Effects of socioeconomic status on enrollment in clinical trials for cancer: A systematic review

Abstract Background To achieve equitable access to cancer clinical trials (CCTs), patients must overcome structural, clinical, and attitudinal barriers to trial enrollment. The goal of this systematic review was to study the relationship between socioeconomic status (SES), assessed either by direct or proxy measures, and CCT enrollment. Methods The review team and medical librarian developed search strategies for each database to identify studies for this systematic review, which was conducted according to PRISMA guidelines. Inclusion criteria were as follows: studies published in relevant scientific journals between January 2000 and July 2022, primary sources, English literature, and studies conducted in the US. Sixteen studies fulfilled the inclusion criteria and were reviewed. The risk of bias assessment was conducted independently by two reviewers using the Newcastle Ottawa scale. Results The initial search yielded 4070 citations, and 16 studies were included in our review. Four of the studies included used patient reported annual income as a measure of SES, while the remaining 12 studies used patient zip code as a proxy measurement of SES. Consistent with our hypothesis, 13 studies showed a positive association between high SES (patient‐reported or proxy measurement) and CCT enrollment. Two studies showed a negative association, and one study showed no relationship. Conclusions The existing literature suggests that low SES is associated with lower participation in CCT. The small number of studies identified on this topic highlights the need for additional research on SES and other barriers to CCT participation.


| INTRODUCTION
Cancer clinical trial (CCT) enrollment at the initial course of treatment for oncology patients may increase survival rates, highlighting the importance of equitable access to CCTs. 1 However, current CCT enrollment favors young, healthy, white patients treated at academic centers. 1 Although CCTs present patients with access to novel treatments, fewer than 1 in 20 adult cancer patients enroll, raising concerns about generalizability of trial results. 2 Low accrual rates have a negative impact on CCTs because they can prolong duration, delay analysis, and lead to early closure of studies.
Barriers to CCT participation have been studied at length, yet the rate of trial participation as not changed substantially over time. 3Furthermore, disparities in CCT enrollment for racial and ethnic minorities have increased over time. 4To increase enrollment and achieve equitable access to CCTs, patients must overcome structural, clinical, and attitudinal barriers to trial enrollment. 2 A 2014 study found that the highest rated barriers to CCT participation included fear of side effects, insurance coverage concerns, fear of low efficacy, and fear of being randomized to placebo. 5 This underscores the importance of addressing costs and benefits in the decision-making process.
A patient's socioeconomic status (SES) reflects their relative position within a social hierarchy and their ability to consume resources.Studying the impact of SES on CCT enrollment is important because it can guide policy, advance treatment discovery, and increase survival.The goal of this systematic review is to study the relationship between SES and CCT enrollment.We hypothesized that low SES is a major barrier to participation in CCTs.

SEARCH
We followed the PRISMA 2020 checklist.Our population of interest in this systematic review was patients with cancer who were offered enrollment in a clinical trial (CT).We excluded secondary sources, non-English literature, and studies conducted outside of the US.This review was not registered.Since this review did not comprise human subjects research, we did not obtain IRB approval.
The review team and Clinical Informationist (JKN) developed search strategies for each database to identify studies for this systematic review.The search was restricted to studies published from 2000 to present and was conducted on July 27th, 2022.Of note, review articles were included in the search.Databases searched were PubMed, Embase, Cochrane, Web of Science, and Scopus.Limitations on the search were from January 1, 2000 to current.The search template (below) was translated into the individual databases with controlled vocabulary applied as applicable.
All article titles, abstracts, and full texts identified were reviewed by two reviewers through Emory University's subscription to Covidence, a screening and extraction tool for systematic reviews, to limit bias in inclusion.Disagreements were resolved by consensus.Since all identified articles were case-control or retrospective case studies, the risk of bias assessment was further evaluated independently by two reviewers (MWD and NCS) using the Newcastle Ottawa scale (Table S1).

| RESULTS
The initial search yielded 4070 citations, and 16 of these studies were included in our review (Figure S1).Characteristics of included studies are summarized in Table 1.
Historically, the study of SES and CCT enrollment has been hindered by a lack of data on income.While the measure of current income provides insight into a person's access to material goods and services, income is typically age dependent and does not include access to all assets such as wealth, insurance coverage, and disability benefits. 6In the absence of data on individual level income, area level income as a proxy measure of individual level SES is substituted. 7our of the studies included in this systematic review used patient reported annual income as a measure of SES, while the remaining 12 studies used patient zip codes to develop proxy measurements of SES.Of the 12 studies that used proxy measurements for SES, 8 used area level income, 3 used area level poverty, and 1 used an area level material deprivation score considering the percentage of persons living in poverty, households without a car, unemployed persons aged 16 and older, and owner-unoccupied housing.For studies that used income to measure SES, the cutoff for low SES ranged from a household income below $38,000-$50,000.For studies that used area poverty level as a proxy for SES, the cutoff for a high poverty level ranged from greater than 0.13%-4% of persons living in poverty.
Consistent with our hypothesis, 13 studies showed a positive association between high SES and CCT enrollment.Two studies showed a negative association, and one study showed no relationship.The results of these conflicting studies are further discussed below.
A 2019 study on CCT enrollment of children with acute myelogenous leukemia (AML) measured SES by calculating zip code-based poverty.This study found that patients with low zip code-based poverty rates (less than 4%) were less likely to enroll in AAML 1031 than patients who had higher zip-based poverty rates, differing from previous literature which has shown that pediatric patients from low-income areas have no differences in CCT enrollment. 8hile pediatric CCTs are inherently different from adult CCTs, this study suggests that perhaps efforts to increase equity in pediatric CCTs may have been successful.
A 2020 adult glioma study found that minority race and SES did not impact CCT enrollment, rather patient location was the most significant predictor of trial enrollment for patients with newly diagnosed gliomas. 9This was consistent with another study in this review, which concluded that reduced distance from the trial site was associated with increased odds of trial participation. 10he 2020 adult glioma study also found that additional socioeconomic factors known to play a role in CCT enrollment for non-glioma patient populations including employment status, insurance type, household income, and percent below poverty line were not associated with trial enrollment for glioma patients. 9This contrasts with a 2014 breast cancer study that found accrual is favorably associated with certain measures of socioeconomic disadvantage such as Medicaid eligibility and low educational attainment independent of income. 11he study by Fayanju and colleagues on CCT participation among breast cancer patients found that trial enrollment for the 12-year study period was associated with white race, higher levels of education, and lower levels of area-based income. 12Over the course of the study period, overall trial participation declined sharply, reflecting the trend to smaller, and fewer trials over time.This study is unique in demonstrating complex and varying interactions between income, race, and trial availability over 12 years.At the beginning of the study period, highincome Asian/Pacific Islander (API), Hispanic, and white patients had higher rates of participation than their lowincome counterparts.By the end of the study, low-income API and Hispanic patients were more likely to participate than their high-income counterparts.The likelihood of trial participation among all racial/ethnic groups has decreased over time, but gains among low-income API, Hispanic, and white patients appear to have occurred in proportion to their high-income counterparts, suggesting that efforts to diversify trial enrollment have been effective.

| DISCUSSION
Expanding CCT enrollment is advantageous for patients, researchers, and future generations, given it's crucial role in driving breakthroughs in cancer treatment discovery.Although CCTs provide patients with access to novel treatments, close monitoring, and increased survival, 1 enrollment among adults remains low.Generalizability is important to consider when applying trial results to real world patients.A lack of socioeconomic diversity in CTs limits generalizability because SES influences health both directly and indirectly through social, environmental, and medical factors such as chronic stressors, dietary options, and housing quality, and pollutants. 13o our knowledge, this is the first attempt at a systematic review to assess the relationship between SES and CCT enrollment.Consistent with our hypothesis, most included studies (13/16) concluded that lower SES is associated with decreased rates of CCT enrollment.While this systematic review does not explain why this trend exists, it suggests that low SES is associated with barriers that play a significant role in CCT enrollment.
The findings in our study were expected, but we were surprised to see that most studies used proxy measures of SES, most commonly zip code, rather than direct measures of SES.Although zip code may be an adequate proxy measure when individual or household SES data are not available, direct measures of SES are preferred. 14ultiple studies included in this systematic review found that lower rates of CCT enrollment were associated with increased distance from the clinical site, lower education levels, and minority race/ethnicity.Eskander et al. 15 concluded that geography remains a significant barrier to CCT access: Participants in that study who lived in the south needed to travel twice as far and had half the odds of enrolling in a CCT when compared to their northeastern counterparts.Baquet et al. 16 studied CCT recruitment and participation for adult Marylanders and found that respondents of rural Western Maryland and the rural Eastern Shore were significantly less likely to be recruited to CCT.These findings are consistent with data on growing health disparities that exist between rural and urban communities, with rural communities having lower life expectancy and higher mortality rates from most top causes of death. 17The effects of rural versus urban location on CCT enrollment are likely to be underestimated in the included studies, considering neighborhood level measures of SES inadequately account for individual SES in rural settings. 18Seventeen percent of the US population lives in rural or remote communities, but only 9% of physicians practice in rural areas. 19Additionally, rural communities have higher poverty rates and lower educational attainment when compared to urban communities. 19he role of race and ethnicity in CCT enrollment has been an area of frequent study, yet racial/ethnic minorities are still not adequately represented in CCT.Behrendt et al. 11 discovered that adjustment of SES helped reveal ancestry related disparities that would otherwise remain obscure, demonstrating that studying the intersection between race/ethnicity and SES in addition other vulnerabilities is key to understanding disparities and achieving equity in CCT enrollment.
When examining the relationship between SES and CCT enrollment, it is important to acknowledge the relationship between SES and race.Researchers have long assumed that racial differences in SES substantially contribute to racial disparities in health.However, complex patterns emerge when studying the relationship between race, SES, and various health outcomes.
In the US, racial groups capture differences in power, status, and resources because the structure of segregation and its consequences have remained relatively intact over time. 20Strikingly high levels of racial inequality in SES remain today, contrasting with the perception that racial inequalities have narrowed over time. 21Poverty varies greatly by race, with black and Hispanic levels of poverty being two to three times higher than that of whites. 21Even after controlling for SES, racial disparities in health often remain, with white males and females living 2-4 years longer than their black peers at every level of income. 21hile this paper does not aim to study the relationship between race, SES, and CCT enrollment, it is important to acknowledge the co-founding concerns given overlapping populations.
In a study by Unger and colleagues on patient income level and CCT participation, lower income predicted lower CCT participation in patients age ≥ 65, indicating an income disparity even in a population with universal access to Medicare. 22This finding suggests that while the cost of a CCT itself may not be a barrier to enrollment, other structural and attitudinal factors may play a significant role in the decision to enroll.Lower income patients are more likely than higher income patients to be concerned about how to pay for the nonmedical costs of CCT participation, which is consistent with the idea that low SES is a barrier to CCT enrollment. 23ancer is the most expensive chronic disease in the United States and the cost of care has increasingly shifted to patients in the form of premiums, copays, coinsurance, and deductibles leading to financial toxicity. 24In addition to direct costs, loss of wages also contributes to financial burden, with one study reporting that approximately one third of patient annual wages were lost within 1 year of an early stage cancer diagnosis. 25,26Financial toxicity is defined as the objective and subjective patient-level impact of the costs of cancer care and can afflict individuals-even those with health insurance-from any socioeconomic background. 24,274][35] The Affordable Care Act (ACA) requires coverage of routine care costs for patients enrolled in approved CTs, with exceptions for insurance programs predating the ACA; however, coverage of additional and indirect costs is not required.36,37 Indeed, patients enrolled in CTs spend over $600 per month out of pocket for nonmedical costs alone, with one out of five patients spending at least $1500 per month. 36These high out of pocket expenditures disproportionately effect non-white and Latinx/Hispanic patients 36 and may represent one factor contributing to the lack of diversity in CTs.Implementation of programs that address financial toxicity through a multipronged approach that includes financial reimbursement, increased accessibility, and patient advocacy may increase CT enrollment, particularly among patients with lower SES.
Although we found only one study in our review involving children, it suggests that SES may have different effects on CCT enrollment in children versus adults.Most children with cancer in the United States are enrolled in CCTs, in part because CCTs are considered the standard of care. 38The Children's Cancer Group (CCG), a cooperative CT group, was established in 1955 to address poor pediatric cancer outcomes. 39In 2000, four large cooperative CT groups merged to create the Children's Oncology Group (COG), which is estimated to treat 90% of children with cancer in the United States today. 39Since the implementation of cooperative CT groups, pediatric cancer survival has gone from 30% in the 1960s to over 80% today. 38he improvement in pediatric oncology survival parallels improvements in multidisciplinary care and research, mediated by cooperative CT groups. 39Additionally, free housing and financial assistance provided to pediatric oncology patients through programs such as such as Ronald McDonald House and St. Jude may have a unique influence on the pediatric CCT participation.
Our study has several limitations.Data on income is often not available in large data sets, and as a result, most studies included in this systematic review use proxy measures of SES based on zip code.Due to heterogeneity in study methodology, we were not able to conduct a metaanalysis.Despite these limitations, this systematic review highlights the importance of considering SES along with additional barriers to CCT enrollment.The surprisingly small number of studies also highlights the need for additional research on SES, other patient-level barriers, and strategies for equitable enrollment.

| CONCLUSIONS
SES is a major barrier to enrollment in CTs for cancer, at least in a subset of patients.Further research is needed on the role of SES in CT enrollment for different cancers, ideally with more direct (rather than surrogate) measures of SES.