Abstract
Introduction
For lung cancer screening, available data are often derived from patients enrolled prospectively on clinical trials. We therefore investigated lung cancer screening patterns among individuals eligible for but not enrolled in a screening trial.
Patients and Methods
From February 2017 through February 2019, we enrolled subjects to a trial examining telephone-based navigation during low-dose computed tomography (LDCT) for lung cancer screening. We identified patients for whom LDCT was ordered, and who were approached but not enrolled in the trial. We categorized non-enrollment as patient declined or unable to contact. We compared characteristics and LDCT completion rates among these groups and the enrolled population using two-sample t-test and Chi-squared test.
Results
Among 900 individuals approached for participation (mean age 62 years, 45 percent female, 53 percent black), 447 patients were enrolled to the screening clinical trial. There were no significant demographic differences between the enrolled and non-enrolled cohorts. Among the 453 individuals not enrolled, 251 (55 percent) declined participation, and 202 (45 percent) were unable to be contacted despite up to six attempts. LDCT completion was significantly associated with enrollment status: 81% among enrolled individuals, 73% among individuals who declined participation, and 49% among individuals unable to be contacted (P<0.001).
Conclusion
In this single-center study, demographic factors did not predict participation in a lung cancer screening trial. Lung cancer screening adherence rates were substantially lower among individuals not enrolled in a screening trial, particularly among individuals unable to be contacted. These findings may inform the broader implementation of screening programs.
Keywords: adherence, communication, computed tomography, demographics, underserved
MicroAbstract
Data on lung cancer screening comes from prospective clinical trials. We determined characteristics and screening adherence among individuals eligible for but not enrolled in a trial. Although demographics did not differ, adherence was significantly lower among non-enrolled individuals, particularly those who could not be reached. These findings may inform the broader implementation of screening programs.
In multiple medical scenarios, clinical recommendations are derived from prospective clinical trials. Generally, clinical trials are designed to demonstrate efficacy—that is, the potential benefit in a highly controlled environment and well-characterized population. Effectiveness—the utility of a treatment in a real-world setting—may differ substantially.1 Patients enrolled to clinical trials may be highly selected and have fewer co-morbidities.2 Indeed, individuals who seek their clinical care at centers with greater clinical research activity, such as National Cancer Institute (NCI)-designated cancer centers, appear to have distinct characteristics from the broader population, including educational levels and socioeconomic status.3,4
Lung cancer screening with low-dose computed tomography (LDCT) in high-risk populations was endorsed by the United States Preventive Services Task Force (USPSTF) in 2013 and has been covered as a benefit by the Centers for Medicare and Medicaid (CMS) since 2015.5 Evidence for a reduction in lung cancer mortality and overall mortality from LDCT comes from the National Lung Screening Trial (NLST).6,7 This 50,000-plus person study randomized individuals ages 55–74 with ≥30 pack-year smoking history to annual LDCT or chest x-ray. Conducted at 33 primarily academic sites, the National Institutes of Health (NIH)-sponsored NLST enrolled patients who were more educated, less likely to be current smokers, and more likely to be white than the broader U.S. population meeting the same age and smoking eligibility criteria.8 Participation in the NLST required successful contact by study personnel and provision of written, informed consent. Similarly, participants in the European NELSON lung cancer screening trial had better self-reported health, were younger, more physically active, higher educated, and more often former smokers compared with eligible individuals who did not participate.9
The extent to which clinical trial cohorts are representative of broader populations remains a major question for cancer screening modalities. We therefore determined the characteristics and behavior of individuals who were eligible for but did not enroll in a prospective study of patient navigation for LDCT-based lung cancer screening.
Methods
Study design and setting
We invited individuals for whom a screening low-dose CT was ordered between February 27, 2017 (study activation) through February 11, 2019 (enrollment completion) to participate in a pragmatic randomized trial of telephone-based patient navigation for lung cancer screening within the Parkland Health and Hospital System (Parkland). Parkland is the integrated safety-net health system for Dallas County, Texas, providing care for more than one million under- and uninsured county residents through a central, 982-bed tertiary care hospital, specialty clinics, and 12 community-based primary care clinics.10,11 Parkland’s neighborhood-based clinics provide critical outreach for Dallas County, which has a population of 2.6 million and is one of the most ethnically diverse counties in the country (41% Hispanic, 24% African American).12 This diverse but highly vulnerable population has substantial risk factors for lung cancer. Parkland has an enterprise-wide EMR system (EPIC; Verona, WI) that allows electronic tracking of a wide array of patient characteristics and outcomes.
Standard of care at Parkland is system-wide; however, the process is largely opportunistic, with separate clinic structures for screening and follow-up, and lacks systematic measures to ensure appointment scheduling and receipt of care. Patients are identified for screening by the ordering clinician, most commonly the primary care provider. The order is placed using a designated template in which relevant details (age, tobacco history) are pre-populated when available. Once the order is placed, both patients and providers have the option to contact the Radiology Call Center to set up an appointment (inbound call). However, this option is employed rarely. In most cases, radiology department schedulers make outbound calls to the patient, using any and all available contact phone numbers and—as required—interpreters. If the patient is reached by phone, the patient is scheduled within the template and scheduling guidelines. If the outbound call goes to voicemail, a message is left providing the patient with an appointment date and time, as well as a call back phone number if any changes are needed. If there is no voicemail option, an appointment date and time are provided by mailed letter, with the scheduling date accounting for mail delivery time. Once an appointment is scheduled, reminders are sent through Audiocare (telephone) and MyChart (electronic patient portal), depending on patient opt-in. If no telephone information or portal account, the reminder is sent by letter. If a patient does not complete the appointment, the ordering provider is automatically notified via an EHR inbasket notification and provided the Radiology Call Center phone number to reschedule the patient or have the patient call to reschedule. Screening results are communicated to the referring physician. Providers have the option of referring patients with radiographic abnormalities suspicious for lung cancer to a central Lung Diagnostics Clinic staffed by Pulmonary Medicine physicians. Patients diagnosed with lung cancer are referred to thoracic surgery, radiation oncology, and/or medical oncology clinics.
Our randomized trial (NCT02758054) identified individuals for whom LDCT orders had already been placed.13 To limit screening of ineligible patients, the Parkland EMR-based order set required input of qualifying patient age and smoking history (data imported automatically from demographic and social history EMR fields when available). Study personnel received a listing of ordered LDCT on a daily basis. Letters were mailed to those patients who appeared eligible for the trial (i.e., eligible for LDCT and English- or Spanish-speaking) announcing the trial and providing a telephone number to opt out of participation by calling the study telephone number. The letter briefly described the trial and was jointly signed by the study principal investigator (D.E.G.) and the Senior Medical Director of Ambulatory Services at Parkland (N.S.). Study staff contacted by telephone those patients who had not opted out to discuss participation. Staff made up to six attempts, distributed across various times and days (Monday-Friday, 8 am to 8 pm; Saturday and Sunday, 9 am to 4 pm), to reach them. During these attempts, study personnel rotated the telephone numbers used, including work, home, mobile, and designated patient contacts, such that all available telephone numbers were tried. Staff described the trial in broad terms, including tracking of screening-related events and administration of surveys. Individuals who agreed to participate completed a baseline telephone survey of health-related beliefs and behaviors and were then randomized to usual care or telephone-based patient navigation.
Data collection
This study was approved by the UT Southwestern Institutional Review Board (IRB STU 122015–046). For all patients evaluated for participation in the randomized trial, we collected the following data: demographics (age, gender, race/ethnicity, primary language), tobacco history, comorbidities (scored by the Charlson-Deyo index14). We also recorded the date of LDCT order placement and date of LDCT completion. We categorized individuals who did not enroll in our trial as either (1) unable to contact, or (2) declined participation. Those who declined participation were further categorized according to whether they declined by contacting the study team after receiving an informational letter (opt-out) or declined at the time of invitational phone call.
Statistical analysis
We used descriptive statistics (means ± standard deviation [SD), number [%]) to report baseline characteristics and completion of initial ordered LDCT of individuals (1) enrolled on the trial, (2) did not enroll due to inability to contact, and (3) did not enroll due to declining participation. These characteristics between individuals enrolled and not enrolled were compared in univariable analysis using two-sample t-test for continuous variables and Chi-squared test for categorical variables.
Results
Among 900 individuals planned for initial LDCT approached for participation in the prospective screening navigation trial, a total of 447 patients (49 percent) were enrolled and 453 (51 percent) were not enrolled (Figure 1). In the overall population, mean age was 62 years, 45 percent were female, and 53 percent were black. Characteristics of individuals enrolled and not enrolled in the trial are shown in Table 1. There were no significant differences in age, race, gender, primary language, insurance status, or smoking history between the enrolled and non-enrolled cohorts.
Figure 1.
CONSORT diagram
Table 1.
Characteristics of (a) the enrolled subset (N=447), (b) the non-enrolled subset (N=453), and (c) all approached individuals (N=900).
| Enrolled N=447 n (%) | Not Enrolled N=453 n (%) | Overall N=900 n (%) | P_value | |
|---|---|---|---|---|
| Age, Years, Mean (SD) | ||||
| 62.0 (4.6) | 62.3 (4.8) | 62.2 (4.7) | 0.51 | |
| Sex | ||||
| Female | 204 (45.6) | 201 (44.4) | 405 (45.0) | 0.70 |
| Race/Ethnicity | ||||
| White (non-Hispanic) | 134 (30.0) | 154 (34.0) | 288 (32.0) | 0.31 |
| Black | 243 (54.4) | 236 (52.1) | 479 (53.2) | |
| Hispanic | 65 (14.5) | 54 (11.9) | 119 (13.2) | |
| Other/Unknown | 5 (1.1) | 9 (2.0) | 14 (1.6) | |
| Language | ||||
| English | 410 (91.7) | 418 (92.3) | 828 (92.0) | 0.76 |
| Spanish/Other | 37 (8.3) | 35 (7.7) | 72 (8.0) | |
| Marital Status | ||||
| Single | 197 (44.1) | 195 (43.0) | 392 (43.6) | 0.65 |
| Married/Common Law | 92 (20.6) | 112 (24.7) | 204 (22.7) | |
| Divorced/Legally Separated | 118 (26.4) | 109 (24.1) | 227 (25.2) | |
| Widowed | 36 (8.1) | 34 (7.5) | 70 (7.8) | |
| Other/Unknown | 4 (0.9) | 3 (0.7) | 7 (0.8) | |
| Insurance | ||||
| Charity/self-pay | 437 (97.8) | 444 (98.0) | 881 (97.9) | 0.79 |
| Commercial/Medicare/Medicaid | 10 (2.2) | 9 (2.0) | 19 (2.1) | |
| Smoking History† | ||||
| Current | 354 (79.2) | 356 (78.6) | 710 (79.0) | 0.90 |
| Former | 12 (2.7) | 13 (2.9) | 25 (2.8) | |
| Never | 49 (11.0) | 45 (9.9) | 94 (10.4) | |
| Missing | 32 (7.2) | 39 (8.6) | 71 (7.9) | |
Based on imported electronic health record data (which may have been over-ridden by ordering clinician)
Among the 453 individuals who were not enrolled in the trial, 202 (45 percent) were never contacted despite up to 6 attempts, and 251 (55 percent) declined participation. Compared to individuals who declined participation, individuals who we were unable to contact were either significantly more likely or have a near-significant trend to be male, and to have have commercial insurance/Medicare/Medicaid (Table 2). Among the 251 who declined participation, 58 (23 percent) contacted the study team proactively as an “opt out” after receipt of the invitation/informational letter, and 193 (77 percent) declined at the time of initial telephone contact, and.
Table 2.
Characteristics of non-enrolled cases according to whether they were unable to be contacted or declined participation.
| Unable to be contacted N=202 n (%) | Declined Participation N=251 n (%) | Overall N=453 n (%) | P value | |
|---|---|---|---|---|
| Age, Years, Mean (SD) | ||||
| 62.3 (5.0) | 62.3 (4.7) | 62.3 (4.8) | 0.51 | |
| Sex | ||||
| Female | 81 (40.1) | 120 (47.8) | 201 (44.4) | 0.10 |
| Race/Ethnicity | ||||
| White (non-Hispanic) | 75 (37.1) | 79 (31.5) | 154 (34.0) | 0.48 |
| Black | 97 (48.0) | 139 (55.4) | 236 (52.1) | |
| Hispanic | 26 (12.9) | 28 (11.2) | 54 (11.9) | |
| Other/Unknown | 4 (2.0) | 5 (2.0) | 9 (2.0) | |
| Language | ||||
| English | 184 (91.1) | 234 (93.2) | 418 (92.3) | 0.40 |
| Spanish/Other | 18 (8.9) | 17 (6.8) | 35 (7.7) | |
| Marital Status | ||||
| Single | 91 (45.0) | 104 (41.4) | 195 (43.0) | 0.07 |
| Married/Common Law | 44 (21.8) | 68 (27.1) | 112 (24.7) | |
| Divorced/Legally Separated | 56 (27.7) | 53 (21.1) | 109 (24.1) | |
| Widowed | 9 (4.5) | 25 (10.0) | 34 (7.5) | |
| Other/Unknown | 2 (1.0) | 1 (0.4) | 3 (0.7) | |
| Insurance | ||||
| Charity/self-pay | 195 (96.5) | 249 (99.2) | 444 (98.0) | 0.04 |
| Commercial/Medicare/Medicaid | 7 (3.5) | 2 (0.8) | 9 (2.0) | |
| Smoking History† | ||||
| Current | 168 (83.2) | 188 (74.9) | 356 (78.6) | 0.12 |
| Former | 15 (7.4) | 30 (12.0) | 45 (9.9) | |
| Never | 4 (2.0) | 9 (3.6) | 13 (2.9) | |
| Missing | 15 (7.4) | 24 (9.6) | 39 (8.6) | |
Based on imported electronic health record data (which may have been over-ridden by ordering clinician)
Adherence to LDCT differed significantly among enrolled and non-enrolled groups (P<0.001). Among the 447 enrolled individuals, 361 (81%) completed the initial ordered LDCT. Among the 250 individuals who declined participation, 181 (72%) completed the initial ordered LDCT. Among the 203 individuals we were unable to contact for participation, 99 (49%) completed the initial ordered LDCT.
Table 3 displays characteristics of non-enrolled individuals according to LDCT completion status. Overall, 281 of 453 individuals (58%) completed the initial ordered LDCT. Those who did not complete the initial ordered LDCT were significantly more likely to have commercial/Medicare/Medicaid insurance and to be active smokers. There were no significant differences in age, sex, race/ethnicity, language, or marital status.
Table 3.
Characteristics of non-enrolled individuals according to LDCT completion.
| Completed N=279 n (%) | Not Completed N=174 n (%) | Overall N=453 n (%) | P value | |
|---|---|---|---|---|
| Age, Years, Mean (SD) | ||||
| 62.6 (4.8) | 61.8 (4.8) | 62.3 (4.8) | 0.41 | |
| Sex | ||||
| Female | 126 (45.2) | 75 (43.1) | 201 (44.4) | 0.67 |
| Race/Ethnicity | ||||
| White (non-Hispanic) | 101 (36.2) | 53 (30.5) | 154 (34.0) | 0.39 |
| Black | 144 (51.6) | 92 (52.9) | 236 (52.1) | |
| Hispanic | 30 (10.8) | 24 (13.8) | 54 (11.9) | |
| Other/Unknown | 4 (1.4) | 5 (2.9) | 9 (2.0) | |
| Language | ||||
| English | 257 (92.1) | 161 (92.5) | 418 (92.3) | 0.87 |
| Spanish/Other | 22 (7.9) | 13 (7.5) | 35 (7.7) | |
| Marital Status | ||||
| Single | 121 (43.4) | 74 (42.5) | 193 (43.0) | 0.94 |
| Married/Common Law | 67 (24.0) | 45 (25.9) | 112 (24.7) | |
| Divorced/Legally Separated | 66 (23.7) | 43 (24.7) | 109 (24.1) | |
| Widowed | 23 (8.2) | 11 (6.3) | 34 (7.5) | |
| Other/Unknown | 2 (0.7) | 1 (0.6) | 3 (0.7) | |
| Insurance | ||||
| Charity/self-pay | 278 (99.6) | 166 (95.4) | 444 (98.0) | 0.002 |
| Commercial/Medicare/Medicaid | 1 (0.4) | 8 (4.6) | 9 (2.0) | |
| Smoking History† | ||||
| Current | 209 (75.0) | 147 (84.5) | 356 (78.6) | 0.01 |
| Never | 5 (1.8) | 8 (4.6) | 13 (2.9) | |
| Former | 35 (12.5) | 10 (5.7) | 45 (9.9) | |
| Missing | 30 (10.8) | 9 (5.2) | 39 (8.6) | |
Based on imported electronic health record data (which may have been over-ridden by ordering clinician)
Discussion
The generalizability of cancer clinical trials has come under question in recent years.15 Increasingly numerous and complex eligibility criteria and procedures result in study cohorts that have fewer comorbidities, better functional status, and greater knowledge base than the broader disease population.16,17 For interventional, therapeutic clinical trials, these discrepancies can have major effects on treatment feasibility, tolerability, and benefit. These differences may be particularly pronounced in oncology, where patients tend to be older and treatments tend to have greater toxicities. For other types of studies, such as diagnostic tests, differences between trial and non-trial populations may have different implications. For instance, low-risk interventions such as mammography are likely to be equally tolerated across patient groups. However, behaviors such as uptake and adherence might differ,18,19 particularly in association with clinical procedures that may be perceived as new or unfamiliar, as the case for LDCT-based lung cancer screening.
In the present study, we examined and compared the characteristics and behaviors of eligible individuals who were enrolled or not enrolled in a trial of lung cancer screening. The study was conducted in a medically underserved population, thereby providing an added opportunity to understand more fully potential differences between trial and non-trial patients. In this analysis, non-enrollment did not reflect failure to meet all inclusion/exclusion criteria, but rather inability to establish contact or declined participation. Because trial eligibility was was based on LDCT order placement—which in turn reflected USPSTF guidance for lung cancer screening (apart from the single additional requirement that participants speak English or Spanish)—both the enrolled and non-enrolled populations may be representative of lung cancer screening-eligible patients in other safety-net settings.
Notably, we found no significant differences in multiple demographic characteristics between enrolled and non-enrolled individuals. This finding echoes earlier observations that race, income level, and education level are not associated with interest in research opportunities.20 By contrast, some earlier studies have demonstrated that individuals from lower socioeconomic groups are less likely to participate in clinical trials.21 It has also been noted that, over time, older individuals are less likely to participate in clinical trials,22 which may represent a major limitation when studying diseases that occur predominantly in the elderly, such as lung cancer. Reasons for this discrepancy include a relatively socioeconomically homogeneous population and lack of direct income data, as well as a relatively limited age range, in the current study. Additionally, enrollment in our LDCT screening study entailed only additional phone calls rather than more complex interventions typical of therapeutic clinical trials, which may also impact these findings.
Overall, the non-enrolled group had lower adherence to LDCT than the enrolled cohort, suggesting that individuals interested in and committing to a research study are also more likely to commit to standard clinical care. Examining the non-enrolled group further, we found clear differences according to whether the individuals could not be reached or whether they declined participation. Importantly, those we could not reach were less likely to be married, a demographic characteristic associated with late disease presentation, under-treatment, and increased risk of death from cancer.23 They also had the lowest rates of adherence, with fewer than half of individuals completing the initial LDCT ordered. Unfortunately, we cannot determine from available data whether individuals were not reached because the listed phone number was incorrect, they did not have regular access to the phone, they actively chose not to answer the calls, or other factors. Clearly more information about this sizeable and high-risk population is needed.
What drives interest in clinical trial participation? Motivators may include access to promising, cutting-edge therapies; the opportunity to benefit others; access to free healthcare; and compliance with a clinician’s request.24,25 Conversely, perceived inconvenience, concerns about blinding or randomization, protocol complexity and stringency, or the unknowns of experimental therapy efficacy and toxicity may limit participation.26,27 It seems likely that the present trial may have come across to invited, eligible candidates as a low-risk, low-reward study. These individuals would have access to LDCT regardless of trial participation. While unlikely to pose physical harm, the serial surveys and increased tracking inherent to trial participation may not have been perceived as particularly beneficial either. Accordingly, it seems that understanding, convenience, and a wish to contribute to knowledge advances may have been key driving factors.
The main limitations of this study include the lack of data on reasons for LDCT non-completion and the single-center setting. Specifically, a safety-net medical system in Texas is likely to serve a different population demographically than is a safety-net system in other regions, or a medical system for insured persons in the same geographical area. Importantly, however, safety-net populations appear to have the greatest risk for lung cancer and may therefore stand to receive the greatest benefit from screening, prevention (via tobacco cessation), early detection.28,29 A key strength of our study is the closed nature of the medical system under investigation. Patients generally receive all aspects of their care within PHHS and are unlikely to undergo LDCT elsewhere.30 Thus our data on LDCT completion is likely representative of actual screening adherence.
In conclusion, we found meaningful differences in screening adherence among individuals according to participation in a research study. Furthermore, the reason for non-participation had a major impact on screening completion. While those individuals who declined participation were somewhat less adherent than the enrolled cohort, those who were never reached despite multiple attempts were far less likely than other individuals to complete the initial LDCT. Thus, ensuring reliable communication may be a discrete yet impactful means of improving adherence in this population. These findings may have potential applicability to care provision and clinical trial interpretation not only in the context of lung cancer screening, but in other settings as well.
Clinical Practice Points.
What is already known about this subject?
Low-dose computed tomography (LDCT)-based lung cancer screening decreases lung cancer-specific and all-cause mortality. However, these data are based on individuals eligible for and enrolled in screening clinical trials.
What are the new findings?
We studied a cohort of individuals eligible for but not enrolled in a lung cancer screening study. Compared to the enrolled population, this group had similar demographic characteristics. However, these individuals were significantly less likely to complete the initial ordered LDCT. Adherence was particularly low among those who could never be reached by trial staff.
How might it impact on clinical practice in the foreseeable future?
These real-world observations may impact the interpretation of screening trials. By identifying potential barriers in the screening process, they may also inform the broader implementation of screening programs.
Acknowledgements
The authors thank Ms. Dru Gray for assistance with manuscript preparation, and Helen Mayo, MLS, for assistance with literature searches.
Funding:
This work was supported by the Cancer Prevention Research Institute of Texas (CPRIT; RP160030 and PP190052 to HAH, SJCL, CDO, and DEG), a National Cancer Institute Midcareer Investigator Award in Patient-Oriented Research (K24CA201543-01; to DEG), and a Mentored Research Scientist Development Award (K01CA234425; to CDO). Additional support was received through the Harold C. Simmons Comprehensive Cancer Center (5P30 CA142543) and UT Southwestern Center for Patient-centered Outcomes Research (R24 HS022418).
Footnotes
Trial registration: NCT02758054
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
References
- 1.Glasgow RE, Lichtenstein E, Marcus AC. Why don’t we see more translation of health promotion research to practice? Rethinking the efficacy-to-effectiveness transition. Am J Public Health 2003;93:1261–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Kennedy-Martin T, Curtis S, Faries D, Robinson S, Johnston J. A literature review on the representativeness of randomized controlled trial samples and implications for the external validity of trial results. Trials 2015;16:495. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Onega T, Duell EJ, Shi X, Demidenko E, Goodman D. Determinants of NCI Cancer Center attendance in Medicare patients with lung, breast, colorectal, or prostate cancer. J Gen Intern Med 2009;24:205–10. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Onega T, Duell EJ, Shi X, Demidenko E, Gottlieb D, Goodman DC. Influence of NCI cancer center attendance on mortality in lung, breast, colorectal, and prostate cancer patients. Med Care Res Rev 2009;66:542–60. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Moyer VA, Force USPST. Screening for lung cancer: U.S. Preventive Services Task Force recommendation statement. Ann Intern Med 2014;160:330–8. [DOI] [PubMed] [Google Scholar]
- 6.Aberle DR, Berg CD, Black WC, et al. The National Lung Screening Trial: overview and study design. Radiology 2011;258:243–53. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Aberle DR, Adams AM, Berg CD, et al. Reduced lung-cancer mortality with low-dose computed tomographic screening. N Engl J Med 2011;365:395–409. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Aberle DR, Adams AM, Berg CD, et al. Baseline characteristics of participants in the randomized national lung screening trial. J Natl Cancer Inst 2010;102:1771–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Yousaf-Khan U, Horeweg N, van der Aalst C, Ten Haaf K, Oudkerk M, de Koning H. Baseline Characteristics and Mortality Outcomes of Control Group Participants and Eligible Non-Responders in the NELSON Lung Cancer Screening Study. J Thorac Oncol 2015;10:747–53. [DOI] [PubMed] [Google Scholar]
- 10.Pickens S, Boumbulian P, Anderson RJ, Ross S, Phillips S. Community-oriented primary care in action: a Dallas story. Am J Public Health 2002;92:1728–32. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Anderson RJ, Amarasingham R, Pickens SS. The quest for quality: perspectives from the safety net. Frontiers of health services management 2007;23:15–28. [PubMed] [Google Scholar]
- 12.Dallas County Profile. 2018. (Accessed September 17, 2019, at http://www.txcip.org/tac/census/profile.php?FIPS=48113.)
- 13.Gerber DE, Hamann HA, Santini NO, et al. Patient navigation for lung cancer screening in an urban safety-net system: Protocol for a pragmatic randomized clinical trial. Contemp Clin Trials 2017;60:78–85. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis 1987;40:373–83. [DOI] [PubMed] [Google Scholar]
- 15.Gerber DE, Pruitt SL, Halm EA. Should criteria for inclusion in cancer clinical trials be expanded? Journal of comparative effectiveness research 2015;4:289–91. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Garcia S, Bisen A, Yan J, et al. Thoracic Oncology Clinical Trial Eligibility Criteria and Requirements Continue to Increase in Number and Complexity. J Thorac Oncol 2017;12:1489–95. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Ellis PM, Butow PN, Tattersall MH, Dunn SM, Houssami N. Randomized clinical trials in oncology: understanding and attitudes predict willingness to participate. J Clin Oncol 2001;19:3554–61. [DOI] [PubMed] [Google Scholar]
- 18.Caplan LS, Wells BL, Haynes S. Breast cancer screening among older racial/ethnic minorities and whites: barriers to early detection. J Gerontol 1992;47 Spec No:101–10. [PubMed] [Google Scholar]
- 19.Duke SS, Gordon-Sosby K, Reynolds KD, Gram IT. A study of breast cancer detection practices and beliefs in black women attending public health clinics. Health Educ Res 1994;9:331–42. [DOI] [PubMed] [Google Scholar]
- 20.Rasco DW, Xie Y, Yan J, et al. The impact of consenter characteristics and experience on patient interest in clinical research. Oncologist 2009;14:468–75. [DOI] [PubMed] [Google Scholar]
- 21.Unger JM, Gralow JR, Albain KS, Ramsey SD, Hershman DL. Patient Income Level and Cancer Clinical Trial Participation: A Prospective Survey Study. JAMA oncology 2015:1–3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Ludmir EB, Mainwaring W, Lin TA, et al. Factors Associated With Age Disparities Among Cancer Clinical Trial Participants. JAMA oncology 2019. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Aizer AA, Chen MH, McCarthy EP, et al. Marital status and survival in patients with cancer. J Clin Oncol 2013;31:3869–76. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Ellis PM. Attitudes towards and participation in randomised clinical trials in oncology: a review of the literature. Ann Oncol 2000;11:939–45. [DOI] [PubMed] [Google Scholar]
- 25.Bevan EG, Chee LC, McGhee SM, McInnes GT. Patients’ attitudes to participation in clinical trials. Br J Clin Pharmacol 1993;35:204–7. [PMC free article] [PubMed] [Google Scholar]
- 26.Mills EJ, Seely D, Rachlis B, et al. Barriers to participation in clinical trials of cancer: a meta-analysis and systematic review of patient-reported factors. Lancet Oncol 2006;7:141–8. [DOI] [PubMed] [Google Scholar]
- 27.Ross S, Grant A, Counsell C, Gillespie W, Russell I, Prescott R. Barriers to participation in randomised controlled trials: a systematic review. J Clin Epidemiol 1999;52:1143–56. [DOI] [PubMed] [Google Scholar]
- 28.Laaksonen M, Rahkonen O, Karvonen S, Lahelma E. Socioeconomic status and smoking: analysing inequalities with multiple indicators. Eur J Public Health 2005; 15:262–9. [DOI] [PubMed] [Google Scholar]
- 29.Bach PB, Cramer LD, Warren JL, Begg CB. Racial differences in the treatment of early-stage lung cancer. N Engl J Med 1999;341:1198–205. [DOI] [PubMed] [Google Scholar]
- 30.Lee SJC, Jetelina KK, Marks E, et al. Care coordination for complex cancer survivors in an integrated safety-net system: a study protocol. BMC Cancer 2018;18:1204. [DOI] [PMC free article] [PubMed] [Google Scholar]