- © 2010 by American Society of Clinical Oncology
Time to Activate Lung Cancer Clinical Trials and Patient Enrollment: A Representative Comparison Study Between Two Academic Centers Across the Atlantic
- Andrea Wang-Gillam,
- Kristina Williams,
- Silvia Novello,
- Feng Gao,
- Giorgio V. Scagliotti and
- Ramaswamy Govindan
- From the Washington University School of Medicine, St Louis, MO; University of Torino, Torino; and San Luigi Hospital, Orbassano, Italy
- Corresponding author: Ramaswamy Govindan, MD, Washington University School of Medicine, 660 South Euclid Ave, Campus Box 8056, St Louis, MO 63110; e-mail: rgovinda{at}dom.wustl.edu.
-
Presented in part at the 45th Annual Meeting of the American Society of Clinical Oncology, May 29-June 2, 2009, Orlando, FL, and at the 13th World Conference on Lung Cancer July 31-August 4, 2009, San Francisco, CA.
Abstract
Purpose Activation of clinical trials is a lengthy process. We studied the procedures and time required to activate lung cancer clinical trials in a US academic center compared with a European center.
Methods A retrospective review was performed of all thoracic oncology therapeutic trials submitted for regulatory review between 2001 and 2008 at Washington University School of Medicine (WUSM; St Louis, MO) and the University of Torino (UT; Torino, Italy). A process map was drafted by both institutions to establish the order of required events.
Results We reviewed 137 therapeutic thoracic oncology trials from WUSM (n = 83) and UT (n = 54). The median times from submission to opening a trial were 163 days for WUSM and 112.5 days for UT (P = .048). The median times for regulatory approval were 75 days for WUSM and 31 days for UT (P < .001). The difference is more pronounced in a homogeneous subset of phase II, industrial-sponsored trials for the median calendar time from submission to opening a trial (239.5 days for WUSM v 112.5 days for UT; P < .001) and time for regulatory approval (99 days for WUSM v 13.5 days for UT; P < .001). The median number of patients accrued at WUSM was 7.4 patients per study compared with an average of 37 patients per study at UT. The proportion of trials that enrolled 20 patients or more represented 22.2% of trials at UT but only 1.1% of trials at WUSM.
Conclusion It takes additional steps and significantly longer time to activate a therapeutic thoracic clinical trial at a representative US site (WUSM) compared with a European site (UT).
INTRODUCTION
Well-designed clinical trials are the only means to developing rational, effective, and safe therapy. In the evolving field of cancer therapy, clinical trials are critical components of patient care. Clinical trials are often the only option for patients with advanced cancer, for whom standard therapies are often ineffective, toxic, or both. Unfortunately, despite the pivotal role of clinical trials in cancer care, the enrollment of adult patients with cancer in the United States has been consistently low, frequently reported to be approximately 4%.1 A vast majority of adult patients with cancer are never enrolled in a clinical trial. Lack of adequate infrastructure to conduct clinical trials, lack of motivation on the part of busy practicing physicians, and perceived and (to some extent) real regulatory obstacles associated with the conduct of clinical trials may account for the intractable problem of persistent low accrual to clinical trials. Even the committed and motivated clinical trialists are increasingly wary of the length of the time it takes to open a trial to patient enrollment and the mountain of paperwork and processes required along the way before completing patient enrollment.2
The impact of the administrative process on activating oncology clinical trials has been outlined in detail.3–5 Lengthy and laborious procedures are recognized in the United States as barriers to activating oncology clinical trials, requiring up to 110 individual steps performed by as many as 27 participants.4 A previous report revealed that one quarter of the procedures and steps required for clinical trial activation at an academic cancer center do not appear to add value to the overall process.4
The National Cancer Institute (NCI) has established the Clinical Trials Working Group to reconstruct the national cancer clinical trials enterprise. To expand the participation of cancer clinical trials and enhance collaborative efforts, four informatics-focused initiatives were implemented; however, specific guidelines on how to shorten the process for clinical trial activation have not been established. While embracing this effort to improve the clinical trial activation process in the United States, we felt it would be valuable to learn from clinical trial operations at a center outside the United States (University of Torino [UT], Torino, Italy) and compare the processes with an academic center from the United States (Washington University School of Medicine [WUSM], St Louis, MO).
METHODS
Sites
The thoracic oncology divisions of two large academic institutions (UT and WUSM) are compared in this study. The oncology practice of WUSM is carried out at the Alvin J. Siteman Cancer Center, an NCI-designated comprehensive cancer center. It is one of the largest centers in the region, diagnosing 800 new patients with lung cancer each year. The clinical activities of the UT center occur at San Luigi Hospital, where 160 new patients with lung cancer are diagnosed and treated annually.
Process Comparison
The first phase of study began with the identification of the key steps required for the approval of therapeutic clinical trials at each institution. Each institution was asked to identify the major steps required to activate a therapeutic clinical trial and create a simplified flow chart describing the process. Although more detailed graphical representations of the clinical trial process and its intricacies have been created in other studies, the focus of this project was to create simple maps to more readily identify the macro-level activities at each facility and more readily point out the major differences between WUSM and UT. Because both institutions are geographically distant from one another and mapping by a single individual was not feasible, simplifying the map allows for a level of consistency that may not have been guaranteed if each institution had attempted a more detailed chart. Demographic information for each trial, including the enrollment status, phase, and sponsorship, was collected.
Timing and Accrual Comparison
The second phase retrospectively reviewed key time points for all thoracic, therapeutic clinical trials submitted for approval at each institution between January 2001 and November 2008. The calendar date for the execution of each step was extracted, and the duration of each major process was calculated. The data were reported by each institution, and 137 studies were included in this analysis. The major steps established based on the process flow charts created in the first phase were used as modes of comparison. Regulatory approval is defined as the interval from submission to either a scientific review committee (SRC) or institutional review board (IRB) to final IRB approval for patient enrollment. Cancer-related research studies in all NCI-approved cancer centers in the United States need to be approved by the SRC before submission to the IRB. Contract approval is defined as the time from notifying the institutional contracting office to receiving signed contracts from the sponsors. Activating the study was defined as the time point when study was open to patient enrollment. The first accrual was defined as the registration of the first eligible patient to the trial. Accrual data for each trial was also reviewed for each study. We included studies that began patient enrollment from January 2001 through January 2009. Only data for trials that had permanently closed to enrollment at the institution were included in the final analysis for enrollment evaluation.
The majority of trials performed at each institution were phase II, industry-sponsored clinical trials, so an additional subset timing analysis of these studies was performed. Trials that were multiphase were excluded from this subset analysis.
Statistical Analysis
Summary statistics (means, medians, and ranges for continuous variables; counts and proportions for categorical data) were calculated for timing and accrual data in each institution. The between-institution differences were compared using the Kruskal-Wallis rank sum test or Fisher's exact test as appropriate.
RESULTS
Protocol Demographic Comparison
The variation between the two institutions extends beyond the process to the types of trials that each chooses to participate in. Overall, data from 137 trials from both institutions were available for the analysis. The types of trials that were opened at each institution varied, as seen in Table 1. The majority of clinical trials conducted in both institutions were phase II studies and were sponsored by the pharmaceutical industry.
Demographic Description of the Evaluated Studies
Process Mapping
The process maps created by each institution are shown in Figure 1. The more streamlined process used at UT is apparent in Figure 1, where a single committee is responsible for securing both contracts and budgets, and a single review board is used for trial approval process. Overall, WUSM has a total of eight separate processes compared with six at UT.
Process flow developed by (A) Washington University School of Medicine and (B) University of Torino. SRC, scientific review committee; IRB, institutional review board.
Timing Analysis
Figure 2 outlines the timeline differences between the two institutions. The differences between the two groups were statistically significant in five of the six identified intervals. The median time to activate a trial (submission to open enrollment) was 163 calendar days at WUSM and 112.5 calendar days at UT (P = .048; Appendix Table A1, online only). The median time for regulatory approval was 75 days at WUSM and 31 days at UT (P < .001). However, the median time for the contract approval process was identified as the most time-consuming step in the process, with a median of 115 days at WUSM and 97.5 days at UT (P = .011). The time from a protocol being open to enrollment to the time when the first patient is registered on the trial was longer at WUSM compared with UT (51 v 24 days, respectively; P = .005). The only nonstatistically significant interval compared was the time from a protocol receiving regulatory approval to being opened to enrollment at an institution, with a median of 83 days at WUSM and 90 days at UT (P = .75).
Timeline for activating a thoracic oncology clinical trial, showing median number of days per process at each institution (University of Torino, n = 54; Washington University School of Medicine, n = 83).
Phase II, Industry-Sponsored Subset Analysis
A more homogeneous subset of phase II, industry-sponsored studies (29 from WUSM and 34 from UT) was reviewed (Fig 3; Appendix Table A2, online only). The median time required to activate a clinical trial was 239.5 days at WUSM and 112.5 days at UT (P < .001). The median time for regulatory approval was 99 days at WUSM and 13.5 days at UT (P < .001). In this subset, the longest process was the median duration from regulatory approval to activation of the trial, reported as 114 days at WUSM and 99.5 days at UT. These differences are not statistically significant (P = .2075).
Timeline for activating a phase II, industry-sponsored thoracic oncology clinical trial, showing median number of days per process at each institution (University of Torino, n = 34; Washington University School of Medicine, n = 29).
Enrollment Comparison
The average enrollments per study were also compared across both institutions (Table 2). UT enrolled an average of 37 patients per study compared with an average of 7.4 patients per study at WUSM. A significantly greater proportion of studies were closed with no patient enrollment at WUSM (21.8%) compared with only 5.6% at UT. UT also reports a greater percentage of studies with greater than 20 enrollments (22.2% of all studies v 1.1% of studies at WUSM). The data reported by WUSM are similar to previous data reported by other centers (Vanderbilt-Ingram Cancer Center, Ohio State University Medical Center, Fox Chase Cancer Center, University of North Carolina Lineberger Cancer Center) in the United States, as noted in Table 3.
Accrual for Studies Closed to Enrollment
Comparing Accrual Data With Prior Publications
DISCUSSION
The median time from the submission of the protocol for regulatory approval to opening the trials at WUSM (163 days) is comparable to the recently reported times required for activating a clinical trial in an academic center (171 days) and a community cancer center (191 days) in the United States.4 Thus, this prolonged time to activate a trial may closely represent the average time to activate an oncology trial in the United States, and the favorable shorter duration of trial activation at a single European site (UT) merits some attention. Obviously, the extremely wide ranges found for the time to activate a trial for both institutions pose a challenge for an investigator when planning study participation because it is difficult to know with any accuracy when a trial will be open for patient enrollment.
Significant differences were found between the two institutions in the regulatory approval process. In contrast to two sequential steps at WUSM, regulatory approval is simplified into a single-step process at UT. For regulatory approval at WUSM, the protocol needs to receive SRC approval before submission to the IRB. Even if no contingencies or questions are sent back to the investigator by the reviewers from the SRC (a rare occurrence), the process would take a minimum of 2 months. At UT, the single-step method may account for the shorter approval time (31 days for UT v 75 days for WUSM; P < .001).
It is time to seriously rethink the approval process required for activation of clinical trials by improving the efficiency without compromising the integrity of scientific and ethical review processes. A step in the right direction has been taken by the NCI in consultation with the Department of Health and Human Services Office for Human Research Protection in creating the central IRB (CIRB). The CIRB provides the initial full board review of multicenter cancer trials sponsored by cooperative groups, and it can eliminate the duplication of efforts by local IRBs by evaluating the protocol at a central location. Furthermore, the protocols approved by CIRB are handled administratively at the local SRCs, eliminating the lengthy and duplicative peer review process at individual sites. The CIRB handles the continuing review, processes amendments, and evaluates adverse events for these protocols. The local IRBs are still responsible for overall performance of research studies. This initiative has been welcomed by the American Society of Clinical Oncology6 and has been adapted by all phase III cooperative group studies. Although the CRIB is associated with faster review and less IRB staff effort,7 it is uncertain at the moment whether it truly shortens the time of trial activation.
Budgeting and the contract processes are completed by two separate offices at WUSM but by one entity at UT, perhaps contributing to a longer contract approval time at WUSM (115 days v 97 days at UT; P = .011). It is also the most time-consuming activity within the clinical trial activation process when reviewing the entire data set, which is consistent with previous reports.4
Clearly, determining how to effectively work with industry remains a global challenge. Industry-sponsored trials have been reported as constituting as much as 79% of all trials.8 In this data set, these trials account for 94.4% of trials at UT and 53.0% of trials at WUSM. A global solution is needed to shorten the time for budget and contract in general. The master clinical trial agreement (MCTA), a standard contract between an industrial sponsor and an academic institution, has been proposed in the past to simplify the process.9 The benefit of MCTAs is that they speed the process for the industrial trial by initial negotiation of a single, successful MCTA applicable to all studies from the same sponsor. This concept has been adapted in some institutions, including ours, but the impact of having a master contract in improving the efficiency of activating clinical trials is unknown. The contract process varies considerably at different institutions,8 and given the complexity of intellectual property, a master contact may be a necessary, unavoidable step.
Clinical trials are not necessarily activated immediately after receiving regulatory approval. In fact, this phase could be the longest segment in the process of activation of a trial, requiring a median of more than 3 months (114 days at WUSM and 99.5 days at UT). Several factors contribute to this delay, including time to arrange for site initiation visits, training of study personnel at the sites by the sponsor, requirements for approval of new amendments, and arrival of study drugs. Unfortunately, these time-consuming individual processes often can only happen sequentially. No statistically significant difference was found between the two institutions for this phase (P = .2075), which may indicate that this is a universal problem. Working with pharmaceutical sponsors to identify new ways to shorten this time frame and collecting information on the cause of these delays prospectively will help to shorten the length of the overall process.
The poor accrual record at WUSM (21% of studies closed with no enrollment) is similar to records reported earlier from other US sites (Table 3).4,10 Although the reasons for low accruals at these institutions are being explored and barriers to patient enrollment have been discussed in the literature, the low rate of accrual is quite concerning.11 It is known that third-party payers have disapproved patient enrollment in clinical trials.12 At least 8% of eligible patients declined participation in an oncology research study as a result of insurance denial.11 In Europe, the socialized health care system may have a standard approach to clinical trials that removes this additional barrier. For example, the reimbursement structure in Europe was found to contribute to the accrual of lymphoma patients.12 It has also been reported that patients in the United States choose not to participate in clinical trials as a result of a belief that their insurance will not cover the cost of treatment or that they will have to pay significant out-of-pocket expenses.11
With globalization of clinical trials, it is necessary to understand the process of clinical trial activation and accrual in other countries. Biopharmaceutical clinical trials are continuing to spread globally, and increases in the average relative annual growth rate have been noted for many countries outside of the United States and Europe.13 With the current trend toward large international trials, which require multicenter collaboration across national boundaries, streamlined regulatory and contract processes will be extremely advantageous for timely clinical trial activation. Perhaps a standardized global regulatory system can be considered, with minor modifications based on the local ethical and scientific committees. More importantly, development of streamlined processes will be crucial to ensure timely trial activation.
This study is limited by its retrospective nature. Although we selected two academic institutions, the process for clinical trial activation may vary within large academic cancer centers and small- to medium-sized community-based practices, although this difference was reported to be insignificant in a previous study.4 Additional research to explore the impact of administrative barriers on low accrual is warranted, and more importantly, clinical trials operation processes need to be improved and streamlined.
AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST
Although all authors completed the disclosure declaration, the following author(s) indicated a financial or other interest that is relevant to the subject matter under consideration in this article. Certain relationships marked with a “U” are those for which no compensation was received; those relationships marked with a “C” were compensated. For a detailed description of the disclosure categories, or for more information about ASCO's conflict of interest policy, please refer to the Author Disclosure Declaration and the Disclosures of Potential Conflicts of Interest section in Information for Contributors.
Employment or Leadership Position: None Consultant or Advisory Role: None Stock Ownership: None Honoraria: Giorgio V. Scagliotti, Eli Lilly, Roche, AstraZeneca Research Funding: None Expert Testimony: None Other Remuneration: None
AUTHOR CONTRIBUTIONS
Conception and design: Ramaswamy Govindan
Financial support: Ramaswamy Govindan
Administrative support: Ramaswamy Govindan
Collection and assembly of data: Andrea Wang-Gillam, Kristina Williams, Silvia Novello, Ramaswamy Govindan
Data analysis and interpretation: Andrea Wang-Gillam, Kristina Williams, Silvia Novello, Feng Gao, Giorgio V. Scagliotti, Ramaswamy Govindan
Manuscript writing: Andrea Wang-Gillam, Kristina Williams, Silvia Novello, Giorgio V. Scagliotti, Ramaswamy Govindan
Final approval of manuscript: Andrea Wang-Gillam, Kristina Williams, Silvia Novello, Feng Gao, Giorgio V. Scagliotti, Ramaswamy Govindan
Appendix
Timeline for Activating a Thoracic Oncology Clinical Trial
Timeline for a Phase II, Industry-Sponsored, Thoracic Oncology Clinical Trial
Footnotes
-
See accompanying editorial on page 3799
-
Authors' disclosures of potential conflicts of interest and author contributions are found at the end of this article.
- Received January 12, 2010.
- Accepted May 21, 2010.
