Abstract
Background
Characterizing perceptions of clinical trials among the socioeconomically disadvantaged is necessary for understanding how social determinants of health such as socioeconomic disparities in education and income can affect people’s awareness of and exposure to clinical trials
Methods
A survey was distributed in spring 2023 among a survey taking sample stratified by demographic variables to reflect the U.S. population. The survey assessed the socioeconomic status of the respondent and related covariates, as well as outcome measures including interest in joining a clinical trial, concerns relating to participation, and whether the respondent had previously been asked to participate. Multiple and logistic regression were used to assess the relationship between predictor and outcome variables
Results
Here we show the results of outcome measures regressed on main predictors related to socioeconomic status and related demographic predictors. Education, employment status, insurance coverage, and English proficiency were significant predictors of interest in clinical trial participation. Education and the presence of a healthcare professional or former clinical trial participant in the respondent’s personal network were significant predictors of whether the respondent had previously been asked to participate in a clinical trial
Conclusions
The results of the analysis reveal how socioeconomically vulnerable groups, including those from low income and low education groups, are being excluded in clinical research. Analyses also uncovered the impact of clinical trial social influence—the presence of having a family or friend in one’s social network who participated in a clinical trial—on willingness to participate and exposure to clinical trials.
Subject terms: Medical research, Drug development
Plain language summary
Participation in clinical trials has remained largely inaccessible to historically underrepresented communities, which includes groups that are low income and low education. Here, we examine socioeconomic and demographic factors that can influence individuals’ willingness to participate in clinical trials and their experience being asked to participate in clinical trials. Using several types of analysis, we show that those who are low income and less educated are less willing to participate in clinical trials and are less likely to be asked to participate in clinical trials when compared to those with higher income and more education. This highlights the need for improved outreach among healthcare providers and clinical research staff to include these communities and provide individuals with the knowledge, awareness, and opportunity to participate in clinical trials.
Kim et. al explore the impact of socioeconomic vulnerability on clinical trial participation. Findings highlight barriers to trial entry including participant concerns and implications of exclusion of specific groups.
Introduction
Since 2020, the clinical research enterprise has intensified its commitment to diversity, equity, and inclusion, prompted largely by the Black Lives Matter movement and public outcry over disparities in racial and ethnic community participation in the development of and access to COVID-19 vaccinations1–4. The focus of this attention has encouraged the need to not only understand and address the underrepresentation of racial minority communities among clinical trial (CT) volunteers, but also examine the role that social determinants of health such as socioeconomic status (i.e., education, income) play in influencing perceptions about and the willingness to participate in CTs. The research thus far has pointed to large gaps in efforts to collect information related to socioeconomic status (SES). Bierer et al.5 concluded, following a review of the literature, that socioeconomic factors of CT participants are not uniformly collected or reported despite the growing body of research that links lower socioeconomic status to poor health outcomes.
A study conducted by the Tufts Center for the Study of Drug Development (Tufts CSDD) in 2020 found that among pivotal trials supporting drug and biological approvals by the Food and Drug Administration (FDA) between 2007 and 2017, less than 5% contained reference to study volunteer socioeconomic characteristics6. Relatedly, Alegria et al.7 found that a low and declining percentage (14.5% in 2015 vs. 12.3% in 2019) of major medical journal manuscripts even report the socioeconomic characteristics of patients enrolled in government and industry-funded CTs.
The failure to collect and report SES characteristics in CT participation has contributed to an incomplete and inaccurate understanding of who is (under)represented in CTs5,8. Despite growing calls among researchers urging for a deeper examination of how SES, and demographic factors such as employment status, insurance coverage, health literacy, household structure, and English proficiency can affect access to CT, only a few studies have empirically examined the influence of SES on clinical trial participation, with even fewer exploring the intersection of SES and other predictors, such as race and gender9,10. One study on oncology clinical trials showed that participants with lower annual income were significantly less likely to participate in CTs compared to those with higher annual income, confirming that clinical trials may be favoring those who are socioeconomically advantaged while excluding groups who would otherwise benefit from CT participation11. CTs provide an avenue for improving health outcomes for patients with diseases that do not respond to standard treatment12. If patients who are socioeconomically disadvantaged are not participating in CTs at similar rates as those who are socioeconomically advantaged, this trend may be a contributing factor to the health disparity between low- and high-SES groups.
Barriers to clinical trial participation can be structural, clinical, and attitudinal. This study focuses primarily on attitudinal barriers among different SES groups to shed light on the underlying thought process that can affect a person’s decision to participate in a CT. To do so, we chose to examine SES variables, namely education and income, and other factors that could be influenced by one’s SES, including the presence of potential influencers—a healthcare provider (HCP) or a former CT participant—in one’s social network and how this may influence attitudes toward CT. The research has suggested that people may be encouraged to participate in CT by hearing about the experience of an acquaintance who had a positive CT experience, highlighting the potentially important role that word-of-mouth referral through a former CT participant can have on one’s decision-making process13.
In addition to one’s willingness to participate in a CT, we examined participant’s experience with being asked to participate in a CT to determine if there is a gap in who is informed about CTs and who is not. Existing literature has shown that lower education and income levels, in particular, are associated with lower health literacy, which may deter referring clinicians from informing people in these groups about CTs due to biased assumptions, such as assuming that these groups are unable to afford certain procedures or are unable to comprehend basic instructions, and will thus not adhere to the strict study protocols14.
Gaining a clear understanding of reasons related to why individuals hesitate to participate in CTs, especially when segmented by SES, can better inform the design of recruitment strategies, providing insights into who to target and how to reach them. Previous research among the general population has shown that people are concerned about side effects and risks to one’s overall health, yet not much is known about how concerns vary by education or income level15. Therefore, the study also set out to measure general concerns that may contribute to the decision not to participate in a CT by education and income level.
In sum, the study examines the influence of sociodemographic factors on an individual’s interest in participating in a clinical trial, the likelihood of being asked to participate in a clinical trial, and reasons that hinder participation in clinical trials.
Methods
Study design
The study used a survey methodology and was administered by the Center for Information and Study on Clinical Research Participation (CISCRP), a non-profit organization specializing in clinical research education advocacy, among a panel of adult survey takers in the U.S. between March and April 2023. CISCRP used Alchemer, a survey provider that offers panel services to target specific groups in their database of 100 million respondents. CISCRP targeted respondents with no clinical trial experience in the U.S., stratified by age, gender, race, education, and income to better reflect the U.S. population. The survey was administered in English, Spanish, and Chinese. A priori power analysis using the following parameters: f2 = 0.02, , 1−β err prob = 0.95, number of predictors = 10 yielded a minimum sample size of 1229. To reduce Type I error and achieve as much statistical power as possible, we recruited well beyond the suggested 1229, which resulted in a final sample size of 400616. The study received Institutional Review Board (IRB) approval from WCG Clinical Services, and all respondents provided electronic consent to participate.
Participants
The inclusion criteria for participation were the following: (1) be at least 18 years old; (2) reside in the U.S.; and (3) not have previous experience participating in a clinical trial.
Measures
The survey contained multiple-choice questions that measured SES variables selected based on existing literature on SES factors affecting CT participation and retention1–3. SES variables included the primary predictors—education and income—and related covariates, such as work experience, employment status, insurance type, English proficiency, presence of an HCP, and former clinical research participant in one’s network, in addition to race and gender. English proficiency was measured using a 4-point Likert scale (1-Not at all; 4-Very well). See Appendix A for the survey questionnaire.
Outcome measures
The outcome variable—interest in joining a CT—was measured on a 5-point Likert scale (i.e., 1-Not at all willing; 5-Very willing). Whether they had been asked to participate in a CT in the past was measured and treated as a binary variable (Yes/No). Concerns relating to participation were measured by asking participants to choose from a list containing reasons for not participating in a CT. The list was based on previous research15.
Statistical analyses
To determine the influence of SES and other social factors on an individual’s interest in participating in CT, multiple regression was performed where interest in CT participation—the outcome variable—was regressed on the main predictors—education and income—and related covariates—gender, race, employment status, insurance coverage, English proficiency, presence of a healthcare provider, and presence of a former CT participant in one’s network. To measure the influence of SES and other social factors on the outcome variable—an individual’s experience with being asked to participate in a CT (Yes/No) —logistic regression, which measures the log odds ratio as a linear combination of the predictors, was used17. The outcome variable was regressed on the same set of predictors and covariates. Finally, to understand the top reasons for not wanting to participate in a CT by income and education, chi-square tests of independence were performed. All analyses were performed in R18.
Reporting summary
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.
Results
Description of survey respondents
A total of 4006 participants completed the survey. Fifty percent were female. Seventy-four percent were White, followed by 11.3% Black, 3.4% Asian, and 6.6% Latino/a. The average age was 50. Most were employed full-time (40.1%). Most respondents had a high school diploma or higher. Eighty-nine percent had insurance coverage. Thirty-seven percent made $40,000 or less in annual income. The majority (93.5%) of respondents reported speaking English “Very well.” Less than ten percent (9.7%) had previously been asked to participate in a CT. See Table 1 for full descriptive statistics. Because SES variables underpinned the study’s underlying question, which was to understand attitudes and perceptions among research-naive but research-eligible participants who could participate in clinical research, we examined the distribution of the most common medical conditions—cardiovascular disease, oncology, and respiratory—by education and income. For example, 44.6% of those with a HS diploma or less and 45.9% of those with an undergraduate degree reported that they had a cardiovascular condition, while 41.4% of those with an annual income between $10,000 and $40,000 reported having a respiratory illness, and 43.9% of those with an annual income above $80,000 reported having cancer, giving assurance that our sample was appropriate.
Table 1.
Respondent demographics
| Demographics | Percentage (N) |
|---|---|
| Race | |
| Whited | 73.9% (2959) |
| Black | 11.3% (452) |
| Latino/a | 6.6% (263) |
| Asian American/Pacific Islander | 3.4% (135) |
| American Indian/Alaska Native | 0.6% (27) |
| Multi-racial or othera | 4.2% (170) |
| Gender | |
| Femaled | 49.8% (1990) |
| Male | 49.7% (1995) |
| Otherb | 0.5% (21) |
| Age | |
| 18–34 | 23.3% (932) |
| 35–44 | 18.1% (725) |
| 45–54 | 16.2% (650) |
| 55–64 | 17.2% (691) |
| 65+ | 25.2% (1008) |
| Work Experience | |
| 0–5 | 14.3% (573) |
| 5–10 | 12.0% (479) |
| 10+ | 73.7% (2954) |
| Education | |
| High school or less | 48.4% (1938) |
| Undergraduate | 41.6% (1668) |
| Advancedd | 10.0% (400) |
| Insurance Coverage | |
| Insured | 8.6% (346) |
| Uninsured | 88.9% (3561) |
| Don’t knowd | 2.5% (99) |
| Employment Status | |
| Full-timed | 40.1% (1607) |
| Part-time | 12.5% (499) |
| Retired | 27.9% (1116) |
| Unemployed | 13.8% (552) |
| Otherc | 5.8% (232) |
| Annual Income | |
| $10–40 Kd | 37.8% (1516) |
| $40–80 K | 31.4% (1257) |
| $80 K+ | 30.8% (1233) |
| English Proficiency Level | |
| Not at all/Not welld | 1.0% (41) |
| Well | 5.4% (219) |
| Very well | 93.5% (3746) |
| Do you have any family/friends who are healthcare professionals | |
| Yes | 27.9% (1118) |
| Nod | 72.1% (2888) |
| Has anyone in your family/friends/community ever joined a CT? | |
| Yes | 8.9% (356) |
| Nod | 91.1% (3650) |
| Interest in joining a clinical research study | |
| Very Interested | 21.7% (871) |
| Somewhat Interested | 32.0% (1280) |
| Interested | 18.2% (728) |
| Not Very Interested | 19.9% (798) |
| Not at all interested | 8.2% (329) |
| Asked to participate in a clinical research study | |
| Yes | 9.7% (388) |
| No | 90.3% (3618) |
| Reported concerns related to participation in a clinical research study among those uninterested in participation by annual income | |
| $10–40 K | 43.5% (490) |
| $40–80 K | 29.9% (337) |
| $80 K+ | 26.6% (300) |
| Reported concerns related to participation in a clinical research study among those uninterested in participation by education | |
| High school or less | 55.4% (624) |
| Undergraduate | 37.5% (423) |
| Advanced | 7.1% (80) |
aOther racial identity includes respondents who selected more than one race and respondents who selected “other”.
bOther Gender Identity includes transgender male/transgender man; transgender female/transgender woman; non-binary/gender non-confirming; gender queer/gender fluid; another identity.
cOther included people who were not employed (i.e., students, stay-at-home parent, unable to work due to disability).
dReference variable in regression analysis.
Interest in participating in clinical trials
Results revealed that among the primary variables of interest, education was a significant predictor, as were employment status, insurance coverage, and English proficiency. The presence of a friend/family who participated in a CT and presence of a healthcare provider in one’s network were also significant predictors. Specifically, those who had either a HS diploma or less, , or an undergraduate degree, , expressed less interest than those with an advanced degree. Individuals who were employed part-time, , not employed, , unemployed, , or retired, were less interested in participating in CTs than those employed full-time. English proficiency was positively associated with interest level, . Having a friend or family member who participated in a CT increased one’s interest level compared to those who did not, ; however, the presence of a healthcare provider in one’s network decreased one’s interest, . See Table 2 for full regression results.
Table 2.
Regression results showing interest in joining a clinical trial regressed on variables of interest
| Predictor | b | p | b 95% CI [LL, UL] |
|---|---|---|---|
| (Intercept) | 2.60*** | <0.001 | [2.13, 3.07] |
| Race—Black (Ref = White) | 0.03 | 0.6 | [−0.09, 0.16] |
| Race—Asian American/Pacific Islander | −0.03 | 0.8 | [−0.24, 0.18] |
| Race—Latino/a | 0.01 | >0.9 | [−0.15, 0.16] |
| Race—American Indian/Alaska Native | 0.13 | 0.6 | [−0.32, 0.59] |
| Race—Multi-racial or other | 0.05 | 0.6 | [−0.14, 0.24] |
| Gender—Male (Ref = Female) | 0.15*** | <0.001 | [0.08, 0.23] |
| Gender—Other | 0.12 | 0.6 | [−0.40, 0.64] |
| Work experience | −0.00 | 0.7 | [−0.00, 0.00] |
| Annual Income—40-80 K (Ref = 10–40 K) | −0.00 | >0.9 | [−0.09, 0.09] |
| Annual Income—80 K+ | 0.03 | 0.6 | [−0.07, 0.12] |
| Employment—Part-time (Ref = Full-time) | −0.12* | 0.04 | [−0.25, −0.00] |
| Employment—Not employed | −0.35*** | <0.001 | [−0.52, −0.18] |
| Employment—Unemployed | −0.23*** | <0.001 | [−0.36, −0.11] |
| Employment—Retired | −0.54*** | <0.001 | [−0.64, −0.43] |
| Education—College (Ref = Advanced) | −0.18** | 0.007 | [−0.32, −0.05] |
| Education—High school or less | −0.24*** | <0.001 | [−0.38, −0.10] |
| Insurance—Uninsured (Ref = Don’t know) | 0.18 | 0.2 | [−0.09, 0.45] |
| Insurance—Insured | 0.42*** | <0.001 | [0.17, 0.66] |
| Clinical research participant in my network—Yes (Ref = No) | 0.28*** | <0.001 | [0.14, 0.41] |
| HCP in my network—Yes (Ref = No) | −0.10* | 0.02 | [−0.18, −0.01] |
| English Proficiency—Very well (Ref = Not well/not at all) | 0.62** | 0.001 | [0.25, 1.00] |
| English Proficiency—Well | 0.39 | 0.05 | [−0.01, 0.79] |
*p < 0.05, R2 = .062**, 95% CI[0.04, 0.07], ***p < 0.001.
Asked to participate in a clinical trial
Of the SES variables, education was the only significant predictor. The presence of an HCP or a former clinical research participant were also significant predictors. For ease of interpretation, we took the exponentiation of the log odds coefficients to provide the odds ratio of the predictors and covariates. Results indicated that the odds of being asked to participate in a CT decreased by a factor of 0.60 [95% CI, −0.87, −0.13], p = 0.007, for the HS or less group, and by 0.71 [95% CI, −0.68, 0.00], p = 0.04, for undergraduates compared to people with an advanced degree. Among those who had a family/friend who participated in a CT, the odds of being asked to participate in a CT increased by a factor of 5.71, [95% CI, 1.47, 2.01], p < 0.001. Further, having a healthcare provider in one’s network increased the odds of being asked by a factor of 1.78, [95% CI, 0.35, 0.81], p < 0.001. See Table 3 for full results.
Table 3.
Logistic regression results where experience with being asked to participate in a CT is regressed on variables of interest
| Term | estimate | SE | p | Odds Ratio | 95% CI |
|---|---|---|---|---|---|
| (Intercept) | −3.19 | 0.79 | <0.001*** | 0.04 | [−4.91, −1.77] |
| Race—Black (Ref = White) | 0.53 | 0.17 | 0.002** | 1.70 | [0.19, 0.86] |
| Race—Asian American/Pacific Islander | −0.38 | 0.37 | 0.31 | 0.69 | [−1.17, 0.30] |
| Race—Latino/a | 0.36 | 0.23 | 0.10 | 1.44 | [−0.09, 0.79] |
| Race—American Indian/Alaska Native | −0.09 | 0.77 | 0.90 | 0.91 | [−1.97, 1.19] |
| Race—Multi-racial or other | 0.18 | 0.29 | 0.53 | 1.20 | [−0.42, 0.72] |
| Gender—Male (Ref = Female) | 0.47 | 0.12 | <0.001*** | 1.60 | [0.24, 0.70] |
| Gender—Other | −12.82 | 304.98 | 0.96 | 0.00 | [NA, 4.53] |
| Work experience | 0.00 | 0.00 | 0.80 | 1.00 | [−0.01, 0.01] |
| Annual Income—40–80 K (Ref = 10–40 K) | −0.14 | 0.15 | 0.32 | 0.87 | [−0.43, 0.14] |
| Annual Income—80 K+ | −0.10 | 0.15 | 0.53 | 0.91 | [−0.40, 0.20] |
| Employment—Part-time (Ref = full-time) | −0.13 | 0.19 | 0.49 | 0.88 | [−0.52, 0.24] |
| Employment—Other | 0.16 | 0.26 | 0.55 | 1.17 | [−0.38, 0.65] |
| Employment—Unemployed | −0.25 | 0.21 | 0.24 | 0.78 | [−0.68, 0.16] |
| Employment—Retired | 0.11 | 0.16 | 0.51 | 1.11 | [−0.22, 0.43] |
| Education—College (Ref = advanced) | −0.34 | 0.17 | 0.04* | 0.71 | [−0.68, 0.00] |
| Education—High school or less | −0.50 | 0.19 | 0.007** | 0.60 | [−0.87, −0.13] |
| Insurance—Uninsured (Ref = don’t know) | −0.38 | 0.44 | 0.38 | 0.68 | [−1.22, 0.52] |
| Insurance—Insured | 0.07 | 0.39 | 0.85 | 1.07 | [−0.64, 0.89] |
| Clinical research participant in my network –yes (Ref = no) | 1.74 | 0.14 | <0.001*** | 5.71 | [1.47, 2.01] |
| Health care provider in my network –yes | 0.58 | 0.12 | <0.001*** | 1.78 | [0.35, 0.81] |
| English proficiency—very well (Ref = not very/not at all) | 0.51 | 0.64 | 0.42 | 1.66 | [−0.61, 1.99] |
| English proficiency—well | 0.77 | 0.67 | 0.25 | 2.15 | [−0.43, 2.29] |
*p < 0.05, **p < 0.01, ***p < 0.001.
Reasons for not participating in a clinical trial
Next, we report the top reasons why respondents chose not to participate in a CT by income and education. Across all income and education levels, the top concern was about feeling like one was in a medical experiment. This was followed by concerns about medical procedures, cost, worries about one's health, and lack of time. Not wanting to be in a medical experiment was strongly associated with annual household income (χ2 = 216.03, df = 2, p < 0.001). Those earning $40,000 or less, and those in the $40,000–$80,000 income group were more likely to view CTs as a medical experiment (58.4% and 59.6% respectively) compared to those in the $80,000+ income group (54.7%). With regards to secondary concerns, those in the highest income group were more likely to cite not having time (24.0%) compared to those earning less than $40,000 or $40,000–$80,000; this result was significant (χ2 = 81.766, df = 2, p < 0.001). Those earning less than $40,000 or $40,000–$80,000 were more likely to report concerns about medical procedures (35.3% and 35.0% respectively) compared to those in the $80,000+ group (32.0%); this difference was significant (χ2 = 130.25, df = 2, p < 0.001). See Table 4 for concerns by income.
Table 4.
Top concerns among those who were not interested in clinical trial participation by annual incomea
| Annual household income | |||||
|---|---|---|---|---|---|
| Concern | Overall | $10–40 K | $40–80 K | $80 K+ | p valueb |
| N = 1127 | N = 490 | N = 337 | N = 300 | ||
| I do not want to be in a medical experiment | 57.7% | 58.4% (286) | 59.6% (201) | 54.7% (164) | <0.001 |
| I am concerned about the medical procedures | 34.3% | 35.3% (173) | 35.0% (118) | 32.0% (96) | <0.001 |
| I am concerned about the cost (e.g., out-of-pocket costs, lack of insurance coverage) | 27.6% | 25.9% (127) | 30.5% (103) | 27.0% (81) | <0.001 |
| I am worried about my health | 23.1% | 23.7% (116) | 24.3% (82) | 20.7% (62) | <0.001 |
| I don’t have time | 20.1% | 18.4% (90) | 20.5% (69) | 24.0% (72) | <0.001 |
aThis included people who chose “not very/not at all interested in participating in a CT”.
bBased on a Pearson Chi-Square test, the strength of the relationship between annual household income and the concerns about participating in clinical research included in Table 4 was significant among those who are not interested in clinical trial participation.
The relationship between educational attainment and concerns about participating in CTs was also significant. Top concerns included not wanting to be in a medical experiment (χ2 = 216.03, df = 2, p < 0.001), concerns about medical procedures (χ2 = 130.25, df = 2, p < 0.001), concerns about cost (χ2 = 70.841, df = 2, p < 0.001), worries about their health (χ2 = 79.485, df = 2, p < 0.001), and lack of time (χ2 = 81.766, df = 2, p < 0.0001). A higher percentage of undergraduates who reported not being interested in participating in studies viewed CTs as medical experiments (61.5%), compared to those with advanced degrees (58.8%) and those in the HS or less group (55.1%). With regards to secondary concerns, those with an advanced degree cited worries about their health more often (26.2%) compared to those in the undergraduate and HS or less cohorts (21.5% and 24.8% respectively). Undergraduates and the HS or less group were more likely to cite concerns about cost (30.3% and 25.1%, respectively) compared to those with an advanced degree (20.0%). Those with an advanced degree were more likely to report not having time (23.8%) than undergraduates (19.1%) and the HS or less group (21.0%). See Table 5 for concerns by education.
Table 5.
Top concerns among those who were not interested in clinical trial participation by educationa
| Educational attainment | |||||
|---|---|---|---|---|---|
| Concern | Overall | High school or less | Undergraduate | Advanced | p valueb |
| N = 1127 | N = 624 | N = 423 | N = 80 | ||
| I do not want to be in a medical experiment | 57.7% | 55.1% (344) | 61.5% (260) | 58.8% (47) | <0.001 |
| I am concerned about the medical procedures | 34.3% | 30.9% (193) | 40.2% (170) | 30.0% (24) | <0.001 |
| I am concerned about the cost (e.g., out-of-pocket costs, lack of insurance coverage) | 27.6% | 30.3% (189) | 25.1% (106) | 20.0% (16) | <0.001 |
| I am worried about my health | 23.1% | 21.5% (134) | 24.8% (105) | 26.2% (21) | <0.001 |
| I don’t have time | 20.1% | 21.0% (131) | 19.1% (81) | 23.8% (19) | <0.001 |
aThis included people who chose “not very/not at all interested in participating in a CT”.
bBased on a Pearson Chi-Square test, the strength of the relationship between educational attainment and the concerns about participating in clinical research included in Table 5 was significant among those who are not interested in clinical trial participation.
Discussion
This study is one of the first to systematically quantify how perceptions and exposure to CTs differ by SES groups, focusing on socioeconomically disadvantaged groups that are lower income and less educated. We build on previous research that studied the influence of SES and its impact on clinical trial perceptions19,20. We did so by surveying a large, stratified sample of the U.S. population, addressing concerns about lack of external validity, and expanding the generalizability of the study results. Lack of education consistently emerged as a potential barrier to participation. Though a large percentage of people with a HS education or less reported having at least one type of chronic medical condition, such as cardiovascular disease, cancer, and respiratory illness, they were the least likely group to express interest in participating in a CT and were the least likely group to have been asked to participate in a CT. Participating in CT can be positively linked to health outcomes, improving all-cause mortality as well as indication-specific mortality, especially for patients who do not respond well to conventional treatment12. Our findings suggest that less educated people are less likely to be asked to participate in CT due partly to limited access to health services/treatment and bias among referring clinicians who may not think to mention clinical studies to people from lower income levels. Though studies exploring referral bias among referring clinicians are sparse, one qualitative study conducted among key stakeholders across several oncology clinics in the U.S. found that referring clinicians were less likely to refer people whom they viewed as non-adherent to the study protocol21. Given that education is correlated with health literacy and study adherence, providers may prefer to spend their time recruiting college-educated patients assuming that they would be more likely to adhere to the study protocol than less educated, hence less health literate individuals, excluding this group from CT recruitment.
Another barrier that the less educated group faces is the financial burden associated with participating in CTs. This group was more likely than others to list distance to the clinic, out-of-pocket expenses, and need for approval from doctor and family as reasons for not participating in a CT, highlighting that the levers to reduce their hesitancy may be a combination of financial and logistic support. This finding also suggests that this group may be more easily influenced by their physician’s opinion than others, underscoring the critical role that physicians can play to help break down these barriers to participation. Our results spotlight opportunities for healthcare providers, sponsors, and research sites to actively include less educated individuals to increase their awareness of the benefits of CT participation.
Another noteworthy finding was that the presence of a former research participant in one’s network, what we refer to as clinical trial social influence (CTSI), was significantly associated with an increase in respondents’ level of interest in clinical trial participation. People who were socially connected to a former clinical trial participant were also significantly more likely to have been asked to participate in a CT. Previous research has highlighted the influence of CTSI, albeit among a relatively small sample or based on anecdotal evidence22,23. Our study is the first to quantify the effect of CTSI on CT perceptions. The presence of a CT participant in one’s network can serve as a strong incentive for one to consider joining a study and underscores just how influential word-of-mouth referral can be. This influence may be especially important to consider among underserved communities who may not otherwise have opportunities to learn about CT.
Employment status also emerged as a potential barrier. Those who were not working due to various reasons (i.e., retired, disabled, unable to find a job) were less interested in joining a CT compared to those who were fully employed. Existing medical research has pointed out that non-employed individuals were more likely to experience adverse health outcomes and have poor access to healthcare24. Our findings enhance the existing literature by demonstrating how non-employment can also affect CT participation. People who are not employed may have less financial resources, less health coverage and access, and less opportunities to learn about CTs through socialization with one’s colleagues.
Though race was not a focus in this study, results showed that Black participants were more likely than White participants to share that they had been asked to participate in a CT. This finding is puzzling in that it contradicts previous evidence that shows the opposite trend25. However, this finding should be interpreted with caution as it may reflect the mass efforts to recruit hard-hit minorities into COVID-19 vaccine trials to study a CT population that was representative of the U.S. population. Therefore, this result may not be indicative of actual efforts to recruit African American participants into CTs in other therapy areas26,27. A deeper look into how they were asked sheds some additional light. Email was the great unifier in that all racial groups shared that the most common method they were asked to participate in a CT was through email. However, there was a notable difference in the second most common way people were asked. White participants shared that after email (36.3%), they were most likely to be informed about CTs by their HCP (16.1%). This was not the case for racial minority groups. For instance, Black participants shared that they were asked by email (34.4%), followed by social media (20.3%). We found a similar trend among Asian and Latino/a participants who reported email, social media, and telephone as the most common ways they were asked. This trend suggests that HCPs are more likely to inform White patients about CTs than they are their racial minority patients, highlighting a potential bias among HCPs, a tendency that has been documented in healthcare research21,25.
Our study is not without limitations. First, our survey was conducted among a sample of research-naive participants who did not have any CT experience. Though the study provides insights into how SES can influence potential candidates of CTs, it does not reveal much about the SES of actual CT participants. Future research may wish to conduct a similar study among CT participants to create a sociodemographic profile of the typical research participant, which can better inform the barriers and enablers they experience. Another limitation was that the study sampled a broad range of people with different medical conditions instead of focusing on specific conditions, potentially obscuring disparities that exist among specific conditions. While this is a limitation, our broad sampling of different medical conditions helps establish key insights about perceptions and attitudes toward CTs among the general U.S. population that can serve as a benchmark that future studies focusing on specific indications can build upon. We note that a limitation of our study was the under-sampling of Latino/a and Asian American participants in our sample, who represented 6.6% and 3.4% of the participants, well below the 18% and 6.7% they represent in the U.S. even though the survey was administered in Spanish and Mandarin. This suggests that the barriers go beyond language, reflecting instead other barriers that are hindering the participation of Latino/a and Asian American participants and something that can be explored in future studies that examine, for example, gatekeeper bias. Lastly, future studies can also capture how the intersection of SES and other important predictors such as race can influence participants' experience with and attitudes toward clinical trials, helping to explicate how.
Conclusion
Our study is one of the first to systematically examine the role that socioeconomic factors and CTSI can play in influencing people’s perceptions and attitudes toward CTs. We highlight underserved populations using variables such as education and income to spotlight groups that could benefit from bolstered efforts to educate and include them, while also noting the important role that former CT participants can have within their social network to facilitate favorable attitudes and willingness for others to join CT studies.
Supplementary information
Acknowledgements
We thank the CISCRP for their support in survey distribution. This work was funded by Parexel. The funders had a role in reviewing the manuscript but did not have a role in the study design, data collection and analysis, writing of the paper or decision to publish.
Author contributions
Conceptualization: J.Y.K., M.F., E.B., K.G. Methodology, investigation, and writing–original draft: J.Y.K., M.F., E.B. Writing–review and editing: all authors.
Peer review
Peer review information
Communications Medicine thanks Jason Cummings, and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Peer reviewer reports are available.
Data availability
The dataset generated by the survey research during and/or analyzed during the current study are available in the Dataverse repository: 10.7910/DVN/FOP4IN. No restrictions exist on data availability28.
Competing interests
J.Y.K., M.F., E.B., and K.G.: the study was sponsored by Parexel. X.B. and C.G.: sponsored the project.
Footnotes
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary information
The online version contains supplementary material available at 10.1038/s43856-024-00586-9.
References
- 1.Chastain, D. B. et al. Racial disproportionality in COVID clinical trials. N. Engl. J. Med.383, e59 (2020). 10.1056/NEJMp2021971 [DOI] [PubMed] [Google Scholar]
- 2.Webb Hooper, M., Nápoles, A. M. & Pérez-Stable, E. J. COVID-19 and racial/ethnic disparities. JAMA323, 2466 (2020). 10.1001/jama.2020.8598 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Borno, H. T., Zhang, S. & Gomez, S. COVID-19 disparities: an urgent call for race reporting and representation in clinical research. Contemp. Clin. Trials Commun.19, 100630 (2020). 10.1016/j.conctc.2020.100630 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Isaacs, D., Tarnow‐Mordi, W. O. & Sherwood, J. Black lives matter movement: the time for nice words and good intentions is over. J. Paediatr. Child Health56, 1327–1329 (2020). 10.1111/jpc.15100 [DOI] [PubMed] [Google Scholar]
- 5.Bierer B. E. et al. Achieving diversity, inclusion and equity in clinical research. The MRCT Center of Brigham and Women’s Hospital and Harvard; https://mrctcenter.org/diversity-in-clinical-research/guidance/guidance-document/ (2021).
- 6.Getz, K. A., Smith, Z. P. & Peña, Y. Quantifying patient subpopulation disparities in new drugs and biologics approved between 2007 and 2017. Ther.Innov. Regul. Sci.54, 1541–1550 (2020). 10.1007/s43441-020-00181-9 [DOI] [PubMed] [Google Scholar]
- 7.Alegria, M., Sud, S., Steinberg, B. E., Gai, N. & Siddiqui, A. Reporting of participant race, sex, and socioeconomic status in randomized clinical trials in General Medical Journals, 2015 vs 2019. JAMA Netw. Open4, e2111516 (2021). 10.1001/jamanetworkopen.2021.11516 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Yadlapalli, J. S. K. B. & Martin, I. G. Seeking predictable subject characteristics that influence clinical trial discontinuation. Drug Inf. J.46, 313–319 (2012). 10.1177/0092861512440850 [DOI] [Google Scholar]
- 9.Wilkins, C. H., Schindler, S. E. & Morris, J. C. Addressing health disparities among minority populations: why clinical trial recruitment is not enough. JAMA Neurol.77, 1063 (2020). 10.1001/jamaneurol.2020.1614 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Price, K. N., Lyons, A. B., Hamzavi, I. H., Hsiao, J. L. & Shi, V. Y. Facilitating clinical trials participation of low socioeconomic status patients. Dermatology237, 843–846 (2021). 10.1159/000511889 [DOI] [PubMed] [Google Scholar]
- 11.Unger, J. M., Gralow, J. R., Albain, K. S., Ramsey, S. D. & Hershman, D. L. Patient income level and cancer clinical trial participation. JAMA Oncol.2, 137–139 (2016). 10.1001/jamaoncol.2015.3924 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Baker, J. R. et al. Clinical trial participation improves outcome: a matched historical cohort study. Clin. Trials10, 735–743 (2013). 10.1177/1740774513496915 [DOI] [PubMed] [Google Scholar]
- 13.Owens, O. L., Jackson, D. D., Thomas, T. L., Friedman, D. B. & Hébert, J. R. African American men’s and women’s perceptions of clinical trials research: focusing on prostate cancer among a high-risk population in the South. J. Health Care Poor Underserved24, 1784–1800 (2013). 10.1353/hpu.2013.0187 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Sudhakar, S. et al. Health literacy and education level correlates of participation and outcome in a remotely delivered epilepsy self-management program. Epilepsy Behav.107, 107026 (2020). 10.1016/j.yebeh.2020.107026 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Center for Information & Study on Clinical Research Participation. Perceptions and Insights Study 2021: General Perceptions. Center for Information & Study on Clinical Research Participation; 2021. https://www.ciscrp.org/jet-popup/pi-report-participation-download/ (2023).
- 16.Lakens, D. Sample size justification. Collabra: Psychol.8, 33267 (2022). 10.1525/collabra.33267 [DOI] [Google Scholar]
- 17.Lever, J., Krzywinski, M. & Altman, N. Logistic regression. Nat. Methods13, 541–542 (2016). 10.1038/nmeth.3904 [DOI] [Google Scholar]
- 18.R Core Team. R: a language and environment for statistical computing. R foundation for statistical computing. https://www.r-project.org/ (2017).
- 19.Baquet, C. R., Commiskey, P., Daniel Mullins, C. & Mishra, S. I. Recruitment and participation in clinical trials: socio-demographic, rural/urban, and health care access predictors. Cancer Detect. Prev.30, 24–33 (2006). 10.1016/j.cdp.2005.12.001 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Brewer, L. C. et al. African American women’s perceptions and attitudes regarding participation in medical research: the Mayo Clinic/the links, incorporated partnership. J. Women’s. Health23, 681–687 (2014). 10.1089/jwh.2014.4751 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Niranjan, S. J. et al. Bias and stereotyping among research and clinical professionals: perspectives on minority recruitment for oncology clinical trials. Cancer126, 1958–1968 (2020). 10.1002/cncr.32755 [DOI] [PubMed] [Google Scholar]
- 22.Rollins, L. et al. Learning and action in community health: using the health belief model to assess and educate african american community residents about participation in clinical research. IJERPH15, 1862 (2018). 10.3390/ijerph15091862 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Advani, A. S. et al. Barriers to the participation of African-American patients with cancer in clinical trials: a pilot study. Cancer97, 1499–1506 (2003). 10.1002/cncr.11213 [DOI] [PubMed] [Google Scholar]
- 24.Silver, S. R., Li, J. & Quay, B. Employment status, unemployment duration, and health‐related metrics among US adults of prime working age: behavioral risk factor surveillance system, 2018–2019. Am. J. Ind. Med65, 59–71 (2022). 10.1002/ajim.23308 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Ahaghotu, C., Tyler, R. & Sartor, O. African american participation in oncology clinical trials—focus on prostate cancer: implications, barriers, and potential solutions. Clin. Genitourin. Cancer14, 105–116 (2016). 10.1016/j.clgc.2015.12.003 [DOI] [PubMed] [Google Scholar]
- 26.Jaklevic, M. C. Researchers strive to recruit hard-hit minorities into COVID-19 vaccine trials. JAMA324, 826 (2020). 10.1001/jama.2020.11244 [DOI] [PubMed] [Google Scholar]
- 27.Artiga S., Kates J., Michaud J., Hill L. Racial diversity within COVID-19 vaccine clinical trials: key questions and answers. KFF. https://www.kff.org/racial-equity-and-health-policy/issue-brief/racial-diversity-within-covid-19-vaccine-clinical-trials-key-questions-and-answers/ (2023).
- 28.Kim J., et al The impact of socioeconomic status on individual attitudes and experience with clinical trials: How socioeconomically disadvantaged individuals are being left behind. 10.7910/DVN/FOP4IN (2023).
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The dataset generated by the survey research during and/or analyzed during the current study are available in the Dataverse repository: 10.7910/DVN/FOP4IN. No restrictions exist on data availability28.