.

In the SEER registry and in the CiNa registry, the group aged younger than 65 years was smaller (44.0% and 42.6%, respectively) than those aged 65 years and older (56.0% and 57.4%, respectively), and there were slightly more men than women (50.5% and 50.9% vs 49.5% and 49.1%, respectively). Most participants identified as White (80.5% and 79.0%, respectively), followed by Black or African American (11.1% and 10.7%, respectively), Asian or Pacific Islander (6.0% and 3.4%, respectively), and American Indian or Alaska Native (0.5% and 0.4%, respectively); Unknown included 1.9% and 6.6%, respectively. The largest ethnicity was non-Hispanic or Latino (87.0% and 90.9%, respectively) with Hispanic or Latino at 13.0% and 9.1%, respectively. The SEER and CiNA data groups for age, sex, race, and ethnicity follow the same order and have comparable proportions except for Asian or Pacific Islander race (6.0% vs 3.4%, respectively), Unknown race (1.9% vs 6.6%, respectively), and Hispanic or Latino ethnicity (13.0% vs 9.1%, respectively), and, consequently, comparisons of NCI-CC to SEER and CiNA data yield similar conclusions, except for the Asian or Pacific islander race category.

We further studied the enrollment fractions per demographic subgroup (age, sex, race, ethnicity) (Table 2). Enrollment fractions for those aged 65 years and older were less than for those aged younger than 65 years (2.0% vs 8.5%, respectively; OR for trial enrollment in participants aged 65 years and older vs younger than 65 years was 0.23, 95% confidence interval [CI] = 0.22 to 0.24; P < .01). Enrollment fractions were higher for male participants than for female (5.6% vs 4.1%, respectively; OR for trial enrollment was 0.73, 95% CI = 0.71 to 0.74; P < .01). The enrollment fractions for Black or African American participants at the NCI-CC were higher than that for White participants (5.3% vs 4.6%, respectively; OR for trial enrollment for Black or African American vs White was 1.15, 95% CI = 1.11 to 1.18; P < .01). Conversely, Asian or Pacific Islander participants and American Indian or Alaska Native participants had enrollment fractions lower than White participants (3.7% and 3.0%, respectively; OR for Asian or Pacific Islander vs White participants was 0.81, 95% CI = 0.7 to 0.85; P < .01; OR for American Indian or Alaska Native was 0.40, 95% CI = 0.33 to 0.48; P < .01). The enrollment fractions for the non-Hispanic ethnic group were higher than for Hispanic (5.1% vs 2.6%; OR for trial enrollment in Hispanic vs non-Hispanic was 0.52, 95% CI = 0.50 to 0.54; P < .01).

Table 2.

Protocol enrollment, 2005-2020; enrollment fractions in the National Cancer Institute Cancer Center population demographic subgroups (race, ethnicity, sex, age)

Characteristics	Trial enrollees	2018 Surveillance, Epidemiology, and End Results incidence	Enrollment fraction (%)	OR (95% CI)	P
	No. (%)
Age, y
Younger than 65	29 702 (77.1)	348 900 (44.0)	8.5	Referent
65 and older	8829 (22.9)	444 849 (56.0)	2.0	0.23 (0.22 to 0.24)	< .01
Sex
Female	16 074 (41.7)	393 276 (49.5)	4.1	0.73 (0.71 to 0.74)	< .01
Male	22 453 (58.3)	400 473 (50.5)	5.6	Referent
Race
American Indian or Alaska Native	119 (0.3)	3988 (0.5)	3.0	0.40 (0.33 to 0.48)	< .01
Asian or Pacific Islander	1778 (4.6)	47 827 (6.0)	3.7	0.81 (0.77 to 0.85)	< .01
Black or African American	4634 (12.0)	87 997 (11.1)	5.3	1.15 (1.11 to 1.18)	< .01
Other	2661 (6.9)	15 082 (1.9)
White	29 339 (76.1)	638 855 (80.5)	4.6	Referent
Ethnicity
Hispanic	2719 (7.1)	103 266 (13.0)	2.6	0.52 (0.50 to 0.54)	< .01
Non-Hispanic	34 909 (90.6)	690 483 (87.0)	5.1	Referent
Unknown	903 (2.3)	—

a

CI = confidence interval; OR = odds ratio; — = not available.

We reviewed the proportions of sex, race, and ethnicity in a cancer site–specific manner and compared with SEER database. We focused on a selection of the common cancer types (prostate cancer, breast cancer, thoracic cancer, colorectal and anal cancer, gynecologic cancers, renal, ureter or urinary bladder, melanoma) across sexes (24), with the addition of the less common but sex-specific penile and testicular cancer (Supplementary Table 1, available online; one race reported). There were fewer female-specific cancers (2535, 6.6% of total) than male-specific cancers (4488, 11.6% of total US NCI-CC participants), contrasting with SEER where there were more female-specific cancers than male specific (168 144, 21.2% vs 109 275, 13.8%, respectively). In the non-sex-specific cancers we examined, in NIC-CC participants there were also consistently fewer female participants (median 41.9% from 47.9% in thoracic cancer to 35.3% in renal, ureter, or urinary bladder cancer) than male (median 58.1% from 52.1% in thoracic cancer to 64.6% in renal, ureter, or urinary bladder cancer), as in SEER. Compared with SEER, the proportions were comparable for thoracic (female NCI-CC 47.9% vs SEER 48.1%) and colorectal and anal cancers (female NCI-CC 45.8% vs SEER 46.9%), but there were fewer female participants in the NCI-CC vs SEER in melanoma (38% vs 41.2% respectively), while in renal, ureter and bladder, the proportion of female participants in the NCI-CC was higher than in SEER (38.5% vs 30.4%, respectively). The NCI-CC cohort distribution of race across different disease sites was consistent with the overall distribution, in decreasing order: White (median 78.0%, ranging from 72% in colorectal to 96.4% in melanoma), Black or African American (median 12.0%, range from 1.1% in melanoma to 17.3% in prostate), Asian or Pacific Islander (median 4.2%, range from 0.3% in melanoma to 9.9% in thoracic), and American Indian or Alaska Native (median 0.2%, from 0 in penile-testicular to 0.5% in gynecologic) participants. The proportion of Black or African American participants in NCI-CC was higher across all sites in comparison with SEER except in gynecologic cancers (7.9% in NCI vs 12.0% in SEER), while the proportion of Asian or Pacific Islander participants was lower except in thoracic cancer (9.9% in NCI vs 5.7% in SEER). The distribution of ethnicity in NCI-CC was also consistent with the overall pattern, with more non-Hispanic participants (median 93.3%, ranging from 87% in penile and testicular to 96.5% in melanoma) than Hispanic participants (median 5.1%, from 2.1% in melanoma to 12.2% in penile and testicular). In comparison with SEER, the proportion of Hispanic participants was lower across all cancer sites.

We further examined the joint distribution of sex, race, and ethnicity (Supplementary Table 2, available online). This revealed lower proportions for women across all race and ethnic groups in the NCI-CC data, contrary to the SEER registry data, where there were more women in all groups except for the Hispanic ethnicity group, which included more men than women in NCI-CC and SEER data, though notably more so in SEER (3.4% vs 3.6% and 4.4% vs 9.7%, respectively).

We also compared race and ethnicity in the NCI-CC cohort with the findings of a report on enrollment of minorities in trials by NCORP and NCTN from 1999 to 2019 and a report of minority enrollment in all oncology therapeutic clinical trials reported as completed in ClinicalTrials.gov from 2003 to 2016. The ClinicalTrials.gov data included 44.4% pharmaceutical company–sponsored trials, 31% academic or foundation-sponsored trials, and 24.8% NCI-sponsored trials. The NCI NCORP and NCTN minority enrollment report tabulated Hispanic ethnicity with race, so for this comparison we recoded the NCI-CC and SEER data accordingly. Compared with the NCORP and NCTN groups, the NCI-CC cohort had more Black or African American participants (11.9% vs 9.0%) and more Asian or Pacific Islander (4.6% vs 3.6%) but fewer Hispanic participants (7.1% vs 8.0%). Compared with data from ClinicalTrials.gov, the NCI-CC cohort had more Black or African American participants (11.9% vs 6.0%) and more Hispanic participants (7.1% vs 2.6%) but fewer Asian or Pacific Islander participants (4.6% vs 5.3%) (Figure 1).

Figure 1.

Race and ethnicity across oncology clinical trials. CT.gov = ClinicalTrials.gov; NCI-CC = National Cancer Institute Cancer Center; NCORP = NCI Community Oncology Research Program; NCTN = National Clinical Trials Network; SEER = Surveillance, Epidemiology, and End Results.

In terms of temporal enrollment trends over the period studied (2005-2020), we found that the number of new participants per year varied (median = 2420.5, range = 2064-2752; Figure 2, A;Supplementary Table 3, available online). The lowest total enrollment was recorded in 2020 (2064 participants) during the COVID-19 pandemic, and pre–COVID-19, the lowest enrollment was in 2008 (2072 participants). The highest total enrollment was recorded in 2012 (2752 participants), followed by a decreasing trend through 2016, after which a constant increase was observed (interrupted in 2020 likely secondary to the COVID-19 pandemic). The median age at enrollment in 2007 was 50 years, after which the median age gradually increased, reaching a plateau of 55 years in 2013 (Figure 2, B). The proportion of older participants (aged 65 years and older) increased over time from a low of 15.0% in 2006 to a peak of 28.7% in 2018 (Supplementary Table 4, available online). The proportion of male and female participants remained generally stable over time (Figure 3, C; Supplementary Table 3, available online). During the period studied, female participation peaked in 2006 (46.2%) and decreased to a low in 2013 (37.8%); it increased thereafter to a second high in 2020 (43.3%). The percentage of Black or African American participants peaked in 2012 (14.6%) (Figure 3, A; Supplementary Table 5, available online), and Asian or Pacific Islander participation increased toward the end of the period studied and peaked in 2020 (6.7%), a year of low enrollment. The proportion of Hispanic participants increased over time (12.5%), peaking in 2020 (Figure 3, B; Supplementary Table 6, available online).

Figure 2.

(A) Annual enrollment and (B) median age over time for the National Cancer Institute Cancer Center population.

Figure 3.

Temporal trend of (A) race, (B) ethnicity, and (C) sex (National Cancer Institute Cancer Center population).

Discussion

Clinical trials are vital to the development of novel cancer treatments and scientific innovation, but equitable and adequate representation by age, sex, race, and ethnicity are imperative to conduct trials that more accurately reflect the US cancer population (25). Characterization of enrollment enables identification of inequalities and potential barriers such as access and cost. Enrollment demographic reports have been published for trials by the NCI NCTN, cooperative groups, and trials in ClinicalTrials.gov, but to our knowledge, this is the first report for the NCI-CC. In this retrospective analysis of the demographics of participants in cancer clinical trials at the NCI-CC, we identified disparities in patient representation by age, sex, race, and ethnicity compared with the US cancer population (represented in the SEER and CiNA registry data) and differences with the representation of minority groups in oncology clinical trials (from the NCORP and NCTN and from ClinicalTrials.gov). We also found improvements over time in enrollment of several underrepresented groups. Collectively, findings from this work highlight the need for multifaceted efforts to improve equitable access to care across the United States for clinical research.

In the NCI-CC cohort, older adult participants (aged 65 years and older) were fewer than those aged younger than 65 years and were underrepresented in comparison to SEER and CiNA data (Table 1) and had a lower enrollment fraction than those aged younger than 65 years (Table 2). Potential reasons for the underrepresentation of older adults may be restrictive eligibility requirements with regard to performance status and comorbidities, underreferral from oncologists, and patient and caregiver factors (26). Clinical trial availability for pediatric oncology at the NCI-CC may also skew the age distribution toward younger age groups. Geriatric assessment and other aging-associated tools could be a plausible strategy for future efforts to enhance accrual of older adults to clinical trials (27).

Regarding the representation of women, although the proportion of women is lower in all 3 datasets—NCI-CC, SEER, and CiNA data—there is underrepresentation in the NCI-CC relative to SEER and CiNA (Table 1), and the enrollment fraction for female participants was lower than for male participants (Table 2). Further review of the proportion of women across race and ethnicity in NCI-CC and SEER (Supplementary Table 1, available online) revealed that, in the NCI-CC, the proportion of women was lower across all racial and ethnic groups. This was contrary to the SEER data, where despite the higher overall proportion of men, the proportion of women was higher in all groups except for the Hispanic ethnicity group (which included fewer women than men) and, in fact, notably so than NCI-CC (4.4% vs 9.7% and 3.4% vs 3.6%, respectively). This disparity may possibly reflect systemic biases, such as underreferral of women, but may also reflect variations in the activity of clinical trial programs addressing sex-specific cancers. Our findings highlight the well-known pervasiveness of female underrepresentation in clinical trials (28).

Review of the representation of race categories in the NCI-CC population and in comparison with the general cancer population represented by the SEER and CiNA registries data revealed a similar distribution in terms of frequency order with higher numbers of Black or African American participants in the NCI-CC over SEER and CiNA, while Asian or Pacific Islander participants were fewer in the NCI-CC relative to SEER, although slightly higher than in CiNA. The higher representation of Black or African American participants, which has also been reported for NCI-sponsored NCTN trials compared with pharmaceutical companies (16), may reflect demographic characteristics of the NCI-CC referral base, given that in the NCI-CC participant population from the District of Columbia, Maryland, and Virginia areas, which constitutes close to half (48%) of the NCI-CC participants (29), there is higher proportion of Black or African American participants than in the non–District of Columbia, Maryland, and Virginia areas (Supplementary Figure 1, available online). Notable in the examined datasets is the underrepresentation of American Indian or Alaska Native participants, considering that this group makes up 2.1% of the US population (US Census 2018) and that the standardized age-adjusted cancer mortality rate of the American Indian or Alaska Native population is equal to that of all US races (30). This may be due to the geographic distribution of American Indian or Alaska Native participants: only 60% of the American Indian or Alaska Native population lives in metropolitan areas, the lowest percentage of any US population group. Of this population, 13% lives on reservations or other trust lands, which can be rural, with limited access to major health-care institutions and opportunities for referral to clinical trials (30). In addition, mistrust in health care is prevalent in many US populations underrepresented in clinical trials, including the American Indian or Alaska Native population (31).

Although in all 3 datasets studied, there was a higher proportion of non-Hispanic than Hispanic individuals, there were fewer Hispanic participants in the NCI-CC dataset compared with SEER and CiNA data, consistent with the known underrepresentation of Hispanic and Latino participants in clinical trials and biobanking (32). This may reflect barriers in access to care and treatment as well as lower referral rates of Hispanic participants for clinical trials and the most commonly diagnosed cancers among the Hispanic population (33).

We performed a review of the distribution of sex, race, and ethnicity across cancer sites and compared with SEER database, examining sex-specific cancers and the most common non-sex-specific cancers (thoracic cancer, colorectal and anal cancer, renal, ureter and urinary bladder, melanoma). We found that in the NCI-CC there were fewer total female-specific cancer participants than male-specific participants, contrary to SEER, whereas for non-sex-specific cancers there were fewer total female than male individuals in both datasets. We also found a similar ranking relative to race as with the overall data in the NCI-CC and SEER, with a higher NCI-CC proportion of Black or African American individuals for all examined sites except women’s malignancies and a lower NCI-CC proportion of Asian or Pacific Islander individuals with the exception of thoracic cancer. The NCI-CC had lower proportion of Hispanic participants than non-Hispanic participants across all sites and compared with SEER across all sites, though with an appropriate increase in the case of increased Hispanic race proportion in the cancer population, such as in the penile and testicular category (34). Consequently, the findings across cancer sites are consistent with the overall findings for sex, race, and ethnicity and identify an underrepresentation of women’s malignancies. Women’s malignancies is also the only group in NCI-CC with less Black or African American participants than SEER.

We further compared the NCI-CC cohort with race and ethnicity enrollment data for oncology clinical trials by external NCI-related groups (NCORP and NCTN) and to previously published data for oncology clinical trials aggregated from ClinicalTrials.gov. We found that the representation of Black or African American participants in the NCI-CC cohort compared favorably against the NCORP and NCTN trials data and the ClinicalTrials.gov data, similarly to the comparison against the US cancer population. Notably, the NCORP and NCTN data ranked higher than the ClinicalTrials.gov data, in concordance with the previously reported NCTN trials’ better representation compared with pharmaceutical company–sponsored trials (16), given the latter constituting the majority of the trials in the referenced ClinicalTrials.gov data (21). In addition, between the NCI-CC and the NCORP and NCTN group, there were small differences in the representation of Asian or Pacific Islander (higher in NCI-CC) and Hispanic participants (higher in NCORP and NCTN), while compared with the ClinicalTrials.gov data, the NCI-CC data had better representation of Hispanic participants and a small difference in Asians or Pacific Islander participants (higher in ClinicalTrials.gov).

Temporal analysis of enrollment for the NCI-CC cohort revealed a peak of overall enrollment in 2012 and a trough in 2020 during the COVID-19 pandemic. An increase in the proportion of older adults (aged 65 years and older) was noted from 2006 to 2018, along with a 5-year increase in median age at enrollment from 2007 to 2013. These increases likely reflect the increasing age of patients with cancer, partially because of the aging of the population in general and the notion that cancer represents a disease associated with aging (35). Further, over the period studied, the predominant enrollment groups in racial, ethnic, and sex subgroups remained unchanged. However, in addition to the older adults, there has been an improvement in the representation of Black or African American and Asian or Pacific Islander individuals, as well as Hispanic individuals, particularly toward the end of the period studied, contrary to the representation of women, which has not improved over time.

The NCI is committed to developing strategies to increase the diversity of patients on clinical trials that will serve as a guide for clinical trials in the United States. The NCI Center for Cancer Research (CCR) is engaging in ongoing efforts to combat barriers to participation and to provide information about clinical research and access to clinical trials to underrepresented groups. For participants enrolled in a study at the NCI-CC, testing, appointments, imaging, and treatments are provided at no cost to participants or are billed to third-party insurers. Furthermore, the CCR is aiming to offset financial hardship by covering transportation expenses and providing a lodging stipend for participants traveling from within the United States, including transportation costs for the first visit to the CCR. This policy was expanded to include travel expenses for consultation and enrollment visits to increase access. The CCR has also engaged with federal and nonfederal entities to facilitate forming institutional partnerships to increase awareness of clinical trials and coordinate rapid referral processing for potential clinical trial participants. These referrals also include consultations for rare oncologic tumors. Efforts to create partnerships with community medical providers and community leaders are underway to build relationships with patients and their communities. During the COVID-19 pandemic, the NCI Telehealth Concierge Service was expanded to support telehealth services for clinical trials, including remote consenting. This has continued postpandemic and is expected to contribute to addressing disparities by allowing decentralization and access into communities that would otherwise be limited because of travel, such as underrepresented minorities and participants in rural and underserved populations (36). The CCR is also modernizing the eligibility criteria for clinical trials toward less restriction for participants with brain metastases, abnormal laboratory values, comorbidities, and secondary cancers while critically evaluating clinical trial design for impediments to accrual (37).

Our work has limitations, including possible inaccuracies in the self-reported race and ethnicity data captured in the Biomedical Translational Research Information System and other examined databases, which highlight the need for ensuring correct and inclusive data collection relevant to racial, ethnic, and sex populations early in the development of clinical trials. Furthermore, our dataset does not capture the patients who were referred but only those who were enrolled, which limits our ability to fully explore the impact of demographic factors to participation. Additionally, although the SEER database includes approximately 48% of the US population (19), it does not include the District of Columbia, Maryland, and Virginia (Supplementary Table 7, available online), where close to half (48%) of the NCI-CC participants live (29); however, results from the wider-covering CiNA database were generally consistent with those from SEER.

To our knowledge, the current study is the first to examine representation by age, sex, race, and ethnicity in the population of participants in clinical trials at the NCI-CC. We show underrepresentation of older adults and women and Asians, Hawaiian or Pacific Islander, American Indian or Alaska Native, and Hispanic or Latino participants, but higher representation of Black or African American participants in comparison with the US cancer population, with recent improvement for all groups except women and additional differences in representation of race and ethnicity groups in comparison with other clinical trial mechanisms. Our findings can be partially attributed to demographics of the catchment area and systemic factors and can further inform the ongoing initiatives aimed at increasing diversity in cancer clinical trial enrollment at the NCI-CC. Further research is needed for improved understanding of the observed differences in representation of demographic groups and the impact of focused efforts required to assure equity in access to cancer clinical trials and advances in inclusive treatments.

Data availability

The data from National Cancer Institute underlying this article will be made available online in Zenodo, at https://dx.doi.org/10.5281/zenodo.8193221.

Author contributions

Nirmal Choradia, MD (Data curation; Formal analysis; Methodology; Visualization; Writing—original draft; Writing—review & editing), Fatima Karzai, MD (Methodology; Writing—review & editing), Ryan Nipp, MD, MPH (Writing—review & editing), Abdul Rafeh Naqash, MD (Writing—review & editing), James L. Gulley, MD, PhD (Methodology; Resources; Writing—review & editing), and Charalampos S. Floudas, MD, DMSc, MS (Conceptualization; Data curation; Formal analysis; Methodology; Writing—original draft; Writing—review & editing).

Funding

This work was supported by the Intramural Research Program, National Institutes of Health, National Cancer Institute, Center for Cancer Research.

Conflicts of interest

JG, who is a JNCI associate editor and co-author on this paper, was not involved in the editorial review or decision to publish the manuscript. The authors have no potential conflicts of interest to disclose.

Acknowledgements

The funder did not play a role in the design of the study; the collection, analysis, and interpretation of the data; the writing of the manuscript.

The interpretation and reporting of these data are the sole responsibility of the authors.

Results in this manuscript have been presented in part as a poster presentation at the American Society of Clinical Oncology 2022 Annual Meeting.