Demographic and Health Behavior Factors Associated With Clinical Trial Invitation and Participation in the United States

Key Points

Question  What person-level factors are associated with US adults’ invitation to and participation in clinical trials?

Findings  In this cross-sectional study of 3689 adults, 9% were invited to participate in a clinical trial, and of those, 47% participated. Respondents had higher odds of clinical trial invitation if they were non-Hispanic Black, college educated, single, or urban-dwelling or had medical conditions; non-Hispanic Black respondents had lower odds of clinical trial participation.

Meaning  In this study, clinical trial invitation and participation differed by person-level demographic and clinical characteristics, reinforcing the need for strategies encouraging generalizable and equitable translation of research to practice.

Importance  Representative enrollment in clinical trials is critical to ensure equitable and effective translation of research to practice, yet disparities in clinical trial enrollment persist.

Objective  To examine person-level factors associated with invitation to and participation in clinical trials.

Design, Setting, and Participants  This cross-sectional study analyzed responses from 3689 US adults who participated in the nationally representative Health Information National Trends Survey, collected February through June 2020 via mailed questionnaires.

Exposures  Demographic, clinical, and health behavior–related characteristics.

Main Outcomes and Measures  History of invitation to and participation in a clinical trial, primary information sources, trust in information sources, and motives for participation in clinical trials were described. Respondent characteristics are presented as absolute numbers and weighted percentages. Associations between respondent demographic, clinical, and health behavior–related characteristics and clinical trial invitation and participation were estimated using survey-weighted logistic regression models.

Results  The median (IQR) age of the 3689 respondents was 48 (33-61) years, and most were non-Hispanic White individuals (2063 [59%]; non-Hispanic Black, 452 [10%]; Hispanic, 521 [14%]), had more than a high school degree (2656 [68%]), were employed (1809 [58%]), and had at least 1 medical condition (2535 [61%]). Overall, 439 respondents (9%) had been invited to participate in any clinical trial. Respondents with increased odds of invitation were non-Hispanic Black compared with non-Hispanic White (adjusted odds ratio [aOR], 1.85; 95% CI, 1.13-3.02), had greater than a high school education compared with less than high school education (eg, ≥college degree: aOR, 4.84; 95% CI, 1.89-12.39), were single compared with married or living as married (aOR, 1.68; 95% CI, 1.04-2.73), and had at least 1 medical condition compared to none (eg, 1 medical condition: aOR, 2.25; 95% CI, 1.32-3.82). Respondents residing in rural vs urban areas had 77% decreased odds of invitation to a clinical trial (aOR 0.33; 95% CI 0.17-0.65). Of invited respondents, 199 (47%) participated. Compared with non-Hispanic White respondents, non-Hispanic Black respondents had 72% decreased odds of clinical trial participation (aOR, 0.28; 95% CI, 0.09-0.87). Respondents most frequently reported “health care providers” as the first and most trusted source of clinical trial information (first source: 2297 [59%]; most trusted source: 2597 [70%]). The most frequently reported motives for clinical trials participation were “wanting to get better” (2294 [66%]) and the standard of care not being covered by insurance (1448 [41%]).

Conclusions and Relevance  The findings of this study suggest that invitation to and participation in clinical trials may differ by person-level demographic and clinical characteristics. Strategies toward increasing trial invitation and participation rates across diverse patient populations warrant further research to ensure equitable translation of clinical benefits from research to practice.

Evidence-based evaluation of new health care treatments and interventions via clinical trials is critical for improving individual and population health. As of April 2021, ClinicalTrials.gov listed 291 087 ongoing trials, including 158 827 drug or biologic interventions and 96 449 behavioral interventions.^1,2 However, clinical trial participants infrequently represent the real-world population in which the intervention will be applied.^3,4 Several patient population subgroups remain underrepresented in trials, including older adults, people who belong to racial or ethnic minority groups, persons with multiple comorbid conditions, and individuals residing in rural locations.^5-7 Disparities between trial participants and real-world populations are problematic, since results from these trials provide the evidence base for real-world clinical practice guidelines and public health services.⁸

Previous research has shown that person-level factors, such as sociodemographic characteristics, may affect whether individuals are invited to participate in clinical trials, while motivations and beliefs may affect decisions to participate if invited.^9-14 Much of the existing literature on trials tends to be condition-specific, focused on participation rather than invitation, and specific to subpopulations of the US public. To build on the existing literature base, investigation into the larger US public’s recent invitation to, participation in, and motivations surrounding clinical trials is needed. This study seeks to understand person-level factors associated with invitation to and participation in clinical trials in a large, nationally representative sample of US adults. We also describe respondent primary trial information sources, trust in information sources, and motives for participation in clinical trials to inform future efforts addressing disparities in clinical trial invitation and participation.

This cross-sectional study used data from the National Cancer Institute’s Health Information National Trends Survey (HINTS) 5 Cycle 4, collected February through June 2020 via mailed questionnaires. HINTS is a publicly available, cross-sectional, nationally representative survey of civilian, noninstitutionalized US adults that assesses knowledge of, attitudes toward, and use of health-related information.^15,16 HINTS 5 Cycle 4 included 7 new questions regarding clinical trials (eFigure in the Supplement).¹⁵ Initial constructs and associated items were included based on published literature representing a range of clinical conditions,^9,17 racial and ethnic groups,¹⁷ and nationalities.^17-19 Two rounds of cognitive testing²⁰ were conducted with a diverse sample of 30 individuals (15 per round). Results informed the refinement of items and response options for inclusion in the survey. A plain language definition of clinical trials with 2 examples was included to standardize respondent interpretation and to specify that trials include drug and behavioral interventions. HINTS data are deidentified and thus exempt from review by the US National Institutes of Health Office of Human Subjects Research Protections. Respondents provided consent to participate in HINTS. This study follows the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guidelines.²¹

Invitation to a clinical trial was assessed with a single yes or no item, “Have you ever been invited to participate in a clinical trial?” If the respondent answered yes, they were asked the single yes or no item, “Did you participate in the clinical trial?”

Respondents were asked about their current knowledge of clinical trials, where they would first go to get information about a hypothetical trial, and who they would most trust as a source of information about a hypothetical trial. Respondents were also asked to rate factors (eg, helping others, payment for participation) that would influence their decision to participate in a hypothetical clinical trial on a 4-point scale (a lot, somewhat, a little, not at all). Influential factors in trial decision-making were dichotomized as a lot vs the combined scores of somewhat, a little, and not at all.²² Separated scores for all response options are shown in eTable 1 in the Supplement.

Self-reported demographic, clinical, and health behavior–related respondent characteristics were included from the HINTS survey data for descriptive comparison and multivariable analysis. Demographic variables included age, sex, race and ethnicity (race: White, Black or African American, American Indian or Alaska Native, Asian Indian, Chinese, Filipino, Japanese, Korean, Vietnamese, other Asian, Native Hawaiian, Guamanian or Chamorro, Samoan, other Pacific Islander; ethnicity: not of Hispanic, Latino/a, or Spanish origin; Mexican, Mexican American, Chicano/a; Puerto Rican; Cuban; another Hispanic, Latino/a, or Spanish origin), education, feelings about present income, marital status, rural or urban residence,²³ Census region, employment status, and health insurance status. Clinical variables included number and type of diagnosed medical conditions and self-reported health. Health behavior–related variables included smoking status, hazardous drinking,²⁴ weekly exercise, and annual health care visits.

All analyses were weighted to provide nationally representative estimates using complex survey methodology with jackknife replicate weights for accurate standard errors.¹⁵ Mean differences, or effect sizes, were calculated using unweighted Cohen d or Cramer V to determine the magnitude of associations in bivariate associations.²⁵ Associations between respondent demographic, clinical, and health behavior–related characteristics and clinical trial invitation and participation were estimated using adjusted odds ratios (aOR) and 95% CIs from exploratory survey-weighted logistic regression models. All covariables were included in the model exploring clinical trial invitation. Due to reduced sample size limiting power to determine magnitude of bivariate associations in the model estimating clinical trial participation, covariables were restricted based on past research demonstrating differential rates of clinical trial participation associated with age, race and ethnicity, sex, education, income, urban or rural residence, health insurance status, and health status.^26-30 Covariable multicollinearity was examined using the variance inflation factor. Sensitivity analyses were conducted excluding respondents who reported knowing nothing about clinical trials. Trial participation motives, information seeking, and trust were described by (1) the overall sample, (2) respondents who participated in a clinical trial, and (3) respondents who were invited but did not participate in a clinical trial. All analyses were performed using SAS version 9.4 (SAS Institute, Cary, NC).

A total of 15 347 HINTS 5 Cycle 4 surveys were mailed. Of mailed surveys, 3977 were returned, 3865 were eligible for analysis (weighted response rate, 37%), and 3689, with complete information on clinical trial items, were included in our sample. The 176 respondents excluded because of missing trial invitation status more frequently had less education (eTable 2 in the Supplement). Of the 3689 included respondents, half were female (2107 [50%]), the median (IQR) age was 48 (33-61) years, and most were non-Hispanic White (2063 [59%]; non-Hispanic Black, 452 [10%]; Hispanic, 521 [14%]), had more than a high school degree (2656 [68%]), were employed (1809 [58%]), and had at least 1 medical condition (2535 [61%]) (Table 1).

Overall, 439 respondents (9%) reported being invited to participate in a clinical trial (Table 1). In exploratory models, respondent factors associated with increased odds of reported clinical trial invitation included being non-Hispanic Black compared with non-Hispanic White (aOR, 1.85; 95% CI, 1.13-3.02), having some college education or a college degree or higher compared with less than high school education (some college: aOR, 2.48; 95% CI, 1.07-5.77; ≥college degree: aOR, 4.84; 95% CI, 1.89-12.39), being single compared with married or living as married (aOR, 1.68; 95% CI, 1.04-2.73), and having at least 1 more medical condition compared with none (1 condition: aOR, 2.25; 95% CI, 1.32-3.82; 2 conditions: aOR, 2.67; 95% CI, 1.56-4.57; ≥3: aOR, 3.76; 95% CI, 2.01-7.03) (Table 2). Conversely, respondents residing in a rural area had 77% decreased odds of invitation to a clinical trial than those residing in an urban area (aOR 0.33; 95% CI 0.17-0.65).

Of those invited to a clinical trial, 199 respondents (47%) reported participating (Table 3). In models estimating odds of reported clinical trial participation, respondents who were non-Hispanic Black compared with non-Hispanic White had 72% decreased odds of participation after invitation (aOR, 0.28; 95% CI, 0.09-0.87) (Table 2). In sensitivity analyses excluding 1345 respondents who reported knowing nothing about clinical trials, education and marital status were no longer associated with trial invitation, while dual-eligible beneficiaries had higher odds of invitation compared with those with private insurance (eTable 3 in the Supplement). Trial participation model results were similar.

When respondents were asked to imagine a need for clinical trial information, “health care providers” were most frequently reported as their first source of information (2297 [59%]), followed by the internet (644 [21%]) (Figure 1A). “Health care providers” were also reported most frequently as respondents most trusted information source (2597 [70%]), followed by health organizations or groups (456 [13%]) (Figure 1B). Among those invited to a clinical trial, primary information sources differed between those who did and did not participate. Respondents who reported participating in clinical trials, compared with those who did not report participation, less frequently reported physicians (94 of 199 [46%] vs 145 of 230 [57%]) and more frequently reported the internet as a first source of information (44 [28%] vs 34 [18%]). Among those invited to a clinical trial, trust in information sources did not differ substantially between those who did and did not participate.

When all respondents were asked to imagine being invited to a clinical trial for any health issue, the most influential factors in the decision to participate in a trial included “wanting to get better” (2294 [66%]), the standard of care not being covered by insurance (1448 [41%]), and the chance to try a new kind of care (1085 [30%] (Figure 2). Among those invited to a clinical trial, influential factors in the decision to participate differed between those who did and did not participate. Respondents who reported participating in trials, compared with those who did not report participation, were more motivated by helping other people (98 of 199 [45%] vs 70 of 230 [29%]) and getting paid (63 [39%] vs 49 [27%]).

Results of this nationally representative survey suggest that approximately 1 in 10 Americans report being invited to participate in a clinical trial, with a little less than half of those invited reporting participation. In multivariable models, respondents had higher odds of clinical trial invitation if they were non-Hispanic Black, had a college education, were single, were urban-dwelling, or had at least 1 medical condition. However, non-Hispanic Black respondents had lower odds of self-reported trial participation than non-Hispanic White respondents. These results reinforce the need for future research promoting effective clinical trial communication via health care professionals and the internet, as these were identified as important sources of clinical trial information. Our results also highlight a need for further research on ameliorative, financial, and altruistic factors as potential engagement strategies for trial participation.

Non-Hispanic Black respondents in our study had higher odds of reported invitation to clinical trials than non-Hispanic White respondents. These results contrast with prior literature proposing trial design- and researcher-level explanations for racial and ethnic disparities in clinical trial representation, such as restrictive eligibility criteria and researcher bias, that may hinder equitable invitation.^31-33 Our results may reflect researcher-level uptake of initiatives to increase the diversity of clinical trial participants. Since the 1993 National Institutes of Health (NIH) Revitalization Act³⁴ mandating inclusion of racial and ethnic minority groups in NIH-funded research, increased attention has been placed on greater clinical trial inclusivity.^35,36 More recently, the US Food and Drug Administration (FDA) issued recommendations for greater inclusion of certain segments of the population, including those identifying as Black, Indigenous, or another minority racial or ethnic group as well as analysis of study results by race and ethnicity.³⁷ Continued commitment to equity in trial invitation may help ensure equitable representation in clinical trials.

Despite greater odds of reported invitation to clinical trials in this study, non-Hispanic Black respondents had lower odds of reported participation in clinical trials than non-Hispanic White respondents. This contrasts with previous work showing Black and White individuals enroll in clinical trial at similar rates if invited.^30,31,38,39 Although our study controlled for comorbidity status, health care engagement, and financial burden, other items influencing barriers to clinical trial participation, such as ineligibility or lack of access, may result in lower participation rates. Furthermore, historical research mistreatment toward the Black community, current disparities in health care access and quality, and negative encounters with health care professionals may also affect trial participation.^40,41 Multilevel institutional efforts addressing eligibility, access, and systemic racism are needed to decrease racial and ethnic health inequities in clinical trial participation, such as diversifying the health care workforce, increasing community engagement, enhancing transparency of research practices, and training health care professionals in structural competency, defined as identifying and clinically addressing structural drivers of health inequity.^42-45 Building an environment of trust and facilitating effective communication between patients and the health care system could also be helpful. Participation in therapeutic cancer clinical trials nearly doubled among Black patients receiving trial education by nonclinical patient navigators compared with patients not receiving navigation services.⁴⁶ Other multilevel efforts, such as information dissemination via community health advisors,⁴⁷ warrant further research.

In our study, respondents residing in rural compared with urban areas had decreased odds of invitation to clinical trials. These results may stem from the historical lack of trial availability for individuals receiving medical care in rural settings. Recent efforts have focused on remediating this discrepancy. Current NIH-funded initiatives expand trial availability in community-based settings by dismantling patient-level barriers to trial participation related to residing far from care settings. The National Cancer Institute Community Oncology Research Program,^29,48 the National Institute on Drug Abuse Clinical Trials Network,⁴⁹ and the NIH Collaboratory Distributed Research Network⁵⁰ bring more research into community settings, offering increased opportunities for participation in clinical trials closer to their residence. To ensure equitable access to trials, increased awareness of trial availability is needed, especially among underserved and diverse patient populations.

Health care professionals may serve as the greatest opportunity to increase clinical trial enrollment, given that they were both the first and most trusted clinical trial information source reported by 59% and 70% of all respondents, respectively. This suggests patient-clinician communication is key to raising patient awareness of and participation in clinical trials. Clinicians must first be aware of actively recruiting trials as well as their eligibility criteria to communicate potential enrollment with their patients.^51-53 Furthermore, communicating complex trial protocols to patients with differing health literacy is challenging because of time-related constraints during clinical encounters.³³ The internet was reported as a trustworthy source of information by 28% of respondents, suggesting future research in web-based education and counseling to increase efficiency of in-clinic trial communications.⁵⁴ Culturally appropriate, plain-language tools could also facilitate patient-clinician communication, especially for patients without internet access or with low health literacy.^55,56 Historically underrepresented potential research participants found simple paper flashcards to be effective communication tools given that they quickly communicate research information.⁵⁷ Other modalities, like video, do not increase trial knowledge, but help facilitate extended discussions between patients and clinicians about clinical trial participation.^58,59

Financial incentives were commonly reported by study respondents as influential in their clinical trial participation decision, including having treatment payment covered and getting paid for trial participation. Although payment for treatment and routine care received while participating in trials may act as a facilitator to clinical trial participation, indirect costs related to trial participation, such as transportation to the clinic or time off work to participate, may act as barriers. Barriers related to indirect trial expenses may explain why respondents enrolled in Medicaid or dually enrolled were more often invited to trials but less often participated. Our results are consistent with previous studies regarding patient financial issues surrounding trial participation.^27,60,61 To potentially remove financial barriers to participation, recent recommendations from advocacy, research, and governmental organizations, including the FDA, have encouraged patient reimbursement for direct and indirect trial-related costs.^62-66 In December 2020, Congress passed the Clinical Treatment Act, which requires Medicaid to cover routine clinical care costs associated with clinical trial participation,⁶⁷ removing some financial barriers and furthering representation of disparate patient groups. However, quantification of and reimbursement for indirect trial costs in a trial environment remains unclear. Furthermore, trialists should ensure financial incentives are ethical and do not result in coercion or undue influence for study participation.⁶⁸ Efforts to reduce financial hardship for individuals willing to participate in clinical trials should be considered an ethical priority for trialists to increase equitable access for all.

The results from our study should be considered within the context of several limitations. The data for this study were subject to selection biases including mailed survey coverage error, nonresponse, and voluntary response potentially associated with likelihood to participate in a clinical trial. However, HINTS data weights were specifically calibrated to compensate for these biases.⁶⁹ Data were self-reported by HINTS respondents and therefore subject to recall bias. Although items were developed using 2 rounds of cognitive testing to maximize understandability, variation in respondent interpretation of these questions is possible. In particular, rates of invitation to clinical trials may be overreported in the current data, as respondents may have difficulty differentiating invitation to a clinical trial from prescreening or eligibility assessment for a trial, a more general discussion of a trial, or even invitation to other health research (eg, survey or qualitative studies). Future efforts to decrease measurement error are needed for the development of interventions targeting equitable clinical trial invitation and enrollment. The complexity surrounding the multilevel factors influencing clinical trial invitation and participation was not captured. Our results may be influenced by the low general knowledge of clinical trials in our sample, especially since this survey was fielded outside of a usual care setting. However, sensitivity analyses excluding respondents reporting no knowledge of clinical trials revealed similar results in our multivariable models. Respondents reported financial access to standard of care medications as a primary motivator to trial participation. This may be a uniquely US phenomenon given the high cost of health care, even for individuals with insurance. Because our survey was fielded during spring 2020, results may also be influenced by increased awareness of trials by the US public due to the COVID-19 pandemic. However, rates of invitation and participation were similar for surveys returned both before and after March 11, 2020, the date the pandemic was declared. Additionally, our results are not disease specific but rather provide a generalized snapshot of clinical trial engagement nationwide.

In this nationally representative sample, approximately 1 in 10 respondents reported being invited to participate in a clinical trial, and nearly half of those respondents reported participating in a trial after invitation. Self-reported clinical trial invitation was associated with respondent race and ethnicity, residence, education, marital status, and comorbidity status, while self-reported trial participation was associated with race and ethnicity alone. Respondents reported “health care providers” and the internet as important sources of clinical trial information, and health improvement, financial incentives, and altruism as key motivations for clinical trial participation. Strategies for increasing trial participation rates across diverse patient populations, such as increasing trial availability across diverse geographic care settings, facilitating patient-clinician trial information communication, and providing financial reimbursement for trial participation, warrant further research to ensure equitable translation of clinical benefits from research to practice.

Accepted for Publication: July 30, 2021.

Published: September 29, 2021. doi:10.1001/jamanetworkopen.2021.27792

Open Access: This is an open access article distributed under the terms of the CC-BY License. © 2021 Williams CP et al. JAMA Network Open.

Corresponding Author: Courtney P. Williams, DrPH, Division of Cancer Control and Population Sciences, National Cancer Institute, 9609 Medical Center Dr, Rockville, MD 20850 (courtney.williams@nih.gov).

Author Contributions: Dr Williams had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: All authors.

Acquisition, analysis, or interpretation of data: Williams, Senft Everson, Shelburne.

Drafting of the manuscript: Williams, Norton.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Williams, Senft Everson.

Administrative, technical, or material support: Shelburne.

Supervision: Senft Everson, Norton.

Conflict of Interest Disclosures: None reported.

Disclaimer: The opinions expressed by the authors are their own and this material should not be interpreted as representing the official viewpoint of the US Department of Health and Human Services, the National Institutes of Health, or the National Cancer Institute.

Additional Contributions: We would like to acknowledge Jennifer Alvidrez, PhD, David Chambers, DPhil, Andrea Denicoff, RN, MS, ANP, and Holly Massett, PhD, all of the National Institutes of Health, for their thoughtful input on earlier versions of the manuscript. No compensation was provided for their assistance.