Customize your JAMA Network experience by selecting one or more topics from the list below.
In the current issue of JAMA Oncology, Unger et al1 provide an overview from SWOG in response to the recommendations to modernize eligibility criteria for clinical trials developed by the American Society of Oncology (ASCO), Friends of Cancer Research (Friends), and the National Cancer Institute (NCI) to support less restrictive eligibility criteria to broaden participation in clinical trials.2-6
At this point, there is little doubt among oncology practitioners that addressing unnecessarily restrictive inclusion criteria in clinical trials not only would facilitate enrollment but also would provide an opportunity for improving patient diversity so that trial participants are representative of the patient treatment population. Although the spirit of the restrictions has been the fundamental protection of patient safety—primum non nocere or “first do not harm”—ongoing scientific and clinical advances highlight the need to modernize eligibility criteria.7 As mentioned by Unger et al,1 the joint statement by ASCO and Friends described the positive aspects of inclusiveness and issued detailed recommendations to safely build ways to include patients with brain metastases, HIV infection, prior cancer, organ dysfunction, and pediatric age in the clinical research enterprise.2,7
Unger et al1 describe results from a survey conducted between 2007 and 2011. At that time, 77 752 surveys were sent to patients using the online NexCura treatment decision-making tool, and 5499 responded to the surveys (7% response). Participation was limited to patients with breast, colorectal, lung, and prostate cancer. Breast and prostate cancer were the most prevalent cancer types (53% and 28%, respectively). Comorbid diseases included hypertension (35.0%), vision loss (16.6%), arthritis (15.3%), asthma (11.5%), hearing loss (11.2%), and prior cancer (10.2%). Comorbidities were inversely associated with the outcome of being offered the opportunity to participate in a clinical trial. The authors performed elegant analyses to describe the associations between the comorbidity summary measures and outcomes, simulations of trial participation rates, and sensitivity analyses controlled for demographics and socioeconomic status.
The authors found an inverse correlation between offers of clinical trial participation and the presence of comorbid conditions, even after they adjusted for demographics and socioeconomic factors. Nonetheless, the analysis was skewed by the patients they were able to analyze from the data set. Unger et al1 also noted that the survey required self-reporting, which may limit the scope of information provided, but they thought that the data provided were reliable. Once they controlled for demographics and socioeconomic status, comorbidities independently predicted the diminished likelihood that trial discussions, trial offers, or trial participation would occur. From the aggregated analysis, the authors concluded that decreasing barriers to trial participation through the liberalization of eligibility would have a positive influence on trial offers and trial participation.
To provide some context on the limitations of this study, it is important to clearly distinguish the differences between comorbidities and eligibility criteria. Many eligibility criteria are not comorbidities (eg, disease-related requirements and certain laboratory values). Conversely, comorbidities that are not eligibility criteria often contribute to clinician and patient decisions regarding clinical trial participation.8
Unger et al1 analyze the association between comorbidities and 3 outcomes: trial discussions, trial offers, and trial participation. The modeling then focuses on how allowing patients with more comorbidities would change trial participation. Extrapolating trial participation rates is somewhat problematic; the focus should be on trial discussions and trial offers. Although the end goal is to increase trial participation, that goal is dependent on patients being offered the opportunity to make an informed decision about whether to participate in a clinical trial. For some patients, the right decision for their individual circumstances may be to decline trial participation.
In the original publication of analyses from this survey,9 the authors wrote that, because of the number of statistical tests performed, P > .01 should be interpreted with caution. In this analysis, the P values presented for multivariate models, in which demographics and income levels are controlled for, range from P = .01 to P = .02.9 For both the trial discussion and trial offer questions, the P value was .02. Unger et al1 do not describe how multiple testing was controlled for, which is an important limitation when considering these results.
The analysis was based on a convenience sample rather than a defined sample population, and as a result, the demographics of the survey participants are an important limitation when interpreting these results.9 Respondents were overwhelmingly white (94.4%) and younger than 65 years (78%). Among respondents, 32% had an annual income of $100 000 or more, 31% had completed some graduate school, and 53% had breast cancer. Individuals could only participate in the survey if they were using an online decision tool in English, which is indicative of relatively high health literacy. Because participants in this survey were younger, wealthier, more educated, more health literate, and more likely to be white than the general cancer population, it is hard to discern the true effect of changes in comorbidity eligibility criteria. It is possible that comorbidities that disproportionately exclude minority or other underserved populations were not highlighted in this model owing to the limitations of the survey sample. Alternatively, changing eligibility criteria to ensure that patients are not excluded owing to the comorbidities identified in this analysis could have a greater influence among underrepresented populations.
As we look to the future, a variety of stakeholders continue to collaborate to identify ways of broadening trial participation. Several NCI staff participated in the ASCO-Friends effort and brought the final recommendations to NCI clinical trial program leaders and reviewers to discuss incorporation of the recommendations into NCI-supported trials. Various task forces scrutinized the measures incorporated into the Cancer Therapy Evaluation Program Clinical Protocol Templates. These templates are widely used by researchers across the country in the Experimental Therapeutics Clinical Trials Network and the National Clinical Trials Network.10
Expanding clinical trial participation is a dynamic and continuous effort to seamlessly adapt clinical trial design to current knowledge and should not be a 1-time event. Upcoming discussions will encompass evaluation of other eligibility criteria where the potential for improvement exists. These criteria include drug washout periods, number of prior therapies allowed based on the study phase, performance status, exclusion of concomitant medications without drug-drug interactions, and population-based laboratory reference ranges.
Many of the aspects considered in these studies are subject to variability given unique aspects of the disease and the agent considered, and the key aim of this work is to highlight the overall intention of clinical researchers to embed these flexible criteria at the time of clinical trial design and only include those absolutely required for patient safety and scientific validity. Alignment among all stakeholder groups, including the US Food and Drug Administration, ASCO, Friends, NCI, the pharmaceutical industry, and advocacy organizations is necessary to overcome barriers to implementation and increase awareness.
Unger et al1 found that key comorbid conditions are associated with clinical trial discussions, trial offers, and trial participation. Modernizing eligibility criteria by removing restrictions on comorbid conditions represents an important step forward in our efforts to enhance clinical trial participation across the population. However, eligibility criteria represent just 1 of many potential barriers to clinical trial enrollment.11,12 Continued efforts to address these barriers will allow oncology practitioners to discuss clinical trial enrollment with their patients in a more realistic environment and to guide and support them through the decision-making process, with the hope of new and more effective therapies.
Corresponding Author: S. Percy Ivy, MD, National Cancer Institute, 9609 Medical Center Dr, Suite 5W458, Bethesda, MD 20850 (email@example.com).
Published Online: January 10, 2019. doi:10.1001/jamaoncol.2018.5949
Conflict of Interest Disclosures: None reported.
Disclaimer: This editorial represents the opinions of the authors and not those of the National Cancer Institute.
Mishkin G, Arnaldez F, Ivy SP. Drivers of Clinical Trial Participation—Demographics, Disparities, and Eligibility Criteria. JAMA Oncol. Published online January 10, 2019. doi:10.1001/jamaoncol.2018.5949
Create a personal account or sign in to: