key: cord-0924664-2dfi8hbp authors: Unger, Joseph M; Hershman, Dawn L; Till, Cathee; Minasian, Lori M; Osarogiagbon, Raymond U; Fleury, Mark E; Vaidya, Riha title: “When Offered to Participate”: A Systematic Review and Meta-Analysis of Patient Agreement to Participate in Cancer Clinical Trials date: 2020-10-06 journal: J Natl Cancer Inst DOI: 10.1093/jnci/djaa155 sha: d8237a475bb3399524899605b1eb7bab142a9bf3 doc_id: 924664 cord_uid: 2dfi8hbp BACKGROUND: Patient participation in clinical trials is vital for knowledge advancement and outcomes improvement. Few adult cancer patients participate in trials. Although patient decision-making about trial participation has been frequently examined, the participation rate for patients actually offered a trial is unknown. METHODS: A systematic review and meta-analysis using 3 major search engines was undertaken. We identified studies from January 1, 2000, to January 1, 2020, that examined clinical trial participation in the United States. Studies must have specified the numbers of patients offered a trial and the number enrolled. A random effects model of proportions was used. All statistical tests were 2-sided. RESULTS: We identified 35 studies (30 about treatment trials and 5 about cancer control trials) among which 9759 patients were offered trial participation. Overall, 55.0% (95% confidence interval [CI] = 49.4% to 60.5%) of patients agreed to enroll. Participation rates did not differ between treatment (55.0%, 95% CI = 48.9% to 60.9%) and cancer control trials (55.3%, 95% CI = 38.9% to 71.1%; P = .98). Black patients participated at similar rates (58.4%, 95% CI = 46.8% to 69.7%) compared with White patients (55.1%, 95% CI = 44.3% to 65.6%; P = .88). The main reasons for nonparticipation were treatment choice or lack of interest. CONCLUSIONS: More than half of all cancer patients offered a clinical trial do participate. These findings upend several conventional beliefs about cancer clinical trial participation, including that Black patients are less likely to agree to participate and that patient decision-making is the primary barrier to participation. Policies and interventions to improve clinical trial participation should focus more on modifiable systemic structural and clinical barriers, such as improving access to available trials and broadening eligibility criteria. Indeed, a key concern has been the low rate of minority enrollment to clinical trials-especially pharmaceutical companysponsored trials-which may weaken confidence in the applicability of trial findings and demonstrates reduced access to potentially breakthrough treatments (11) (12) (13) . Given the layers of structural, clinical, and physician barriers to patient participation in clinical trials, most patients have very limited opportunity to even consider trial participation as an option for their cancer care. In this context, a key question is, what is the rate of trial participation among patients who are actually offered an opportunity to participate? The answer to this question is important for guiding the research and resources aimed at improving participation in clinical trials. To address this, we conducted a systematic review and metaanalysis synthesizing studies about patient participation in cancer trials published over the past 20 years. We identified studies that evaluated the participation of cancer patients in clinical trials. Studies focused on participation to either treatment trials or cancer control trials were included. Studies must have documented the number of patients offered trial participation and the number who enrolled (the denominator and numerator, respectively, for calculating study-specific rates). Studies were required to have been conducted in the United States. Studies examining individuals at risk of cancer (ie, screening studies) were excluded, as such individuals-in the absence of an actual cancer diagnosis-may have a qualitatively different attitude about study participation. Studies of patient-level interventions to improve the rate at which patients agree to participate in trials were excluded, based on the concern that agreement rates from these studies may not truly represent those commonly observed in trial recruitment. Studies utilizing patient navigators to facilitate enrollment to trials were similarly excluded, as were studies examining the intention to participate, rather than actual participation. Each included study provided a base case assessment of the number of patients offered a trial and the number enrolled. However, some modifications were made to emphasize the agency of patients in determining trial participation and to ensure consistency across the panel of included studies. Patients who were offered a trial but died before enrollment were excluded, because they were not at risk of trial participation. Similarly, patients offered a trial who did not participate because of physician decision or physician barriers were also excluded from the denominator. In contrast, patients reported as not enrolling because of receipt of supportive or palliative care were included if they were initially deemed eligible for trial participation. Further, patients reported as having been offered a trial but not enrolling because they did not return to the site, patients who were lost to follow-up, or patients who were considered to have been seeking a second opinion were included, based on the (conservative) assumption that these reasons are associated with passive refusal to participate in a trial. All other patients offered a trial who did not enroll were included in study-specific calculations. We conducted a computerized literature search using the PubMed, Web of Science, and Ovid Medline databases under Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines for articles published between January 1, 2000, and January 1, 2020 (20 years in total) (14) . We used the search terms "clinical trial accrual," "clinical trial enrollment," "enrollment in clinical trials," "clinical trial enrollment barriers," "patient participation in clinical trials," "patient decision making," or "participation factors" in combination with the term "cancer" (Supplementary Table 1 , available online). The search was conducted in January 2020. Titles, abstracts, and full studies were independently screened by 3 reviewers (RV, CT, and JMU). This reduced the opportunity for subjective interpretation of study-level results and better ensured consistent data collection. Differences between reviewers about the appropriateness of including particular studies were resolved by consensus. To limit the potential for publication bias, both published abstracts and full articles meeting inclusion criteria were included. Web of Science and Ovid Medline search results included published abstracts and posters in addition to full articles, whereas PubMed only included full articles. Treatment trials were those for which cancer patients received any kind of systemic (hormonal, cytotoxic, immunologic, targeted), radiation, or surgical cancer treatment. Cancer control studies included survivorship and symptom management studies. Studies were also described according to whether patients were treated at academic or community sites. Studies that were about participation in both treatment and cancer control trials, or included both academic and community sites, were grouped according to the category comprising more than 75% of patients; otherwise, the study was categorized as mixed. Studies were also described according to multiple design characteristics as follows: 1) requirement for patients to provide consent to study their trial participation decision-making vs not; 2) reliance on patient report of trial participation vs physician report or abstraction from the medical record; and 3) prospective vs retrospective data collection. Race and ethnicity groups included the mutually exclusive categories White, Black, Hispanic, and Asian. We used meta-analysis for single proportions using the R-package "metaphor" (15, 16) . Forest plots were used to summarize individual study effects. Both fixed and random effects approaches were considered for deriving summary rates. The use of fixed effects is predicated on the idea that effect size differences are assumed to vary because of sampling error only; in this case, summary measures are simply weighted by study sample size (17, 18) . We tested this assumption using the Q statistic (to assess betweenstudy heterogeneity) and the I 2 statistic (to assess the proportion of total variation in study estimates due to study heterogeneity) (19) (20) (21) . A statistically significant Q statistic or an I 2 statistic greater than 50% suggests that a random effects approach, which accounts for both within-and between-study variation, is preferable (17, 18, 21, 22) . A restricted maximumlikelihood estimator of the between-study variance was used (23, 24) . We used meta-analysis for single proportions to derive the rate of trial participation among all studies. We also used meta-regression techniques for moderator analyses to compare the rates of trial participation between patients in treatment vs cancer control studies. The absence of statistically significant evidence of a difference in rates between treatment and cancer control studies provided the rationale to aggregate across all included studies when deriving an overall rate of patient participation, as well as rates by cancer care setting and by race and ethnicity. Analyses were also conducted separately within studies about treatment and cancer control, because patients in the treatment vs the survivorship or symptom management phase of a cancer diagnosis may have qualitatively different expectations about the value of participating in a study. Patients who consent to participate in a secondary study about trial decision-making may be more likely to ultimately agree to participate in a clinical trial. To address this, we compared estimates of trial participation between studies requiring vs not requiring consent using meta-regression techniques for moderator analyses (25) . If a statistically significant difference was evident, analyses were conducted separately within groups defined by this variable; otherwise, results were aggregated across all studies. Moderator analyses were also used to assess whether rates differed by community vs academic study setting, given the different commitment to trial research that is prevalent between these care settings and because many more patients (generally estimated at 80% or higher) are treated at community sites (26) (27) (28) . Additionally, we examined whether the results differed by prospective vs retrospective design and by the source of report of trial participation (patient vs physician). Rates of trial participation among race and ethnicity groups were also calculated using meta-analysis for single proportions (15, 16) . Additionally, we tested whether rates of participation for Black, Hispanic, and Asian patients differed from rates for White patients among studies that provided data on participation rates for both the minority race or ethnicity group and White patients. The odds ratio of trial participation (minority group vs White) was estimated and tested whether it was different from 1.0 (15) . All statistical tests were 2-sided. P values of less than .05 were considered statistically significant. We examined whether the findings were sensitive to the influence of individual studies by iteratively excluding each study and recalculating the overall estimates (a "leave one out" analysis). To assess whether patterns of agreement to participate changed over calendar time, we conducted a simple linear regression of the study-specific estimates, indexed by the median year of the specified time period of study conduct, and tested whether the slope of the regression coefficient differed from zero. For studies that did not specify the year(s) of conduct, mean imputation was used, using the mean of the difference between study publication year and specified years of conduct among studies with known data. We also used the Begg rank correlation test to identify any evidence of publication bias using the ranktest function in R (15, 29) . Although the goal was to derive a representative estimate of trial participation rates among those offered a trial, the rate depends in part on analysis assumptions. In sensitivity analysis, we examined the potential lower and upper bounds on the estimate of the trial participation rate by ignoring all anticonservative and conservative assumptions about patient-level exclusions or inclusions, respectively. To estimate the potential lower bound, we included all patients in study-specific denominators who were explicitly indicated in the study publications as having been offered a trial, even if the patients did not meet our definition of being at risk of trial participation. These included patients who died, had no trial available, were ineligible, or did not participate because of physician decision. Conversely, to establish the potential upper bound on the trial participation estimate, we excluded from the study-specific denominators all patients who did not return or were lost to follow-up, because these patients may have participated in a trial elsewhere. Further, we retained the exclusion of patients who died, but also excluded from the denominator patients in supportive or hospice care based on the idea that such patients are at minimal risk of trial participation. Reasons for nonparticipation among patients who did not enroll in clinical trials were described. To enable calculations across studies, category totals for studies that allowed more than 1 reason to be reported for nonenrollment were prorated so that the total number of reasons equaled the number of patients not enrolled. Overall, 4073 studies were flagged by the 3 search engines. Of the studies, 830 studies were duplicates, and 2 studies had mischaracterized publication years that fell outside the prespecified time interval for study inclusion; these records were excluded, leaving 3241 unique studies ( Figure 1 ). Title and abstract review of the 3241 studies yielded 60 potentially relating to participation decision-making for cancer clinical trials. Full articles for these 60 studies were reviewed. Twenty-five were excluded, primarily because the studies included interventions to increase clinical trial accrual or were conducted in a non-US setting ( Figure 1 ; Supplementary Table 2, available online). Thirty-five studies comprised of 9759 patients met our inclusion criteria ( Figure 2 ) (8, 9, . Most studies (n ¼ 30, 85.7%) focused on treatment for cancer (Table 1) (8, 9, (30) (31) (32) 34, 35, 37, 39, 40, (42) (43) (44) (45) (46) (47) (48) (49) (50) (51) (52) (53) (54) (55) (57) (58) (59) (60) (61) (62) . Among the 5 cancer control studies, 4 focused on enrollment to cancer survivorship studies and 1 to a symptom management study ( From 3 studies, 47 patients were identified as not enrolling because of death; these patients were excluded from the studyspecific denominators ( Figure 2 ) (41, 42, 51) . Additional exclusions from the study-specified base case denominators of patients offered trial participation included 87 patients from 2 studies who did not participate because of physician barriers (8,55); 39 patients from 3 studies because of ineligibility (39, 42, 44) ; 14 from 1 study because of lack of trial availability (44) ; and 4 patients from 2 studies because of missing enrollment data (39, 48) . One study excluded 7 patients who were offered a trial but could not be recontacted; these patients were considered passive refusers and were added to the studyspecific denominator for purposes of this analysis (35) . Study-specific estimates are shown in their entirety in Figure 2 . Both the estimated Q (512.4, P < .001) and the I 2 (96.4%) statistics indicated a high degree of heterogeneity across the studies, justifying the use of a random effects model. The overall rate of participation in either treatment or cancer control trials among patients offered participation was 55.0% (95% confidence interval [CI] ¼ 49.4% to 60.5%; Figure 2 ). There were no statistically significant differences in trial participation rates between studies about trial participation that required patient consent (59.9%, 95% CI ¼ 51.0% to 68.5%) compared with studies not requiring patient consent (52.0%, 95% CI ¼ 45.0% to 59.0%; P ¼ .17). Thus, analyses were not reported separately by this variable. Trial participation rates were statistically significantly higher at academic centers (58.4%, 95% CI ¼ 52.2% to 64.5%) vs community centers (45.0%, 95% CI ¼ 34.5% to 55.7%; P ¼ .04). In contrast, there were no differences in trial participation rates between studies with prospective (51.7%, 95% CI ¼ 43.8% to 59.6%) vs retrospective (58.1%, 95% CI ¼ 50.4% to 65.6%; P ¼ . 26) designs or between studies based on patient report (65.7%, 95% CI ¼ 49.8% to 80.0%) vs physician or staff report (53.6%, 95% CI ¼ 47.7% to 59.4%; P ¼ .16) of trial participation status. Among the 30 studies about patient participation in treatment trials (comprised of n ¼ 7915 patients), the rate at which patients participated if a trial was offered was 55.0% (95% CI ¼ 48.9% to 60.9%). The rate of trial participation was marginally statistically significantly higher in patients receiving care at academic centers (58.1%, 95% CI ¼ 51.5% to 64.6%) compared with community centers (44.5%, 95% CI ¼ 32.4% to 56.8%; P ¼ .06). Among the 5 studies about patient participation in cancer control studies (comprised of n ¼ 1844 patients), the overall rate was 55.3% (95% CI ¼ 38.9% to 71.1%). The rate of trial participation trended higher in patients participating in cancer control studies at academic centers (61.3%, 95% CI ¼ 39.0% to 81.4%) compared with community centers (46.5%, 95% CI ¼ 21.1% to 72.9%), although this difference was not statistically significant The participation rates for treatment trials and cancer control studies were not statistically significantly different (P ¼ .98). In the 15 studies that provided data to estimate rates among Black patients offered trial participation (Table 3) , Black patients agreed to participate 60.4% of the time (95% CI ¼ 49.5% to 70.8%; Table 4 ). In the 13 studies that contained data on agreement to participate for both Black and White patients, Black patient participation was slightly higher (58.4%, 95% CI ¼ 46.8% to 69.7%) than White patient participation (55.1%, 95% CI ¼ 44.3% to 65.6%), although the odds of trial participation did not statistically significantly differ between Black vs White patients (odds ratio [OR] ¼ 1.01, 95% CI ¼ 0.90 to 1.13; P ¼ .88). Results were similar in studies about treatment trial participation only (Table 4 ). Similar patterns of higher, but non-statistically significant, rates of participation were evident for Hispanic patients and Asian patients compared with White patients (Table 4 ). For each of Black, Hispanic, and Asian patient groups, rates of participation trended higher at academic compared with community centers; differences were statistically significant among Hispanic patients (P ¼ .04) and especially among Asian patients (P < .001). Half (15 of 30) of the studies about treatment trial participation-comprising 2626 patients-provided reasons for nonenrollment (Table 3) . Treatment-related concerns were most commonly indicated as reasons for nonenrollment, variously described as desire for other treatment, desire to choose own treatment or to avoid protocol treatment, or preference for standard treatment (24.4%). A large portion of patients indicated (19.9%) . Passive refusal to participate-expressed through not returning to the clinic or being lost to follow-up-was the reason 8.3% of patients did not enroll. Other common reasons included fear of side effects (7.9%), financial concerns or insurance denial (6.7%), and a dislike of participating in an experiment, including dislike of having treatment determined by random assignment (6.6%). All 5 cancer studies about participation in cancer control trials provided known reasons for nonparticipation on 959 patients (Table 3) . Not returning to the clinic or being lost to follow-up was the reason nearly half (49.4%) of patients did not participate. Other common reasons for nonenrollment included a dislike of participating in an experiment (12.6%) and lack of interest (11.9%). Travel distance was indicated as a reason for nonenrollment for 4.2% of patients considering a treatment trial and 5.4% of patients considering a cancer control trial. When individual studies were iteratively excluded, in no case did the percentage estimate change by more than 1.2% for all studies combined (primary estimate, 55.0%; range ¼ 54.0%-56.1%) and for the treatment studies (primary estimate, 55.0%; range ¼ 53.8%-56.2%; Figure 3 ). Given fewer available studies, the exclusion of individual cancer control studies resulted in percentage estimate change of up to 6.0% for the overall cancer control estimate (primary estimate, 55.3%; range ¼ 50.9%-61.3%). This analysis indicates that the estimates for all studies combined, treatment trials, and cancer control studies are internally robust. An examination of study-specific estimates suggests that rates of agreement to participate in trials have trended higher in more recent years ( Figure 4 ) for all studies combined (P ¼ .008), treatment trials (P ¼ .007), and cancer control studies (P ¼ .02). The rate of participation was greater among studies with median enrollment year after 2010 (65.5%, 95% CI ¼ 56.0% to 74.4%) vs 2010 or before (50.7%, 95% CI ¼ 44.5% to 56.8%; P ¼ .01). In a sensitivity analysis examining the potential impact of study assumptions, the potential lower bound on the estimated rate of trial participation was 53.4% (95% CI ¼ 48.2% to 58.7%), and the potential upper bound was 57.8% (95% CI ¼ 52.1% to 63.3%). We found no evidence of publication bias using the rank correlation test (P ¼ .24). We found that at least half of patients offered participation in a cancer clinical trial did participate. The findings did not differ between treatment and cancer control trials. Importantly, Black, Hispanic, and Asian patients participated in trials at rates at least as high as White patients. Moreover, the rate of trial participation among those offered a trial may have increased over time. These findings dramatically underscore the willingness of cancer patients to participate in a trial if one is offered. The findings also stand in stark contrast to the commonly cited statistic that only 5% of adult cancer patients participate in trials, a statistic that fails to reflect the many structural and clinical hurdles that stand in the way of trial participation for most patients. Because patients ultimately decide whether to participate in a trial, it is critical to understand why they choose to participate or not. In the studies included in this analysis, the most common reason for not enrolling in a treatment trial was the desire among patients to control their treatment choice, including by avoiding protocol treatment side effects and by avoiding participation in an experiment where treatment may be randomly assigned. Many patients also explicitly (to researchers) or implicitly (through passive refusal, eg, by not following up) expressed a lack of interest in trial participation. Together, these reasons underlay the decision for nearly 7 out of every 10 (69.0%) patients who chose not to participate in either treatment or cancer control studies. An important consideration for researchers and policy makers is understanding the extent to which reasons for nonparticipation in trials are modifiable, as such reasons may be amenable to interventions or policy changes. A patient's desire to control his or her treatment choice or a lack of interest in study participation are unlikely to be easily modifiable; moreover, attempting to do so may tread on the patient protections against undue influence articulated in the US Common Rule for the Protection of Human Subjects. In contrast, other (although less frequent) reasons expressed by patients for nonparticipation may in fact be explicitly addressable through policy. For instance, some patients indicated concerns about finances or insurance. Medicare covers the routine care costs of clinical trial participation, as do many private insurance carriers. State Medicaid programs, in contrast, do not uniformly provide coverage for clinical trials, and coverage provisions in general are highly variable (63, 64) . To address this, legislation currently before Congress would mandate that all state Medicaid programs cover the routine care costs of cancer clinical trials (65). Patients also cited the burden of travel as a barrier. Travel distance may be especially problematic for socioeconomically and geographically disadvantaged populations lacking more proximal access to academic cancer centers where trial conduct is more common (66) (67) (68) . Health-care models that virtually link local providers with oncology specialists could help alleviate the need for cancer patients to travel great distances for care (69) . The recently accelerated adoption of telemedicine approaches (including remote consent and virtual visits) in response to the COVID-19 pandemic can ease the burden of trial participation for cancer patients and, if made permanent, may improve access to trials for patients over the long term (70) (71) (72) . More broadly, external advisory groups, especially that include patient advocates, could help researchers design trials that more readily incorporate elements to make trial participation more attractive to patients (73, 74) . Also includes "did not return" and "second opinions." c One patient who was offered a trial did not participate because of death. This patient was excluded from the study denominator. In total, 47 were not enrolled, but those with trial unavailable (14) , ineligible (8) , and unable to enroll because of paclitaxel shortage (1) were excluded because these were not patient choice factors, leaving n ¼ These concrete steps to improve access to trials for those offered participation are necessary. But only a small portion of patients are offered trial participation, so even very successful strategies will have only limited impact on overall trial participation rates. A much greater impact may be achieved by addressing the numerous and sizeable hurdles to trial participation that occur prior to the physician-patient interaction. Structural barriers to the conduct of trials are endemic in the United States (5) . Clinical trial conduct is a major undertaking for institutions, requiring a commitment of resources that are often poorly reimbursed, especially for nonpharmaceutical company-sponsored trials (6) . Thus, for the majority of patients, no protocol is locally available (5) . In response, governmentsponsored trial mechanisms, such as the National Cancer Institute's Community Oncology Research Program, were designed specifically to enable the conduct of trials outside major academic centers, with notable success in extending the reach and inclusivity of trials (13, 75, 76) . Clinical trial matching services provide clinicians and patients the opportunity to identify clinical trials for which they are potentially eligible. These services have struggled to provide complete and reliable targets, although efforts to standardize and improve these services are ongoing (77) . Even when a trial is available, patients are frequently ineligible. The recognition that trial eligibility criteria are overly restrictive, with limited safety or research benefit, motivated an extensive effort by the American Society for Clinical Oncology, Friends of Cancer Research, and the US Food and Drug Administration to modernize eligibility criteria (7, 78) . One recent study estimated that adoption of these changes to eligibility could generate more than 6000 new registrations to cancer clinical treatment trials annually (79) . Another study estimated that the expanded criteria would double the number of nonsmall cell lung cancer patients eligible for trial participation (80) . Together, these structural and clinical barriers exclude 3 out of every 4 patients. Because many aspects of these barrier domains are potentially modifiable, mitigating these barriers represents an enormous opportunity to increase trial participation rates. We also found that Black, Hispanic, and Asian patients enrolled at rates that were very comparable to rates for White patients. This observation seems surprising given the repeated observations that minority patients are underrepresented in clinical cancer research (2, 12) . Yet, the finding is consistent with other studies showing similarity by race in the willingness to participate in trials if asked (81) (82) (83) . It also strongly suggests that observed racial and ethnic disparities in trial participation manifest earlier in the treatment decision-making process, perhaps because of differential likelihood of meeting restrictive eligibility criteria (55), differential access to cancer centers where clinical research is conducted (84, 85) , and differential access to physicians who offer clinical trials (58) . Indeed, this finding indicates that perhaps the best way to improve enrollment of minority patients to cancer trials is simply to ensure that minority patients are invited to participate. The recognition of this may inform efforts to alleviate potential bias in the provision of health-care resources by race or ethnicity, including trial offers for eligible patients (58, 86, 87) . One concern about conducting secondary studies about patient agreement to participate in clinical trials is that the process of seeking consent for the secondary study is more likely to bias the samples in favor of patients willing to participate in research more generally, including in clinical trials, which could generate an inflated estimate of the rate of clinical trial participation. Recognizing this, many studies about clinical trial decision-making sought waivers of consent. Regardless, we found no statistically significant difference in trial participation rates between studies that did vs did not require consent. Also, the review was limited by the fact that not all studies provided data on enrollment by race and ethnicity. Additionally, estimates of agreement to participate may be biased high if the number of individuals offered a trial was undercounted, although the limited evidence available to examine this suggested a tendency to overestimate the number of individuals at risk of trial participation by including patients who did not participate because of physician or eligibility barriers or lack of trial availability. Further, there is a possibility that publication bias or missed studies could influence the results. Our anticipation is that the influence of one or more missed studies would likely be nominal given the comprehensive search procedures that included abstracts as well as full articles; the fact that 35 studies were included, such that the inclusion of any single study in a random effects model is unlikely to substantially alter the results; and the existing lack of evidence of publication bias based on the Begg rank correlation test. The evaluation of reasons for nonparticipation was a secondary endpoint and was based on the included studies only, rather than on a comprehensive review of the literature about reasons for nonparticipation. Thus, the estimates derived from this component of the analysis may not have been representative and may also have missed some known reasons for nonparticipation that have been previously identified, such as concerns about the consent process or time and effort to participate in a trial (11, 61, 88) . Finally, only 5 studies examined participation in cancer control studies, limiting confidence in the conclusions that can be drawn about participation patterns in this research setting. The findings of this review indicate that patients choose to participate in clinical trials more than half the time when offered the opportunity, irrespective of race and ethnicity. This suggests that the root cause of low trial participation rates in adults with cancer is a clinical trial system beset with structural and clinical barriers, rather than patient disinterest. Research, interventions, and policies to improve trial participation should focus more on these systemic structural and clinical barriers. This work was supported by a grant from the Hope Foundation for Cancer Research (Unger) . Transforming Clinical Research in the United States: Challenges and Opportunities: Workshop Summary Participation in cancer clinical trials: race-, sex-, and age-based disparities How sociodemographics, presence of oncology specialists, and hospital cancer programs affect accrual to cancer treatment trials Representation of African-Americans, Hispanics, and whites in National Cancer Institute cancer treatment trials Systematic review and metaanalysis of the magnitude of structural, clinical, and physician and patient barriers to cancer clinical trial participation What keeps patients out of clinical trials? Broadening eligibility criteria to make clinical trials more representative: American Society of Clinical Oncology and Friends of Cancer Research Joint Research Statement Clinical trial accrual among new cancer patients at a community-based cancer center A prospective analysis of the influence of older age on physician and patient decision-making when considering enrollment in breast cancer clinical trials (SWOG S0316) Use of the National Cancer Institute Community Cancer Centers Program screening and accrual log to address cancer clinical trial accrual Barriers to recruiting underrepresented populations to cancer clinical trials: a systematic review Disparity of race reporting and representation in clinical trials leading to cancer drug approvals from Representativeness of black patients in cancer clinical trials sponsored by the National Cancer Institute compared to pharmaceutical companies Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement Conducting meta-analyses in R with the metaphor package How to conduct a meta-analysis of proportions in r: a comprehensive tutorial A basic introduction to fixed-effect and random-effects models for meta-analysis A comparison of statistical methods for meta-analysis The combination of estimates from different experiments Meta-analysis of prevalence Meta-analysis in clinical trials revisited Quantifying heterogeneity in a meta-analysis A comparison of heterogeneity variance estimators in simulated random-effects meta-analyses Bias and efficiency of meta-analytic variance estimators in the random-effects model Introduction to Meta-Analysis Impact of the National Cancer Institute Community Cancer Centers Program on clinical trial and related activities at a community cancer center in rural Nebraska A community cancer center program: getting to the next level Risk adjusting survival outcomes in hospitals that treat patients with cancer without information on cancer stage Operating characteristics of a rank correlation test for publication bias Barriers exist to patient participation in clinical trials Clinical trial accrual patterns for radiation oncology patients at an academically based tertiary care medical center Abstract P5-13-06: Analysis of barriers to clinical trial accrual in an NCI-designated comprehensive cancer center: Results of identifying clinical trial gaps Tracking clinical trial accrual strategies and barriers via a Web-based screening tool Enrollment of African Americans onto clinical treatment trials: study design barriers Influence of clinical communication on patients' decision making on participation in clinical trials Barriers to recruitment and adherence in a randomized controlled diet and exercise weight loss intervention among minority breast cancer survivors Barriers to enrollment in non-small cell lung cancer therapeutic clinical trials Barriers and facilitators to recruitment to a culturally-based dietary intervention among urban Hispanic breast cancer survivors Factors and outcomes of decision making for cancer clinical trial participation Patient and physician factors associated with participation in cervical and uterine cancer trials: an NRG/GOG247 study Recruitment of low income, predominantly minority cancer survivors to a randomized trial of the I Can Cope cancer education program Barriers to study enrollment in patients with advanced cancer referred to a phase I clinical trials unit Patients' perceptions of physicians communication and outcomes of the accrual to trial process Factors influencing clinical trial enrollment among ovarian cancer patients Identifying barriers associated with enrollment of patients with lung cancer into clinical trials Evaluating the decisions of glioma patients regarding clinical trial participation: a retrospective single provider review Removing barriers to participation in clinical trials, a conceptual framework and retrospective chart review study Barriers to clinical trial participation by older women with breast cancer Comprehension of randomization and uncertainty in cancer clinical trials decision making among rural, Appalachian patients Racial/ethnic differences in clinical trial enrollment, refusal rates, ineligibility, and reasons for decline among patients at sites in the National Cancer Institute's Community Cancer Centers Program Prospective evaluation of cancer clinical trial accrual patterns: identifying potential barriers to enrollment Analysis of factors affecting successful clinical trial enrollment in the context of three prospective, randomized, controlled trials An evaluation of barriers to accrual in the era of legislation requiring insurance coverage of cancer clinical trial costs in California An assessment of age and other factors influencing protocol versus alternative treatments for patients with epithelial ovarian cancer referred to member institutions: a Gynecologic Oncology Group study Barriers to therapeutic clinical trials enrollment: differences between African-American and white cancer patients identified at the time of eligibility assessment Recruitment and retention challenges in breast cancer survivorship research: results from a multisite, randomized intervention trial in women with early stage breast cancer Factors that predict the referral of breast cancer patients onto clinical trials by their surgeons and medical oncologists Factors associated with breast cancer clinical trials participation and enrollment at a large academic medical center Enrollment onto breast cancer therapeutic clinical trials: a tertiary cancer center experience Overcoming barriers to cancer clinical trial accrual: impact of a mass media campaign Patient income level and cancer clinical trial participation Participation and survival of geriatric patients in phase I clinical trials: the Karmanos Cancer Institute (KCI) experience Unsettling scores: a ranking of state Medicaid programs Medicaid: Ensuring Access to affordable health care coverage for lower income cancer patients and survivors At what cost to clinical trial enrollment? A retrospective study of patient travel burden in cancer clinical trials Spatial analysis of adherence to treatment guidelines for advanced-stage ovarian cancer and the impact of race and socioeconomic status Invasive cancer incidence Systematic review of cancer treatment programmes in remote and rural areas regulatory-information/search-fda-guidance-documents/fda-guidance-conduct-clinical-trials-medical-products-during-covid-19-pandemic National Cancer Institute. Coronavirus guidance Oncology practice during the COVID-19 pandemic Evaluating patient and stakeholder engagement in research: moving from theory to practice Integrating comparative effectiveness design elements and endpoints into a phase III, randomized clinical trial (SWOG S1007) evaluating oncotypeDX-guided management for women with breast cancer involving lymph nodes Geographic distribution and survival outcomes for rural patients with cancer treated in clinical trials Letter to Dr. Normal E. Sharpless, Director, National Cancer Institute. Re: NCI Request for Information, Strategies for Matching Patients to Clinical Trials (NOT-CA-18-063) ASCO in Action: initiative to modernize eligibility criteria for clinical trials launched Association of patient comorbid conditions with cancer clinical trial participation Impact of broadening clinical trial eligibility criteria for advanced non-small cell lung cancer patients: real-world analysis Participation in cancer clinical trials: why are patients not participating? The impact of socioeconomic status and race on trial participation for older women with breast cancer Are racial and ethnic minorities less willing to participate in health research Geographic proximity and racial disparities in cancer clinical trial participation Racial and ethnic disparities in the receipt of cancer treatment Barriers to clinical trial enrollment in racial and ethnic minority patients with cancer Unequal treatment: confronting racial and ethnic disparities in health care Accrual to cancer clinical trials: directions from the research literature The data underlying this article are available in the article and in its online Supplementary Material. Role of funder: The funders had no role in the design of the study; the collection, analysis, and interpretation of the data; the writing of the manuscript; and the decision to submit the manuscript for publication.Disclosures: The authors report the following relationships or conflicts of interest: Hershman, DL: Consulting or advisory role (AIM Specialty Health). Till, C (an immediate family member): Patents, royalties, other intellectual property (Roche/Genentech; MustangBio). Osarogiagbon, RU: Consulting or advisory role (Association of Community Cancer Centers; AstraZeneca; American Cancer Society); patents, royalties, other intellectual property (2 US and 1 China patents for lymph node specimen collection kit and method of pathologic evaluation); stock and other ownership interests (Lilly; Pfizer; Gilead Sciences); other relationship (Oncobox). Fleury ME: Research Funding (IQvia; Merck; Genentech). No other relationship were reported.Disclaimer: The content is solely the responsibility of the authors and does not necessarily represent the official views of the study sponsor.