Abstract
Background
Studies using administrative data report that racial/ethnic minority patients enroll in clinical trials less frequently than white patients. We studied a cohort of terminally ill cancer patients to determine a) if racial/ethnic minority patients have lower rates of drug trial enrollment than white patients once socioeconomic characteristics are accounted for and b) what factors most influence drug trial enrollment among patients with advanced canceroverall.
Methods
: Coping with Cancer (CwC) is a National Cancer Institute/National Institute of Mental Health (NCI/NIMH)-funded multisite, prospective, longitudinal study of patients with advanced cancer. Baseline interviews assessed drug trial enrollment as well as socioeconomic characteristics. Logistic regression models estimated associations between drug trial enrollment and baseline characteristics. Stepwise, backward, and subset model selection was applied to select the final model where characteristics significant at α=0.05 remained in the model.
Results
At a median of 4.4 months prior to death, 35 of 358 patients (9.8%) were enrolled in a drug trial. In unadjusted analyses, race/ethnicity, health insurance, performance status, recruitment site, cancer type, preference for life-extending care, and lack of end-of-life care planning were associated (p<0.05) with enrollment. In multivariable analysis, patient race/ethnicity was not significantly associated with enrollment. Patients who reported not having an end-of-life discussion (adjusted odds ratio [AOR], 0.18; 95% confidence interval [CI] 0.04–0.83) and those not wanting to discuss life expectancy (AOR, 0.31; 95%CI 0.12–0.79) were more likely to be trial enrollees.
Conclusion
Patient race/ethnicity was not associated with clinical trial enrollment after adjustment for socioeconomic covariates. Patients with advanced cancer endorsing less engagement in end-of-life planning were more likely to be enrolled in a clinical trial.
Introduction
Enrollment of American cancer patients in clinical trials has remained modest despite significant progress in cancer care as well as the federal government's own commitment to increasing participation over the past two decades.1 Less than 5% of all adult cancer patients participate in a clinical trial.2–4 Low rates of enrollment delay discovery of potentially life-sustaining therapies.
Of particular concern is the oft-cited underenrollment of racial and ethnic minorities. Rates of minority enrollment are conflicting. Some studies conclude that racial minorities enroll in clinical trials at a significantly lower rate than their nonminority counterparts2,3; others suggest that racial minorities enroll in clinical trials at rates that parallel their respective incidence of disease.4
These prior studies used large databases that provide little additional information about the patient's social or economic identity. The lack of such information is problematic given that psychosocial and economic factors may influence patient decision making, particularly when a patient has a limited life-expectancy.
Recent research on cancer care disparities supports this idea. Bhargava and Du5 demonstrated that although black patients were less likely than their white counterparts to receive adjuvant chemotherapy for lymph-node-positive breast cancer at baseline, adjusted analyses that accounted for economic status attenuated the difference in care received. Similarly, Schwartz and colleagues6 found that both access to care and socioeconomic status ameliorated initial disparities observed in the cancer care and cancer-specific survival between black and white prostate cancer patients. Analyses of hospice utilization between white and black patients have also demonstrated similar results.7
The primary objective of this study was to examine the association between the race/ethnicity of patients with advanced cancer and their participation in drug trials. We hypothesized that enrollment in a drug trial would not be associated with racial/ethnic affiliation when socioeconomic and psychosocial factors were taken into account. Our second aim was to determine the factors most associated with drug trial enrollment overall in this cohort of patients with advanced cancer.
Patients and Methods
Coping with Cancer (CwC) is a multisite study of patients with advanced cancer designed to examine the psychosocial attitudes, mental health, and health care preferences of a cohort of terminally ill cancer patients. Patients were recruited from seven outpatient cancer facilities between September of 2002 and August of 2008: Dana-Farber/Partners Cancer Center (Boston, MA), New Hampshire Oncology-Hematology (Londonderry, NH), Memorial Sloan-Kettering Cancer Center (New York, NY), Parkland Hospital (Parkland, TX), Simmons Comprehensive Cancer Center (University of Texas-Southwestern in Dallas, TX), West Haven Veterans Affairs Comprehensive Cancer Center, (West Haven, CT), and Yale Cancer Center (New Haven, CT). The institutional review board of each facility approved the study protocol.
A trained rater interviewed patients and caregivers individually in face-to-face encounters. Each rater received training in interview practices prior to initiation of the study and met preestablished criteria (e.g., kappa ratings of 0.90 or higher on psychiatric diagnoses between new and trained interviewers) before being permitted to conduct interviews for the study. All interviews were conducted in English or Spanish.
To meet criteria for inclusion, patients had to meet the following requirements. First, eligible patients must have had a diagnosis of either an advanced pulmonary, gastrointestinal, or brain cancer. These cancer types were selected to ensure inclusion of patients who, by virtue of tumor biology, would have a limited life expectancy. For the purposes of this study, advanced cancer was defined as either the presence of distant metastases or locoregional disease refractory to first-line chemotherapy. Other requirements for enrollment included: age greater than 20 years to ensure recruitment of an adult cancer population using the National Institutes of Health (NIH) definition of children as an individual age 20 years or below, the use of English or Spanish as the patient's primary language, and the absence of significant cognitive impairment (e.g., dementia, delirium) as determined by physician assessment at the time of the baseline screening interview. Medical staff and the interviewer determined whether at the time of enrollment the patient had adequate stamina to complete the interview. All participants signed written informed consent.
Of 1015 eligible patients, 726 (71.5%) enrolled. For the purposes of this study, only patients who were deceased at the time of study closure and had nonmissing data for clinical trial enrollment at baseline were included to ensure selection of patients with terminal disease. This final cohort of 358 patients did not differ significantly (p<0.05) by cancer type, psychological distress, or rates of psychiatric disorders from the study participants at large. However, this cohort had a worse performance status, a higher symptom burden, and a lower baseline quality of life as would be expected in a cohort with shorter life expectancy. They were also more likely to be younger, female, uninsured, less educated, and to identify as a member of a racial or ethnic minority group.
Within the 358 patients included in this study, a subgroup sample with nonmissing data for the statistically significant confounders included in the multiple logistic regression analysis (n=325) had significantly (p<0.05) lower income and educational levels, was less likely to be married, less likely to have health insurance, less likely to be white or Jewish, and had worse performance status and higher symptom burden compared with the 33 patients not included in the final multiple regression analysis.
Primary outcome
Drug trial enrollment
Enrollment in a clinical trial was determined through chart review performed by a member of the patient's care team to answer the question, “Is the patient on a drug trial? If yes, what phase drug trial?”
Associations of interest
Patient race/ethnicity
As part of the patient interview, participants were asked: “What race or ethnicity do you consider yourself to be?” Responses were: “white,” “black,” “Asian-American, Pacific Islander or Indian,” “Hispanic,” “other,” “refused,” or “don't know.”
Other covariates of interest
Patient factors that could also influence a patient's enrollment in a clinical trial were examined as potential confounders. These included:
Sociodemographic traits
These included patient's self-reported sex, age, religion, marital status, education, income, and health insurance coverage.
Recruitment site
The site was the location where the patient received treatment, for example, the Dana Farber Cancer Institute, Parkland Hospital.
Cancer type
The cancer type was the tissue of origin, for example, lung, pancreas, gallbladder, colon, stomach, or esophagus.
Patient health
The patient's Karnofsky, Zubrod, and Charlson comorbidity index scores were documented by a member of the medical team. A high Karnofsky score implies a better performance status, whereas a low score on the Zubrod or Charlson comorbidity index reflects a higher performance status. The McGill Quality of Life Index was also administered by trained raters at the time of interview, where a high score implies a higher quality of life.8–11
Doctor-patient relationship
Patients were asked several questions about the relationship they had with their physician using The Human Connection Scale. Sample questions included: “How much do you trust your doctor?,” “How much do you respect your doctor?,” and “To what extent do you feel comfortable asking your doctor questions?” Answers were graded on a scale of 1 to 4, where a score of 1 corresponded to “Not at all” and a score of 4 corresponded to “A large extent.”12
End-of-life treatment preferences
Patients were asked about the kind of medical care they would like to receive at the end of their life. For example: “If you could choose, would you prefer: 1) treatment that focused on extending life as much as possible, even if it meant more pain and discomfort, or 2) care that focused on relieving pain and discomfort as much as possible, even if that meant not living as long?” They were also asked specific yes/no questions including: “Would you want to be kept alive if it required being on a feeding tube?” or “Would you want to be kept alive if it required you being on a breathing machine?”
Advance care planning
Patients were asked a series of yes/no questions to ascertain their preferences for end-of-life care. These questions included: “Have you and your doctor discussed any particular wishes you have about the care you would want to receive if you were dying?,” “If your doctor knew how long you had left to live, would you want him or her to tell you?,” “Do you have a signed living will, or health care proxy, or durable power of attorney for health care?,” and “Have you completed a do not resuscitate (DNR) order?”
Statistical analysis
Ttest, Cochran-Mantel-Haenszel test, and χ2 test statistics were used, as appropriate, to test for significant differences between patients who did or did not report clinical trial enrollment at baseline. Multivariable logistic regression modeling was used to examine the association between clinical trial enrollment and the baseline patient characteristics that were significant in bivariate analyses. Stepwise, backward, and subset model selection procedures were applied to select the remaining significant confounders with an entry significance level of α=0.2 and a stay significance level of α=0.1. Only baseline characteristics significant at α=0.05 remained in the model. Factors that were significantly related both to race/ethnicity and to trial enrollment, and that rendered the effect of race/ethnicity on trial enrollment not statistically significant (i.e., p>0.05) were considered confounders.
Statistical analyses were performed with SAS version 9.2 (SAS Institute, Cary, NC). We report two-tailed p values.
Results
Baseline patient characteristics: Univariate analysis
The cohort of patients with advanced cancer died a median of 4.4 months after the baseline assessment. Baseline characteristics are detailed in Table 1. In this study, 9.8% of patients (n=35) were enrolled in a drug trial at some point during their cancer care, of which 8 patients (22.9%) were enrolled in a Phase I study, 9 patients (25.7%) were enrolled in a Phase II study, 7 patients (20%) were enrolled in a Phase III study, and the remaining 11 (31.4%) were not specified.
Table 1.
Associations of Baseline Characteristics with Clinical Trial Enrollment
|
Clinical Trial Experience | |||||
|---|---|---|---|---|---|
| Baseline characteristics | Total (n=358) | Yes (n=35; 9.78%) | No (n=323; 90.2%) | OR (95% CI) | P value |
| Race | |||||
| White | 235 (66.01%) | 29 (12.3%) | 206 (87.7%) | 2.70 (1.09–6.69) | 0.032 |
| Black | 61 (17.13%) | 2 (3.3%) | 59 (96.7%) | 0.27 (0.06–1.15) | 0.077 |
| Hispanic | 53 (14.89%) | 2 (3.8%) | 51 (96.2%) | 0.32 (0.08–1.38) | 0.127 |
| Asian | 4 (1.12%) | 2 (50.0%) | 2 (50.0%) | 9.67 (1.32–70.89) | 0.026 |
| Sociodemographics | |||||
| Age; mean (SD) | 58.58 (12.66) | 56.09 (11.59) | 58.86 (12.76) | 0.98 (0.96–1.01) | 0.219 |
| Male | 194 (54.19%) | 15 (7.73%) | 179 (92.27%) | 0.59 (0.29–1.20) | 0.149 |
| Family Income ≥$31,000 | 113 (31.56%) | 11 (9.73%) | 102 (90.27%) | 1.62 (0.58–4.55) | 0.362 |
| Married | 217 (60.61%) | 24 (11.06%) | 193 (88.94%) | 1.41 (0.67–2.99) | 0.365 |
| Insured | 220 (61.45%) | 32 (14.55%) | 188 (85.45%) | 7.26 (2.18–24.22) | 0.001 |
| Education; mean (SD) | 12.55 (4.09) | 13.54 (3.58) | 12.44 (4.14) | 1.08 (0.98–1.18) | 0.130 |
| Religion | |||||
| Catholic | 132 (36.87%) | 16 (12.12%) | 116 (87.88%) | 1.49 (0.74–3.01) | 0.267 |
| Protestant | 61 (17.04%) | 3 (4.92%) | 58 (95.08%) | 0.43 (0.13–1.44) | 0.168 |
| Jewish | 17 (4.75%) | 5 (29.41%) | 12 (70.59%) | 4.29 (1.42–13.01) | 0.010 |
| Muslim | 4 (1.12%) | 1 (25.00%) | 3 (75.00%) | 3.12 (0.32–30.82) | 0.330 |
| No religion | 16 (4.47%) | 1 (6.25%) | 15 (93.75%) | 0.60 (0.08–4.68) | 0.626 |
| Pentecostal | 9 (2.51%) | 1 (11.11%) | 8 (88.89%) | 1.15 (0.14–9.48) | 0.896 |
| Baptist | 53 (14.80%) | 2 (3.77%) | 51 (96.23%) | 0.32 (0.07–1.38) | 0.127 |
| Recruitment site | |||||
| Yale Cancer Center | 68 (18.99%) | 21 (30.88%) | 47 (69.12%) | 8.78 (4.17–18.46) | 0.000 |
| Veterans Affairs CCC | 12 (3.35%) | 0 (0.00%) | 12 (100.0%) | --------a | 0.983 |
| Simmons Center | 34 (9.50%) | 3 (8.82%) | 31 (91.18%) | 0.88 (0.25–3.04) | 0.840 |
| Parkland Hospital | 141 (39.39%) | 4 (2.84%) | 137 (97.16%) | 0.17 (0.06–0.51) | 0.001 |
| DFCI/MGH | 8 (2.23%) | 1 (12.50%) | 7 (87.50%) | 1.32 (0.16–11.08) | 0.796 |
| New Hampshire Heme/Onc | 65 (18.2%) | 2 (3.1) | 63 (96.9) | 0.25 (0.06–1.07) | 0.061 |
| Cancer type | |||||
| Lung | 76 (21.23%) | 3 (3.95%) | 73 (96.05%) | 0.32 (0.10–1.08) | 0.066 |
| Pancreatic | 33 (9.22%) | 9 (27.27%) | 24 (72.73%) | 4.31 (1.82–10.24) | 0.001 |
| Gallbladder | 7 (1.96%) | 0 (0.00%) | 7 (100.0%) | --------a | 0.987 |
| Colon | 45 (12.57%) | 5 (11.11%) | 40 (88.89%) | 1.18 (0.43–3.21) | 0.747 |
| Brain | 8 (2.23%) | 2 (25.00%) | 6 (75.00%) | 3.20 (0.62–16.51) | 0.164 |
| Stomach | 14 (3.91%) | 1 (7.14%) | 13 (92.86%) | 0.70 (0.09–5.53) | 0.737 |
| Esophageal | 12 (3.35%) | 3 (25.00%) | 9 (75.00%) | 3.27 (0.84–12.70) | 0.087 |
| Performance status; mean (SD) | |||||
| Karnofsky Score | 64.32 (16.54) | 71.18 (18.05) | 63.58 (16.23) | 1.03 (1.01–1.06) | 0.012 |
| Zubrod Score | 1.72 (0.91) | 1.37 (0.94) | 1.75 (0.9) | 0.60 (0.39–0.92) | 0.020 |
| Charlson Index | 8.34 (2.68) | 7.03 (2.23) | 8.48 (2.69) | 0.80 (0.70–0.93) | 0.003 |
| McGill Quality of Life; mean (SD) | |||||
| Physical functioning | 5.61 (2.66) | 5.89 (1.99) | 5.58 (2.72) | 1.04 (0.91–1.20) | 0.524 |
| Symptoms | 5.36 (2.13) | 6.2 (1.6) | 5.27 (2.17) | 1.22 (1.04–1.44) | 0.016 |
| Psychological | 7.22 (2.5) | 7.46 (2.2) | 7.2 (2.53) | 1.04 (0.90–1.21) | 0.563 |
| Support | 8.63 (1.67) | 8.31 (1.44) | 8.67 (1.69) | 0.89 (0.74–1.08) | 0.239 |
| Summative score | 6.79 (1.55) | 6.98 (1.12) | 6.76 (1.59) | 1.10 (0.87–1.38) | 0.431 |
| Doctor-patient relationship | |||||
| How much trust your doctor; mean (SD) (n=121) | 3.34 (1.20) | 3.29 (1.50) | 3.34 (1.19) | 0.96 (0.52–1.78) | 0.90 |
| How much respect your doctor; mean (SD) (n=121) | 3.45 (1.18) | 3.29 (1.50) | 3.46 (1.16) | 0.89 (0.50–1.58) | 0.70 |
| Comfortable asking questions about your care (n=329) | 0.89 (0.24) | 0.87 (0.28) | 0.89 (0.24) | 0.74 (0.19–2.94) | 0.67 |
| EOL treatment preferences | |||||
| Extend life preference | 83 (23.18%) | 13 (15.66%) | 70 (84.34%) | 2.56 (1.15–5.71) | 0.022 |
| Against ICU death | 121 (33.80%) | 12 (9.92%) | 109 (90.08%) | 1.08 (0.51–2.32) | 0.836 |
| Prefer feeding tube | 125 (34.92%) | 10 (8.00%) | 115 (92.00%) | 0.76 (0.34–1.67) | 0.491 |
| Prefer ventilator | 84 (23.46%) | 8 (9.52%) | 76 (90.48%) | 1.04 (0.44–2.43) | 0.930 |
| Prefer chemotherapy | 248 (69.27%) | 27 (10.89%) | 221 (89.11%) | 4.28 (0.99–18.44) | 0.051 |
| Prefer antibiotics | 199 (55.59%) | 21 (10.55%) | 178 (89.45%) | 1.64 (0.70–3.82) | 0.255 |
| Advance care planning | |||||
| Terminal Illness acknowledgement | 125 (34.92%) | 9 (7.20%) | 116 (92.80%) | 0.62 (0.27–1.39) | 0.243 |
| End-of-life discussion | 151 (42.18%) | 6 (3.97%) | 145 (96.03%) | 0.25 (0.10–0.62) | 0.003 |
| Want doctor to tell life-expectancy | 236 (65.92%) | 16 (6.78%) | 220 (93.22%) | 0.39 (0.18–0.85) | 0.017 |
| Completed DNR order | 138 (38.55%) | 8 (5.80%) | 130 (94.20%) | 0.46 (0.20–1.06) | 0.067 |
| Had living will or DPA | 181 (50.56%) | 18 (9.94%) | 163 (90.06%) | 1.08 (0.51–2.28) | 0.843 |
Estimate could not be obtained.
DNR, do not resuscitate; DPA, durable power of attorney; EOL, end of life; ICU, intensive care unit; SD, standard deviation.
In univariate analysis, white or Asian race, Jewish religion, medical insurance, receiving care at Yale Cancer Center, primary pancreatic cancer, a better performance status, and fewer physical symptoms of illness were all significantly (p<0.05) associated with clinical trial enrollment (Table 1). Patients who preferred life-extending care, those who did not want their doctor to tell them their life expectancy, and those who reported not having had an end-of-life discussion with their physician, were also more likely to be enrolled in a clinical trial at baseline (all p<0.05).
Factors associated with clinical trial enrollment: Multivariable analysis
After adjustment for demographic and socioeconomic covariates including age, gender, marital status, insurance, and education, white race was not associated with clinical trial enrollment. Patients who reported that they did not have an end-of-life discussion with their physician (adjusted odds ratio, [AOR] 0.18; 95% confidence interval [CI] 0.04–0.83; p=0.03), those who did not want their doctor to tell them their life expectancy (AOR, 0.31; 95%CI 0.12–0.79; p=0.01), and those treated at Yale Cancer Center were more likely to be enrolled in a clinical trial. Patients with a primary lung cancer were less likely to be enrolled in a clinical trial (Table 2). The remaining factors identified in the unadjusted analysis did not meet criteria (p<0.05) for significance.
Table 2.
Adjusted Associations of Clinical Trial Enrollment with Baseline Predictors
| |
Clinical trial experience |
Adjusted result |
|||
|---|---|---|---|---|---|
| Baseline correlate | Total (n=325) | Yes (n=30, 9.2%) | No (n=295, 90.8%) | OR/HR | P value |
| End-of-Life discussion | 67 (20.6) | 2 (6.7) | 118 (98.3) | 0.18 (0.04–0.83) | 0.03 |
| Want doctor to tell life expectancy | 70 (21.5) | 16 (6.8) | 219 (93.2) | 0.31 (0.12–0.79) | 0.01 |
| Yale Cancer Center | 120 (36.9) | 21 (31.3) | 46 (68.7) | 12.26 (1.09–4.84) | <0.0001 |
| Lung cancer | 235 (72.3) | 3 (4.3) | 67 (95.7) | 0.17 (0.04–0.72) | 0.02 |
Note: 325 patients had nonmissing data with the four baseline correlates.
HR, hazard ratio; OR, odds ratio.
Confounders of the race/ethnicity effect on trial enrollment
Patient race/ethnicity was found to be statistically significant in predicting trial enrollment in the univariate analysis. Analyses were performed to test whether the other factors associated with clinical trial enrollment confounded the relationship between race and trial enrollment. The analyses compared patients who self-identified as “white” versus all other participants (“nonwhite”) because the individual racial/ethnic groups did not have an adequate number of trial participants to yield reliable estimates. A confounding effect of white race on the likelihood of trial enrollment was observed for patient recruitment site (Yale Cancer Center, Parkland Hospital) and health insurance coverage (Table 3). Other factors including education, Jewish and Baptist religion, and pancreatic cancer were found to be partial confounders of trial enrollment.
Table 3.
Associations between White Race and Clinical Trial Enrollment Controlling for Demographic and Psychosocial Characteristics
| No. of subjects | AOR (95% CI)a | P value | |
|---|---|---|---|
| White race (base model) | 356 | 2.70 (1.09–6.69)b | 0.03 |
| Plus age | 356 | 3.15 (1.24–8.00) | 0.02 |
| Plus male | 356 | 2.75 (1.11–6.83) | 0.03 |
| Plus health insured | 350 | 1.33 (0.50–3.58) | 0.57 |
| Plus education | 356 | 2.38 (0.89–6.41) | 0.09 |
| Plus Protestant religion | 356 | 2.94 (1.18–7.33) | 0.02 |
| Plus Jewish religion | 356 | 2.37 (0.94–5.97) | 0.07 |
| Plus Baptist religion | 356 | 2.42 (0.97–6.06) | 0.06 |
| Plus Yale Cancer Center | 356 | 1.96 (0.76–5.07) | 0.17 |
| Plus Parkland Hospital | 356 | 1.13 (0.40–3.22) | 0.82 |
| Plus New Hampshire Heme/Onc | 356 | 3.59 (1.44–9.00) | 0.006 |
| Plus Pancreatic cancer | 356 | 2.30 (0.91–5.79) | 0.08 |
| Plus Brain cancer | 356 | 2.62 (1.05–6.52) | 0.04 |
| Plus Lung cancer | 356 | 2.78 (1.12–6.93) | 0.03 |
| Plus Esophageal cancer | 356 | 2.72 (1.09–6.76) | 0.03 |
| Plus Karnofsky score | 348 | 2.56 (1.02–6.41) | 0.04 |
| Plus Zubrod score | 350 | 2.61 (1.05–6.50) | 0.04 |
| Plus Charlson Index | 355 | 2.96 (1.19–7.41) | 0.02 |
| Plus End-of-life discussion | 355 | 3.18 (1.27–7.97) | 0.01 |
| Plus want doctor to tell life-expectancy | 325 | 3.33 (1.22–9.05) | 0.02 |
| Plus general self-efficacy | 317 | 2.94 (1.09–7.95) | 0.03 |
| Plus extend life preference | 289 | 3.02 (1.08–8.40) | 0.03 |
| Plus prefer chemotherapy | 319 | 3.26 (1.20–8.83) | 0.02 |
| Plus completed DNR order | 322 | 3.73 (1.37–10.19) | 0.01 |
| Plus McGill symptoms | 354 | 2.99 (1.19–7.50) | 0.02 |
The adjusted ORs account for demographic and psychosocial variables.
The OR for the base model is unadjusted.
Sample sizes vary due to missing data.
AOR, adjusted odds ratio; CI, confidence interval; DNR, do not resuscitate.
Bold indicates no statistical significance.
Discussion
Consistent with prior studies of minority clinical trial enrollment, we found white patients were more likely to be enrolled in a clinical drug trial compared to nonwhite patients. Yet, in support of our hypothesis, when demographic and socioeconomic factors were accounted for, race did not remain a significant predictor of enrollment. Specifically, health insurance coverage and recruitment site proved confounders that explained the effect of race on clinical trial enrollment.
Although we are not the first to report the relationship between insurance status or location of care and clinical trial enrollment,3,13–15 our analysis does support the association of each with minority race/ethnicity and serves to emphasize that the discrepancy in drug trial enrollment among racial/ethnic minorities may be one of access to care and not of cultural acceptance.16 However, although the characteristics delineated above appear to mitigate the differences in clinical trial enrollment by patient race, it should not imply that bias, either explicit or implicit, does not exist or that there is no longer a need for clinical staff be vigilant about the need to enroll racial and ethnic minority cancer patients in clinic trials.
With respect to other factors that predicted clinical trial enrollment, patients with a primary lung cancer were significantly less likely to enroll in a clinical trial, whereas patients treated at the Yale Cancer Center, those not reporting having had an end-of-life discussion with one's physician, and those not wanting a physician to relay one's life expectancy were all positively associated with clinical trial participation. It is unclear why patients with lung cancer were less likely to enroll in a trial compared with the cohort at-large, although other logistical factors historically thought to impact trial enrollment and not captured by our study, including physician preference, trial availability, or distance from a participating trial center could have played a role.3 In reference to the association between Yale Cancer Center and trial enrollment, as mentioned in a prior report,17 this site appeared to treat all our enrolled patients more aggressively compared with other participating sites, with both higher rates of aggressive care and lower utilization of hospice care than other participating medical centers.
Of particular interest was the finding that all patients who had enrolled in a drug trial were significantly less likely to report having had an end-of-life discussion with their doctor and were also significantly less likely to want to know their life expectancy from their physician even after controlling for numerous potential confounders (Table 1). As all responses in our study relied on patient recall, we cannot know if end-of-life care conversations were never held between patients and their doctors or if patients simply did not remember such discussions. Yet, this is the first study that we know of to find a significant association between clinical trial enrollment and patient lack of interest in life expectancy and reporting of end-of-life discussions. This association does suggest an interaction between terminal illness engagement and the receipt of experimental medical care.
Previously published research about end-of-life care has shown that patients who do not report an end-of-life discussion with their physician are more likely to receive aggressive, costly medical care, and both aggressive care and higher health care costs are associated with a worse quality of death.17–19 Racial minorities in particular are more likely to present with advanced disease and more likely to receive aggressive end-of-life care.18,20 In light of the results of this report, it appears that both minority and majority ethnic status terminally ill patients who participate in clinical trials are likely to benefit from counseling regarding end-of-life care planning.
There are several limitations to this study. First, our study is limited by the modest number of patients who participated in a clinical trial. Consequently, we may have failed to detect some factors that are associated with trial enrollment that might emerge in larger samples. Second, our trial was restricted to patients who spoke either English or Spanish as their primary language. Patients belonging to racial or ethnic groups where a language other than English or Spanish is primarily spoken would have been excluded from our study and therefore, we can draw limited conclusions regarding clinical trial enrollment rates in these populations. Furthermore, the design of this trial did not permit the coding of ethnicity as a separate variable from race. Hence, this could have resulted in the grouping of a subset of patients with Hispanic ethnicity into a different category, for example, white or black. However, as patients were asked to self-identify using a list of choices, they were free to select the option with which they most identified. Therefore, we believe that the potential impact of this choice is unlikely to be profound.
Finally, whereas our study attempted to delineate some of the psychosocial factors that could influence trial enrollment in a cohort of patients with advanced cancer, a multitude of factors remain that could not be addressed within this analysis. For example, this study lacked information regarding patients' eligibility for, or knowledge of, available clinical trials. Therefore, we can draw no conclusions about the reasons why the participants in our study were or were not enrolled in a trial. We also have no information regarding a patient's enrollment in a nondrug trial as all patients were only screened for participation in a drug trial. Additionally, the timing of patient's trial enrollment and their diagnosis of advanced cancer is unknown, so some patients may have enrolled in a clinical trial before they were considered to have advanced disease, or after our baseline assessment, and these data were not captured in the present analysis.
Despite these limitations, our study demonstrates that patients with advanced cancer self-identifying as Asian, black, or Hispanic are equally likely to enroll in a clinical trial as their white counterparts when socioeconomic and clinical factors are taken into account. Additionally, patients with advanced cancer who do enroll in drug trials are less likely to report an end-of-life care discussion or to want to know their life expectancy than patients with no history of drug trial participation. Additional research is required to explore further nuances of the doctor-patient interaction as it relates to end-of-life care discussions and clinical trial decision making. In the interim, this study suggests a tension between advance care planning and drug trial participation. It also underscores the importance of end-of-life care planning for patients participating in clinical trials.
Author Disclosure Statement
No competing financial interests exist.
References
- 1.National Institutes of Health. NIH Policy and Guidelines on the Inclusion of Women and Minorities as Subjects in Clinical Research. http://grants.nih.gov/grants/funding/ women_min/guidelines_update.htm. [Oct 1;2011 ]. http://grants.nih.gov/grants/funding/ women_min/guidelines_update.htm
- 2.Murthy VH. Krumholz HM. Gross CP. Participation in cancer clinical trials. JAMA. 2004;291:2720–2726. doi: 10.1001/jama.291.22.2720. [DOI] [PubMed] [Google Scholar]
- 3.Ford JG. Howerton MW. Lai GY, et al. Barriers to recruiting underrepresented populations to cancer clinical trials: A systematic review. Cancer. 2007;112:228–242. doi: 10.1002/cncr.23157. [DOI] [PubMed] [Google Scholar]
- 4.Tejeda HA. Green SB. Trimble EL, et al. Representation of African-Americans, Hispanics, and whites in National Cancer Institute cancer treatment trials. J Natl Cancer Inst. 1996;88:812–816. doi: 10.1093/jnci/88.12.812. [DOI] [PubMed] [Google Scholar]
- 5.Bhargava A. Du XL. Racial and socioeconomic disparities in adjuvant chemotherapy for older women with lymph node-positive, operable breast cancer. Cancer. 2009;115(13):2999–3008. doi: 10.1002/cncr.24363. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Schwartz K. Powell IJ. Underwood W., III George J. Yee C. Banerjee M. Interplay of race, socioeconomic status, and treatment on survival of patients with prostate cancer. Urology. 2009;74:1296–302. doi: 10.1016/j.urology.2009.02.058. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Greiner KA. Perera S. Ahluwalia JS. Hospice usage by minorities in the last year of life: Results from the National Mortality Followback Survey. J Am Geriatr Soc. 2003;51:970–978. doi: 10.1046/j.1365-2389.2003.51310.x. [DOI] [PubMed] [Google Scholar]
- 8.Karnofsky DA. Burchenal JH. The clinical evaluation of chemotherapeutic agents in cancer. In: MacLeod CM, editor. Evaluation of Chemotherapeutic Agents. New York: Columbia University Press; 1949. pp. 199–205. [Google Scholar]
- 9.Oken MM. Creech RH. Tormey DC, et al. Toxicity and response criteria of the Eastern Cooperative Oncology Group. Am J Clin Oncol. 1982;5:649–55. [PubMed] [Google Scholar]
- 10.Charlson ME. Pompei P. Ales KL, et al. A new method of classifying prognostic comorbidity in longitudinal studies: Development and validation. J Chronic Dis. 1987;40:373–383. doi: 10.1016/0021-9681(87)90171-8. [DOI] [PubMed] [Google Scholar]
- 11.Cohen SR. Mount BM. Strobel MG, et al. The McGill Quality of Life Questionnaire: A measure of quality of life appropriate for people with advanced disease. A preliminary study of validity and acceptability. Palliat Med. 1995;9:207–219. doi: 10.1177/026921639500900306. [DOI] [PubMed] [Google Scholar]
- 12.Mack JW. Block SD. Nilsson M. Wright A. Trice E. Friedlander R. Paulk E. Prigerson HG. Measuring therapeutic alliance between oncologists and patients with advanced cancer: The Human Connection Scale. Cancer. 2009;115:3302–3311. doi: 10.1002/cncr.24360. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Baggstrom MG. Wagar SN. Sezhiyan AK, et al. Barriers to enrollment in non-small cell lung cancer therapeutic clinical trials. Thorac Oncol. 2011;6:98–102. doi: 10.1097/JTO.0b013e3181fb50d8. [DOI] [PubMed] [Google Scholar]
- 14.Klamerus JF. Bruinooge SS. Ye X, et al. The impact of insurance on access to cancer clinical trials at a comprehensive cancer center. Clin Cancer Res. 2010;15(16):5997–6003. doi: 10.1158/1078-0432.CCR-10-1451. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Lara PN. Higdon R. Lim N, et al. prospective evaluation of cancer clinical trial accrual patterns: Identifying potential barriers to enrollment. J Clin Oncol. 2001;19:1728–1733. doi: 10.1200/JCO.2001.19.6.1728. [DOI] [PubMed] [Google Scholar]
- 16.Wendler D. Kington R. Madans J. Van Wye G. Christ-Schmidt H. Pratt LA, et al. Are racial and ethnic minorities less willing to participate in health research? PLoS Med. 2006;3:e19. doi: 10.1371/journal.pmed.0030019. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Wright AA. Mack JW. Kritek PA, et al. Influence of patients' preferences and treatment site on cancer patients' end-of-life care. Cancer. 2010;116:4656–4663. doi: 10.1002/cncr.25217. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Loggers ET. Maciejewski PK. Paulk E, et al. Racial differences in predictors of intensive end-of-life care in patients with advanced cancer. J Clin Oncol. 2009;27:5559–5564. doi: 10.1200/JCO.2009.22.4733. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Zhang B. Wright AA. Huskamp HA, et al. Health care costs in the last week of life: Associations with end-of-life conversations. Arch Intern Med. 2009;169:480–488. doi: 10.1001/archinternmed.2008.587. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Barnato AE. Chung-Chou HC. Saynina O, et al. Influence of race on inpatient treatment intensity at the end of life. J Gen Intern Med. 2007;22:338–345. doi: 10.1007/s11606-006-0088-x. [DOI] [PMC free article] [PubMed] [Google Scholar]