Clinical Trial Perceptions Among Patients Enrolled in a Precision Oncology Protocol at an Appalachian‐Serving Cancer Center

ABSTRACT

Objective

Low rates of clinical trial participation, with notable disparities among rural populations, slow cancer treatment advances. Precision oncology protocols are designed to alleviate trial access barriers by consenting patients to sequencing with results matched to national databases of available trials. Precision oncology protocols represent a potential avenue for supporting patient informed trial decision‐making before they are offered a specific trial but little is known about clinical trial perceptions and intentions of patients who join such protocols.

Methods

We administered surveys to 116 patients with a cancer diagnosis enrolled in a precision oncology protocol at an Appalachia‐serving, academic cancer center. Analyses tested associations between clinical trial perceptions (e.g., beliefs, attitudes, knowledge), sociodemographic covariates, and trial intentions.

Results

Participants (mean age = 64.5; median education = high school; 97.4% white; 51.7% female; 60% Appalachian residence) held overall favorable clinical trial perceptions and overall positive intentions. Trial knowledge was significantly lower among participants with lower education, lower income, and those residing in Appalachian counties. Intentions were significantly associated with attitudes, beliefs around trying new treatments, and perceived family support for trial participation in both bivariate and multivariate regression analyses.

Conclusions

This study contributes to the literature by describing clinical trial perceptions among patients in a precision oncology prospective cohort from a cancer center that serves a predominately white, Appalachian population. Addressing patients' perceptions and concerns related to trials and involving family or support networks in the decision making process is likely to be a key element of supporting informed patient decision‐making for future trial opportunities.

1. Background

Many cancer clinical trials fail due to insufficient enrollment [1]. A range of multi‐level trial enrollment barriers have been identified ranging from structural [1] (e.g., a lack of trial availability at treatment clinics or ineligibility for available trials), interpersonal [2, 3, 4] (e.g., provider knowledge of trial opportunities or unconscious biases, lack of support from patients' family and caregivers) [5, 6], and patient‐level (e.g., mistrust, misinformation, and poor understanding) [6, 7]. Patients from rural and under‐served areas experience a high degree of barriers, leading to insufficient trial accrual [1, 2, 8].

The field of precision oncology relies on clinical trials to develop and test targeted therapies [9]. Precision oncology protocols, such Total Cancer Care (TCC) are designed to reduce key structural barriers through utilizing national clinical trial databases (i.e., the Oncology Research Information Exchange Network (ORIEN) which includes samples from over 400,000 patients enrolled in partner institutions across the US) [10, 11]. All ORIEN alliance member organizations use the standard TCC protocol, a prospective cohort study in which participants agree to: sequencing (exome tumor, RNA, and germline), ongoing monitoring of clinical data, and permission for contact if an appropriate trial is identified within ORIEN. TCC has promise in reducing enrollment disparities with high rates of patient participation in the protocol. For example, our institution's (University of Kentucky Markey Cancer Center (UKMCC)) TCC protocol has successfully consented 90% of approached patients with no differences in consent rate based on Appalachian residence [8].

Although precision oncology protocols have tremendous promise to reduce access barriers, such data‐based solutions alone may not be sufficient for addressing the trial enrollment problem as the ultimate decision to enroll in a specific clinical trial lies with the patient. Patients typically join precision oncology protocols such as TCC early in treatment and they are not yet likely to have been invited to join a clinical trial. Their perceptions of trials at that time may be influential of their likelihood of agreeing to participate if offered a trial in the future. Although it is possible that patients who opt to enroll in precision oncology protocols may be more favorable to or inclined to enroll in trials, emerging studies of the implementation of such protocols have shown limited clinical impact on trial enrollment rates [12, 13]. Patient perceptions also have implications for designing clinical educational and decision‐aid programs to address the unique context of precision oncology cohorts in which patients are likely to receive trial invitations in the future rather than the prevailing approaches to supporting patient trial participation by delivering educational materials or decision aids at the time of offering a specific trial.

In the current study, we conducted a cross‐sectional survey to describe trial decision‐making factors and intentions at TCC enrollment. There is a lack of attention in the literature to clinical trial decision‐making variables among patients in the unique context of precision oncology protocols, perhaps due to the assumption that these patients would be positively inclined to participate in future trials. Our study is also unique in sampling patients from a cancer center with a high proportion of residents from underserved and predominantly rural Appalachian counties home to more than 26 million people (8% of the U.S. population) [8]. Rural and Appalachian patients have historically lower rates of clinical trial enrollment and are under‐represented in precision oncology research [8] yet trial decision‐making research is lacking among these populations.

2. Methods

The University of Kentucky Institutional Review Board approved the data collection. Survey inclusion criteria included: (1) age 18 years or older, (2) TCC enrollment, (3) primary diagnosis of gastrointestinal, head/neck, lung, or ovarian cancer (cancer types with a high availability of trials at UKMCC), and (4) no concern for significant cognitive impairment. Participants were recruited at time of enrollment in the TCC prospective cohort study at UKMCC. The UKMCC TCC program enrolls an average of 1200 patients annually [8]. Data was collected from August 2023 to July 2024. Study staff conducted electronic medical record (EMR) pre‐screening and provided eligible patients with a study recruitment packet at their next scheduled visit. Recruitment packets included a study information sheet, informed consent/HIPAA authorization, and instructions for completing the survey by paper (to return by mail) or online (QR code or unique password). The survey was designed to take 30 minutes with a $50 gift card incentive. A series of reminder letters and phone calls were used.

2.1. Survey Measures

Our survey is grounded in constructs drawn from Theory of Reasoned Action behavior shown to influence clinical trial decisions (i.e., beliefs, attitudes, knowledge, and social norms) and included valid and reliable survey items [7, 14, 15, 16, 17] (see Table S1: Table for items and responses). The clinical trial perceptions of interest were beliefs, attitudes, and knowledge of clinical trials. Beliefs about clinical trials related to family, trust, and costs were measured using five statements [14]. Attitudes about clinical trials were measured using items with a semantic differential scale responses including between bad/good, unimportant/important, and unsafe/safe [15]. Knowledge about clinical trials was measured with four statements [15]. A knowledge sum score was created by adding the number of correct responses from the statements with a range from 0 to 4. A normative belief is defined as the perception of what others expect or approve of and often impact behavior [17]. A normative belief score was computed as a product of responses to two questions about a participant's perceptions of favorability toward clinical trials among an important referent and their motivation to comply with this perception [17]. Intention of participating in a clinical trial was measured with a single item.¹⁶Appalachian residence was determined by applying Appalachian Regional commission county codes to participants' primary residence address. Cancer location and stage were obtained from EMR review.

2.2. Data Processing and Analysis

We first examined item response distributions for non‐normality and recoded response options to allow for adequate cell sizes. We recoded/created composite scales in accordance with best measurement practices and prior literature. We conducted one‐way ANOVAs to examine the relationship between each decision‐making construct and sociodemographic covariates. We then used linear regression analyses to examine the bi and multi‐variate relationships between intentions (outcome), sociodemographic variables, each clinical trial perception (beliefs, attitudes, knowledge, social norms). We utilized listwise deletion as there was minimal missing data (all variables had less than 6% missing).

3. Results

155 patients were provided with study packets: 117 participants completed the survey and 4 declined to participate (78% response rate). 116 patient surveys were analyzed (1 survey was removed for a repetitive response pattern). Most participants were White (97.4%), female (51.7%), older than 65 years (56.0%), and lived in Appalachian counties (59.5%) (Table 1).

TABLE 1.

Participant characteristics (n = 116).

Variable	N	%
Sex
Female	60	51.7
Male	56	48.3
Age
37–54	20	17.2
55–64	31	26.7
65–74	44	37.9
75+	21	18.1
Race
Black	2	1.7
White	113	97.4
Multi‐racial	1	0.9
Ethnicity
Hispanic	3	2.6
Non‐hispanic	112	96.6
County appalachian status
Appalachian	69	59.5
Non‐appalachian	47	40.5
Household income
$0–$19,999	24	21.8
$20,000–$39,999	25	22.7
$40,000–$79,999	34	30.9
≥ $80,000	27	24.5
Education
Less than high school	10	8.6
High school	46	39.7
Some college	35	30.2
4 or more years of college	25	21.6
Health insurance status
Private	51	45.1
Public/Uninsured a	62	54.9
Employment status
Employed b	30	26.5
Unemployed	9	8.0
Receiving disability	25	22.1
Retired	49	43.4

^a 4 participants were uninsured and were combined with public insurance to allow for adequate sample size for analyses.

^b Includes full‐ and part‐time employment and homemaker.

3.1. Trial Perceptions and Sociodemographic Correlates

The supplemental displays the survey items and sample means. Mean scores for beliefs related to trying new treatments (M = 3.0, SD = 1.3) and not trusting drug companies (M = 3.0, SD = 1.1) were neutral. However, participants, on average, disagreed with the belief regarding not trusting doctors (M = 1.8, SD = 1.0). Mean responses to the item describing worry about costs of trials were neutral (M = 3.1, SD = 1.4). Participants disagreed that they were worried their family wouldn't want them to participate in a trial (M = 2.3, SD = 1.1). Attitude items suggested most participants viewed clinical trials as good (M = 5.3, SD = 1.6), important (M = 5.4, SD = 1.5), and safe (M = 4.9, SD = 1.6). For knowledge, patients lacked knowledge related to finding trial information in ways other than asking doctors. Slightly more than half of participants knew they could leave a study without penalty. Most participants knew that trials are necessary to discover treatments and nearly 2 in 3 correctly identified that the statement about patients not needing to receive all trial information was false. For normative beliefs, participants felt important others viewed trials favorably and felt these opinions mattered. For clinical trial intentions, participants on average agreed that they would consider joining a clinical trial if offered (M = 3.5, SD = 1.1) (only 14% disagreed somewhat or disagreed strongly).

Comparisons of trial perceptions by sociodemographic factors revealed non‐significant associations by sex (Table 2). A significant association of age and the belief “I don't trust doctors” reflected more less mistrust among the oldest category of respondents (i.e., those aged 75+) compared to younger respondents. Compared to participants outside of Appalachia, Appalachia residents reported significantly higher agreement that their families would not want them to go on a trial (p = 0.02), lower overall knowledge scores (p = 0.02), and less favorable normative beliefs (p = 0.03). Both income and education were significantly associated with the belief item for worry about treatment costs and knowledge such that lower income and education were associated with significantly higher levels of worry about costs (income p = 0.04; education p = 0.02) and less knowledge about trials (income p = 0.008; education p = 0.04).

TABLE 2.

Sociodemographic correlates of clinical trial perceptions.

Survey measures		Perceptions and sociodemographic variables a (one‐way ANOVAs)
Construct	Item/scale	Sex	Age	Appala‐chian	Income	Edu‐cation	Insurance status
Beliefs	I don't like to try new treatments until they've been around for a while.	ns	ns	ns	ns	ns	ns
	I don't trust drug companies.	ns	ns	ns	ns	ns	ns
	I don't trust doctors.	ns	2.78* η 2 = 0.07	ns	ns	ns	ns
	I'm worried that I wouldn't be able to afford the costs of treatment on a clinical trial.	ns	ns	ns	2.80* η 2 = 0.08	3.30* η 2 = 0.08	ns
	I'm worried that my family wouldn't want me to go on a clinical trial.	ns	ns	5.28* η 2 = 0.05	ns	ns	ns
Attitudes	Mean scale b	ns	ns	ns	ns	ns	ns
Knowledge	Mean sum score c	ns	ns	5.72* η 2 = 0.05	4.17** η 2 = 0.11	2.91* η 2 = 0.08	ns
Normative belief	Normative belief X motivation to comply d	ns	ns	4.78* η 2 = 0.04	5.71*** η 2 = 0.15	ns	4.62* η 2 = 0.04

Abbreviation: ns = non‐significant (p > 0.05).

^a The coding of each sociodemographic variable response is consistent with the presentation in Table 1.

^b The three attitude items (show in Table S1) were averaged to create a single mean score for analyses (α = 0.84).

^c A total knowledge score for each participant was computed by summing the 4 knowledge items. The mean knowledge score across the 4 items was 2.4 (SD = 1.1).

^d Normative beliefs are calculated as the product of the two survey items related to normative belief and motivation to comply.

^*** p < 0.001.

^** p < 0.01.

^* p < 0.05.

3.2. The Association of Sociodemographic Variables, Trial Decision‐Making Factors, and Intentions

Table 3 shows the results from a series of bi‐ and multi‐variate linear regression analyses of the impact of clinical trial decision‐making factors on intention. In bivariate models, no sociodemographic variables were significantly associated with intentions. Agreement with the belief of not liking to try new treatments (p < 0.001, model R ² = 0.18) and agreement about worry with family trial preferences (p < 0.001, model R ² = 0.20) were negatively associated with clinical trial intentions. More favorable attitudes (p < 0.001, model R ² = 0.26) and higher knowledge scores (p < 0.001, model R ² = 0.10) were positively associated with intention. In a multivariate model, knowledge was no longer significantly associated with intentions and the association with attitudes was attenuated but remained significant. Of note, both education and worried about not being able to afford the costs of treatment on a clinical trial were significantly positively associated with intentions in the multivariate model.

TABLE 3.

Association of sociodemographic variables and clinical trial perceptions with trial intentions.

Survey measures		Regression analyses (intention as outcome)
		Bivariate				Multivariate
Construct	Variable/survey item/scale a	β (95% CI)	SE	p	R 2	β (95% CI)	SE	p
Socio‐demographic	Sex	−0.02 (−0.44, 0.41)	0.21	ns	—	0.02 (−0.33, 0.38)	0.18	ns
	Age	−0.08 (−0.30, 0.13)	0.11	ns	—	−0.05 (−0.24, 0.14)	0.09	ns
	Appalachian county of residence	−0.21 (−0.64, 0.22)	0.22	ns	—	0.18 (−0.19, 0.56)	0.19	ns
	Annual household income	0.13 (−0.07, 0.32)	0.10	ns	—	0.07 (−0.12, 0.26)	0.10	ns
	Education	0.20 (−0.03, 0.42)	0.12	ns	—	0.25 (0.03, 0.46)	0.11	0.025
	Insurance status	0.01 (−0.42, 0.45)	0.22	ns	—	0.00 (−0.38, 0.38)	0.19	ns
Beliefs	I don't like to try new treatments until they've been around for a while.	−0.37 (−0.51, −0.22)	0.07	< 0.001	0.18	−0.32 (−0.48, −0.17)	0.08	< 0.001
	I don't trust drug companies.	−0.15 (−0.33, 0.04)	0.09	ns	—	−0.04 (−0.23, 0.16)	0.10	ns
	I don't trust doctors.	−0.18 (−0.39, 0.02)	0.10	ns	—	0.05 (−0.14, 0.24)	0.10	ns
	I'm worried that I wouldn't be able to afford the costs of treatment on a clinical trial.	0.02 (−0.13, 0.17)	0.08	ns	—	0.28 (0.14, 0.42)	0.07	< 0.001
	I'm worried that my family wouldn't want me to go on a clinical trial.	−0.43 (−0.60, −0.27)	0.08	< 0.001	0.20	−0.33 (−0.51, −0.15)	0.09	< 0.001
Attitudes	Mean scale score	0.42 (0.28, 0.55)	0.07	< 0.001	0.26	0.21 (0.04, 0.37)	0.08	0.013
Knowledge	Mean sum score	0.33 (0.14, 0.53)	0.10	< 0.001	0.10	0.08 (−0.10, 0.27)	0.09	ns
Normative belief	Normative belief X motivation to comply	0.02 (−0.00, 0.05)	0.01	ns	—	−0.01 (−0.03, 0.01)	0.01	ns

Abbreviation: ns = non‐significant (p > 0.05).

^a Item coding and score computation for scales are described within Table 2 notes.

4. Discussion

This survey study is unique with the focus on patient clinical trial decision‐making in the context of a precision oncology protocol and the sampling of patients from the predominately rural and under‐served Appalachian region of the US. Study participants provided a range of responses to trial beliefs related to trying new treatments, mistrust of drug companies, and worry about costs. Most respondents indicated trust toward doctors, felt support from family members, held favorable attitudes toward trials, and were knowledgeable about key elements of trials. However, slightly more than half of participants reported positive intentions to join a trial with a third reporting a neutral intention. There were few systematic differences in perceptions based on key socio‐demographic characteristics except for income and education which were both associated with increased worry about the costs of trials, less knowledge, and lower perceived norms. Individuals from Appalachian regions were less knowledgeable and had less favorable normative beliefs.

Findings from prior studies, primarily conducted in samples of individuals living in urban areas or outside of the US, are largely consistent with current results with overall positive views of clinical trials among patients, gaps in some important barriers (e.g., mistrust, lack of knowledge), and some differences with regard to age, and socio‐economic status indicators [1]. Overall, our findings suggests that patients who join precision oncology protocols may not necessarily be more favorable to or have higher intentions to join trials compared to the general patient population, highlighting the need for expanding the data solutions in precision oncology to also consider patient trial decision‐making factors.

4.1. Clinical and Research Implications

Since patients typically enroll in TCC and similar precision oncology protocols early in treatment and before they are offered a trial, these protocols represent a unique access point for proactive patient education opportunities to engage and prepare patients to actively participate in informed consent discussions if they are presented with a research opportunity later in their course of treatment. This contrasts with prevailing approaches to supporting patient trial participation by delivering educational materials or decision aids at the time of offering a specific trial.

Lower knowledge levels were found among those with lower education, lower income, and Appalachian residence. Prior studies have shown trial participation is less likely among individuals who report a lack of knowledge of existing trials or how to get involved with one [18, 19]. In this study, knowledge was positively associated with trial intentions in a bivariate analysis but was no longer significant after controlling for other factors. Of note, the item “I'm worried that I wouldn't be able to afford the costs of treatment on a clinical trial” was not associated with intentions in bivariate models but was significantly positively associated in the multivariate model. This is an unexpected result as this item is designed to capture financial concerns which are a well‐established barrier for trial enrollment [1]. The meaning of this specific association is difficult to interpret in the context of the multivariate model that controls for socioeconomic variables, but it is possible that participants interpreted the prompt as a trial would be beneficial if they were unable to afford the costs of treatment. Prospective research is needed to understand the impact of trial knowledge, education status, and financial concerns on trial intentions as well as how intentions and treatment response relate to eventual trial enrollment.

Despite the well‐established intention‐behavior gap, intentions are among the best proxies for understanding future behavior [17]. Reluctance to try new treatments was associated with lower intentions while positive attitudes were associated with higher intentions. Patient education and decision‐support programs that are delivered early in the treatment process, rather than at the time of trial invitation as is typical, may provide an opportunity to prepare patients to ask questions and actively participate in informed consent discussions if presented with a research opportunity later in their course of treatment. Patients who feel more prepared or confident are likely to hold more positive attitudes and intentions to try trials if offered which could lead to higher enrollment rates.

Worry about family support for trial participation, highest among Appalachian patients and those with lower income, was negatively associated with trial intentions. Prior research demonstrates the role that family and caregivers can play in supporting or dissuading clinical trial engagement [20] and is likely to be beneficial to integrate family members and caregivers into precision oncology efforts through informational materials, educational programs, and communication resources designed to support patients in their decision making.

5. Limitations

Study participants were enrolled in a precision oncology protocol and may be positively inclined to participate in clinical trials. However, our cancer center's TCC program is universally offered and more than 97% of patients approached consent to the protocol. Also of note, the decline rate is not significantly different for patients from Appalachia versus other areas [8]. Our sample was economically diverse but, as is typical of the region, had few participants who were racial/ethnic minorities and findings may not generalize beyond the Appalachian region or similar rural areas. This research is unlikely to generalize to all individuals with cancer but rather is designed to explore perceptions of patients in precision oncology protocols to better understand this unique context. As precision oncology protocols become more common, it is important to replicate this study in other contexts. The study is cross‐sectional and there is an established intentions behavior gap. Future research is needed to explore the impact of perceptions and intentions at the time of joining a precision oncology protocol on actual future trial enrollment behaviors.

6. Conclusion

Clinical trial participation rates remain low despite the expansion of precision oncology programs aimed at reducing structural and clinical barriers. This highlights the need to ensure attention to patient‐centered factors related to trial decision making, particularly for historically underrepresented populations. This study contributes to the literature by examining the perceptions of trials and decision‐making factors associated with trial enrollment intentions within a sample patients enrolled in a precision oncology prospective cohort at a cancer center serving rural Appalachia.

Author Contributions

We attest that each listed author contributed to each of these 3 conditions: (1) made substantial contributions to conception and design, acquisition of data, or analysis and interpretation of data (details below), (2) drafted the article or revised it critically for important intellectual content, and (3) provided approval of the submitted manuscript. Substantial contributions to J.L.S., A.L., A.O., J.K., E.B., M.J.C.: conception and design. A.L., J.L.S., E.B.: acquisition of data. J.L.S., A.L., A.O., C.D., T.J.W.: analysis and interpretation of data.

Funding

This work was supported by National Cancer Institute awards P30CA177558 and R34CA313778 to the University of Kentucky. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Table S1: Item response percentages and means.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.