Adapting clinical trials networks to promote cancer prevention and control research†
Reviewed and approved by the Institutional Review Board, School of Public Health, University of North Carolina at Chapel Hill (IRB 02-1812).
Since 1987, cancer centers and clinical cooperative groups serving as ‘research bases’ for the National Cancer Institute's (NCI) Community Clinical Oncology Program (CCOP) have been required to design and conduct cancer prevention and control (CP/C) clinical trials in addition to therapeutic trials. The study describes the structural and strategic adaptations that CCOP research bases have made to incorporate CP/C research into their scientific agendas and operations.
Approaches to CP/C research in four cooperative groups with different scientific agendas were investigated: the Eastern Cooperative Oncology Group, the National Surgical Adjuvant Breast and Bowel Project, the North Central Cancer Treatment Group, and the Southwest Oncology Group. Sixty-five individual interviews with group leaders and investigators were conducted. Data were coded and thematically analyzed from transcribed interviews, observations of scientific sessions, and secondary documents.
The cooperative groups have tailored CP/C research programs to fit the interests, skills, and practice settings of their investigators without making significant adaptations in administrative and statistical infrastructures. Ongoing challenges include finding ways to broaden the pool of investigators designing CP/C protocols, involve more nononcologists in protocol design, and secure more stable and adequate funding for CP/C research.
The cooperative groups' experiences with CP/C research suggest that, with adequate resource investments, existing clinical research networks can expand into new areas of scientific investigation. Cancer 2006. © 2005 American Cancer Society.
Aided by sophisticated new technologies, scientists are generating vital information on cancer processes at the genetic, molecular, and cellular levels that can guide the development of interventions to prevent, suppress, or reverse carcinogenic progression; detect cancers at the earliest stages; and improve quality of life for those living with the disease.1, 2 New insights into the molecular pathways driving the carcinogenesis process are blurring the distinction between cancer treatment and cancer prevention.3, 4 Targeted therapies for advanced cancer are being tested in adjuvant settings and then modified for use as preventive agents. Tools from molecular genetics and molecular oncology hold similar promise for alleviating cancer-related and treatment-related symptoms.
As new agents and interventions for preventing, detecting, and managing cancer become available, the infrastructure for evaluating their effectiveness in Phase III clinical trials will require significant expansion. For example, compared with most Phase III trials the testing of molecularly targeted chemoprevention agents will require access to larger population bases to identify and enroll healthy, but high-risk individuals with specific genetic/environmental susceptibilities.5 Greater collaboration between oncologists and nononcologists also will be needed to design and conduct prevention trials targeting cancer-free populations and those at risk for second primary cancers. For the oncology community, two options exist: 1) building new prevention-focused clinical research networks, or 2) adapting historically therapy-oriented research networks to conduct cancer prevention and control (CP/C) research.
The question of whether to build or adapt clinical research networks has important policy and resource implications for the nation's biomedical research efforts. For example, the National Institutes of Health's (NIH) “Roadmap for Re-engineering the Clinical Research Enterprise” recommends the creation of integrated networks of academic centers and community-based physicians to accelerate the translation of biomedical research from bench to bedside and beyond.6 Recognizing the time and cost involved in building new research networks, the NIH is investigating the feasibility of broadening the research scopes and memberships of existing clinical research networks to address new types of scientific questions.
This article explores the potential for using existing clinical research networks to advance CP/C research. Building on earlier research,7 our study describes the structural and strategic adaptations that clinical cooperative groups participating in the National Cancer Institute's (NCI) Community Clinical Oncology Program (CCOP) have made to incorporate CP/C research into their scientific agendas and organizational structures. Four questions guided our study. 1) How have cooperative groups defined the scope and priority of CP/C research? 2) How have they built internal capacity to design and conduct CP/C research? 3) How have they cultivated investigator interest in designing and conducting CP/C clinical trials? 4) How have they budgeted for CP/C research and obtained external financial support? We conclude with a discussion of three necessary conditions for successfully integrating CP/C research into cancer clinical trials networks.
The CCOP involves a three-way partnership between the NCI's Division of Cancer Prevention (NCI/DCP), 14 participating clinical cooperative groups and cancer centers (CCOP research bases), and 63 local cancer care provider networks (CCOPs). In this partnership, NCI/DCP provides overall direction and funding; CCOP research bases design and implement clinical trials; and CCOPs assist with patient accruals, data collection, and dissemination of study findings. Although this program initially focused on the evaluation of cancer therapies, NCI/DCP expanded its scope in 1987 by requiring CCOP research bases to design and conduct CP/C clinical trials and requiring CCOPs to meet annual cancer control accrual targets. Since then, the CCOP network has demonstrated its ability to design and conduct CP/C trials, including chemoprevention trials involving tens of thousands of study participants. Several CP/C trials have garnered national attention and changed the standard of care for at-risk populations and cancer patients.8 Despite these overall achievements, CCOP research bases vary in CP/C research productivity. Examining how four of these research bases have designed and implemented CP/C research programs provides important insights about the organizational changes and resource investments needed for clinical research networks to expand into areas of investigation involving new disciplines, research methods, and patient populations.
MATERIALS AND METHODS
Between October 2002 and August 2003, we conducted case studies of four CCOP research bases: the Eastern Cooperative Oncology Group (ECOG), the National Surgical Adjuvant Breast and Bowel Project (NSABP), the North Central Cancer Treatment Group (NCCTG), and the Southwest Oncology Group (SWOG). Although 14 cooperative groups and cancer centers currently serve as CCOP research bases, the four groups studied are among the few that have been continuously funded since the program's inception. The selection of this purposive sample9 allowed us to investigate similarities and differences in approaches to implementing CP/C research among groups with distinctive organizational structures, different scientific agendas, and various levels of CP/C research activity.
By scheduling our site visits to coincide with semiannual cooperative group meetings, we conducted 65 1-hour individual interviews with group leaders, chairs and selected members of CP/C related committees, CCOP physicians and staff, and investigators from non-CCOP institutions (Table 1). We used semistructured discussion guides, tailored to the interests and expertise of individual participants, to elicit information on the processes, facilitators, barriers, and outcomes of integrating CP/C research. With permission, we audiotaped and transcribed the interviews verbatim. We also administered a brief questionnaire to each group's chief operations officer, observed CP/C scientific sessions and committee meetings, and reviewed secondary documents such as grant applications and annual progress reports.
Table 1. Distribution of Interview Participants by Cooperative Group
|Group administrators and staff||5||4||6||3|
|CP/C committee chairs||3||2||2||4|
|CCOP physicians and staff||5||7||6||7|
|Investigators from non-CCOP institutions||2||4||4||1|
|Total interview participants||15||17||18||15|
We used a qualitative data analysis software program (Atlas.ti 4.2, Scientific Software Development, Berlin, Germany) to manage study data. Interview transcripts alone totaled 827 single-spaced pages. We coded the data topically by study question and then thematically analyzed the data, seeking at least two mentions of a theme by interview participants who, by virtue of their position or experience, had knowledge about the theme.10 We used questionnaire data and secondary documents for description and corroboration. We then produced detailed case reports of each group, which we shared with group leaders and NCI/DCP officials to correct factual errors, answer remaining questions, and confirm interpretations.
Scope and Priority of CP/C Research
NCI/DCP defines cancer prevention research as studies that evaluate new methods of detecting cancer risk and preventing primary and secondary cancers, and cancer control research as studies that evaluate symptom management, rehabilitation, and continuing care interventions designed to minimize the burden of cancer and improve quality of life. Through a combination of natural evolution and strategic intent, CCOP research bases are differentiating into “centers of excellence” for different types of CP/C research. Each group has tailored its CP/C research program to fit the interests, skills, and practice settings of group members. At NCCTG, for example, symptom management studies have become a mainstay, reflecting the resources and interests of its community investigators. NSABP moved into chemoprevention trials as a “natural extension” of the group's research on the effectiveness of adjuvant chemotherapies in reducing breast cancer recurrence among patients with early-stage disease. ECOG built a strong quality-of-life research program around its therapeutic trials. In the mid-1990s, SWOG departed from this “grow from within” approach by reaching outside the group to change its CP/C committee leadership and reorient its CP/C research program. Since then, SWOG has focused on prevention research, particularly chemoprevention and molecular epidemiology.
The four cooperative groups differ not only in the scope of their CP/C research programs but also in the extent to which CP/C research represents an organizational priority. Because cooperative groups began as therapy-oriented research networks, CP/C research exhibits only a modest fit with most investigators' interests and priorities. However, we observed that group leadership plays an important role in determining the visibility and priority of the CP/C research program. Some group leaders have committed significant resources to CP/C research—at least particular variants of it—while others have moved more cautiously into this arena.
Design and implementation challenges
As cooperative groups have built CP/C research programs, they have had to adapt to the distinctive design and implementation challenges of different types of CP/C research. Chemoprevention protocols, for example, require oncology practices to identify and recruit healthy participants whose ages, medical histories, or other factors place them at increased risk for cancer. To conduct such trials, groups have had to conduct training programs on participant recruitment and adherence strategies and develop extensive promotional materials. While symptom management protocols involve patients routinely seen in oncology practices, they require considerable nursing staff time to screen patient charts for eligibility and to discuss the protocols with patients. To the extent possible, groups have tried to design interventions that can be easily incorporated into the normal course of care. Quality-of-life studies require constant monitoring to ensure that data are collected at the correct intervals. To successfully complete these studies, some groups have implemented compliance initiatives involving monitoring, feedback, incentives, and penalties.
Interview participants described lifestyle and behavior modification clinical trials as particularly challenging because these studies often involve complex batteries of standardized tests or psychosocial interventions that require specialized staff training.11 To address important research questions in a less labor-intensive manner, some groups have adopted a minimalist approach to behavioral intervention or psychosocial assessment. For instance, NCCTG successfully conducted several large smoking cessation trials by evaluating simple interventions, such as nicotine patches, rather than intensive behavioral interventions, such as in-person counseling.
CP/C research productivity
Table 2 provides measures of CP/C research activity for the four cooperative groups. Heterogeneity in the types of CP/C research conducted by the groups, with each type possessing distinctive challenges, precludes a comparative evaluation of CP/C research productivity based on any single metric. For example, between 1997 and 2002, NCCTG activated 27 clinical trials, most of which tested symptom management agents on a few hundred participants. During this same time period, NSABP enrolled more than 15,000 participants on a single breast cancer chemoprevention trial. SWOG accrued more than 18,000 study participants to a large prostate cancer chemoprevention trial but also activated six smaller studies. ECOG activated five CP/C clinical trials and accrued fewer study participants than the other groups. Publication records also vary. For example, between 2000 and 2002, SWOG published 54 abstracts and journal articles on CP/C and quality-of-life studies, as compared with 37 NSABP publications, 31 NCCTG publications, and 10 ECOG publications.
Table 2. Measures of CP/C Research Activitya
|Number of new CP/C concepts submitted|| || || || |
|Number of new CP/C protocols submitted|| || || || |
|Number of CP/C protocols activated|| || || || |
|Accrual to CP/C protocolsb|| || || || |
Building Internal Capacity to Conduct CP/C Research
To accommodate CP/C research, all four groups have altered their organizational structures by adding one or more scientific committees alongside therapy-oriented disease and modality committees. In ECOG and SWOG, CP/C committees have the authority and resources to develop protocols. In NCCTG, the Cancer Control Committee reviews and approves CP/C protocols developed by Mayo Clinic investigators. In NSABP, CP/C committees develop some protocols, while ad hoc committees appointed by the Operations Center develop others. Disease committees occasionally design CP/C protocols but more often play consultative roles in protocol development.
The formation of CP/C research committees offers several advantages. Bringing investigators with similar interests and expertise together encourages concentrated effort and promotes high-quality problem solving.12 Investing these committees with authority and resources also signals an organizational commitment to CP/C research. However, interviewees noted that organizing a CP/C research program this way introduces communication and coordination barriers vis-a-vis the disease and modality committees and subtly reinforces the perception that CP/C research is “something different” from what the rest of the group is doing.
Adaptations in group staffing and operations
All four groups have allocated administrative and statistical staff time to support CP/C research. Staffing levels vary by the scale of the group's CP/C research activity. Although the groups have not made significant changes in protocol development policies and procedures, their statistical center staffing and operations have required greater adaptation. Statisticians accustomed to designing and analyzing therapeutic trials do not necessarily possess the expertise to design and analyze symptom management studies, quality-of-life studies, and behavioral studies with nontraditional endpoints.11 To meet this challenge, the groups have hired new statisticians, cross-trained existing staff, and modified information systems to include new data elements.
Promoting Investigator Interest and Involvement
Finding ways to encourage investigators to design CP/C protocols presents an ongoing challenge for cooperative groups. Regardless of whether CP/C protocol development occurs centrally at the research base or through committees of investigators from member institutions, cooperative groups must 1) recruit investigators with the requisite disciplinary expertise to ask and answer CP/C research questions, 2) equip them with the skills to design and implement CP/C trials, and 3) provide incentives and recognition for CP/C protocol development.
None of the cooperative groups studied have developed a formal, systematic process for identifying and recruiting new CP/C investigators. Recruitment typically occurs informally and opportunistically. For example, a senior investigator may invite a junior colleague at his or her institution to attend a CP/C committee meeting or to present a research concept to the committee. While the groups have managed to assemble multidisciplinary cadres of CP/C investigators, interview participants cited a growing need to involve more nononcologists, such as gastroenterologists, pulmonologists, and urologists, in designing prevention protocols targeted toward cancer-free populations and those at risk for second primary cancers. They reported limited success in attracting such physicians due partly to a lack of focused attention on the issue and partly to their inability to provide salary and travel support.
Despite a recognized need to develop the CP/C research capabilities of junior investigators, the four groups have not implemented formal mentoring systems or training programs specific to CP/C research. NCCTG and NSABP rely on informal mentoring to acquaint new investigators with protocol development and review procedures. In ECOG and SWOG, all investigators planning to develop and coordinate protocols are required to attend study chair workshops. SWOG also offers an intensive 2-week training program to prepare early-career clinical scientists to design protocols, navigate protocols through the cooperative group and NCI review system, and manage multisite trial implementation. Although this training program does not address CP/C clinical trials, SWOG leaders are considering adding a CP/C research component in the future.
The cooperative groups have tried multiple mechanisms to educate investigators about CP/C research and motivate them to design and lead CP/C studies. These mechanisms include presentations at plenary sessions, educational symposia, trial-specific educational sessions and training workshops, newsletter and website articles, and “cross talk” among committee members. Yet even in groups with productive CP/C research programs, the number of investigators designing and coordinating CP/C protocols remains relatively small. Interview participants noted that researchers at academic medical centers have little incentive to design or participate in CP/C trials because they are primarily evaluated based on their therapeutic research. Although cooperative groups can offer public recognition and publication opportunities to study chairs, CP/C studies developed through cooperative groups typically take longer to design, activate, and complete than do CP/C studies pursued through a single institution. This situation represents a formidable barrier for junior investigators interested in CP/C research.
To date, relatively few CCOP investigators have designed CP/C protocols or served as study coordinators. Although all CP/C committees include community oncologists and nurses, CCOP interviewees in three cooperative groups said they would like more opportunities to comment on eligibility criteria, laboratory tests, time requirements, and other feasibility issues during concept and protocol development. They suggested that formal feasibility reviews by small teams of community oncologists and nurses during the design stages would maximize each study's accrual potential and reduce the need for protocol amendments after activation.
Funding for CP/C Research
CCOP research bases receive funding from NCI/DCP in the form of a cooperative agreement that is intended to cover the administrative and statistical center costs associated with CCOP participation in cancer treatment and CP/C trials. Between fiscal year (FY) 1990 and FY 2003, the number of research bases remained fairly constant (mean = 12), but CCOP research base funding increased fourfold, exclusive of funding for the Breast Cancer Prevention Trial13 and the Prostate Cancer Prevention Trial.14 Although NCI/DCP encourages CCOP research bases to allocate some of their funding to support the development and pilot testing of CP/C protocols, only one of the groups in our study reported using CCOP funds for this purpose. Investigators from this group said the earmarked CCOP funds have helped them build a “more robust” CP/C research program. However, like interviewees in other groups, they viewed CCOP research base grants as inadequately covering the staff and information systems support required for CP/C research.
As an additional funding issue, interview participants cited the need for a more stable funding mechanism to cover small cancer control research studies, developmental research for Phase III prevention trials, and the various component costs of conducting large prevention trials. Through a collaborative arrangement with the NCI's Office of Cancer Complementary and Alternative Medicine, NCI/DCP is able to fund some symptom management studies that evaluate complementary and alternative therapies. NCI/DCP tries to support innovative proposals for other small CP/C studies with end-of-year funds, but the availability of these funds is unpredictable. Through a competitive peer-review process, NCI/DCP makes supplemental awards to allow cooperative groups to conduct large Phase III prevention trials. These supplemental awards cover basic study costs, such as data management, drug distribution, and statistical analysis. However, the supplemental awards do not fund laboratory correlative studies or participant recruitment and adherence activities, and they seldom fund tests and procedures that exceed the standard of care covered by health insurance.
To conduct CP/C trials, interviewees said they try to piece together funds from several sources. They noted that each funding source is problematic in some respect. While R01 funding is a possibility, junior investigators often lack experience writing R01 proposals, and heavy clinical workloads make it difficult for them to persevere through the two or three rounds of revision and resubmission that R01 funding requires. Pharmaceutical companies are another important source of support, especially for chemoprevention trials. However, partnering with industry has its own challenges. Contracting for investigational drugs and drug distribution has become increasingly complicated because of ongoing industry consolidation. Even when mergers are not an issue, involving industry partners adds another layer of complexity to the protocol design process.
Beyond the time and effort involved, having to piece together funding from several sources introduces considerable uncertainty into the protocol development process. One investigator offered the following example: You might get 25% from the drug companies, 25% from NCI, and then you are forced to write a grant for the other 50%. You could imagine that if the money's coming from three sources and, to successfully do the study you need the money from all three sources, any one of those that falls through totally derails the whole plan.
Another investigator commented, “We spend years establishing the funding mechanism and that prolongs the development time…. When it takes that long to develop a protocol, the science changes.” All agreed that a more stable funding mechanism would make it easier for cooperative groups to conduct ‘cutting-edge’ CP/C research, particularly in the fast-moving realm of molecular prevention science.
Since 1989, more than 83,000 people have participated in cancer prevention clinical trials through the CCOP research network, and many more have participated in screening, symptom management, and quality-of-life research studies. Although it has been challenging for historically therapy-oriented cooperative groups to design and conduct trials involving new disciplines, research methods, and patient populations, CCOP research bases have shown, with various degrees of success, that they can leverage their substantial operational and statistical infrastructures to meet the demands of CP/C research. Our study findings suggest several actions that CCOP research bases, working collaboratively with NCI/DCP, could take to enhance their CP/C research capabilities (Table 3). These suggested actions complement the recommendations that the Clinical Trials Working Group of the National Cancer Advisory Board recently issued for restructuring the national cancer clinical trials enterprise.15
Table 3. Recommendations for Enhancing the Cancer Prevention and Control CP/C Research Capacity of Cooperative Groups
|• Focus the CP/C research program to fit with members' interests, skills, and practice settings.||Establish a network of Scientific Steering Committees to address design and prioritization of phase III trials that leverages current Intergroup, Cooperative Group, SPORE, and Cancer Center structures and involves the broad oncology community.|
|• Concentrate resources in areas that members see as important; build on knowledge gained from initial CP/C clinical trials.|
|• Streamline internal processes for concept and protocol development, priority setting, and review.||Identify the institutional barriers that prolong time from concept approval to accrual of the first patient, and develop solutions for overcoming these barriers.|
|• Increase internal coordination by assigning formal liaisons that link CP/C and disease committees.|| |
|• Develop and implement plans for CP/C investigator recruitment that promote multidisciplinary participation in CP/C research.||Realign NCI funding, academic recognition, and other incentives to promote collaborative team science and clinical trial participation.|
|• Establish a CP/C research training and mentoring program for young investigators; encourage applications for career awards.|| |
|• Actively solicit CCOP input on feasibility issues during the protocol design stages. Provide feedback on actions taken.||Enhance community oncologist involvement in clinical trial design and prioritization through representation on Steering Committees and creation of community oncologist focus groups.|
|• Disseminate results of completed CP/C studies to all accruing investigators—not just meeting attendees.|
|• Establish cancer control accrual requirements for all member and affiliate institutions.||Restructure the funding model for phase III efficacy trials to incentivize more rapid rates of patient accrual. As part of the implementation plan, conduct a comprehensive financial analysis of clinical trial costs.|
|• Conduct formal cost analysis of CP/C research program to identify areas for cost savings and to document areas needing additional support.|
Beyond these specific actions, our study findings suggest three necessary conditions for successfully integrating CP/C research into cancer clinical trials networks. First, to play a decisive role in advancing scientific knowledge of cancer prevention, screening technologies, and symptom management, cooperative group investigators will need to take ownership of CP/C research. Group investigators are keenly interested in cancer biology and the molecular pathways associated with cancer development and progression. However, more collaboration with basic scientists may be needed to achieve a full portfolio of promising preventive agents. Compared with therapeutic research, prevention research will require more detailed plans for participant recruitment and adherence and longer periods of intervention and follow-up to measure outcomes of interest.5 In the absence of widely accepted standards of symptom management, research protocols will need to be carefully designed to build a documented evidence base.
The evolution of cardiology research shows how research and treatment paradigms can be expanded to include prevention as a key component of science and practice.16 Over time, this field has moved from exclusively treating major cardiovascular disease events to preventing first or second major events through the identification and treatment of risk factors. As noted by one clinical scientist: Cardiologists are very good at both primary and secondary prevention with lipid-lowering agents and aspirin and beta-blockers and other agents. That hasn't stopped them from treating patients once they get their heart attacks. I think that's the kind of paradigm we need to get oncologists to start thinking about, but they are not. They're thinking that prevention is not their bailiwick.
More funding for CP/C research would undoubtedly generate broader interest, produce greater effort, and attract additional investigators. However, to effectively use this funding, each cooperative group needs to decide what it wants to accomplish through CP/C research and how each scientific committee can contribute to these ends. This collective vision can serve as a basis for developing appropriate opportunities, incentives, and resources for CP/C research within the cooperative group.
Second, the successful integration of CP/C research into cancer clinical trials networks requires plans and programs for grooming current and future generations of CP/C investigators. As noted by Sung et al.,17 the number of physician investigators participating in clinical research is declining. Reasons include growing educational debt loads, the prolonged training required for clinical research, difficulties securing research funding, uncertain academic career prospects for physicians with clinical research interests, and increasingly cumbersome regulatory and review processes that have slowed clinical trials implementation.18 Cooperative groups have the scientific expertise and training technologies needed to develop model programs for improving the recruitment, productivity, and retention of junior investigators. Because cooperative groups can request dedicated funding in their CCOP research base budgets to support the development of CP/C protocols in member institutions, one option would be to target these monies toward junior investigators interested in designing small developmental studies.
Finally, the successful integration of CP/C research into cancer clinical trials networks will require an increased investment of financial resources. To take advantage of advances in molecular medicine, cooperative groups have begun recruiting new investigators, reorganizing scientific committee structures, enhancing laboratory capabilities, and developing novel interorganizational relationships. Capitalizing on these efforts will require increased funding for pilot CP/C studies and small targeted intervention clinical trials, CP/C-related laboratory correlative studies, and innovative efforts to expand nononcologists' participation in protocol design and implementation. Over time, the testing of molecularly targeted agents for cancer prevention may require further expansion of existing research networks to obtain sufficient numbers of study participants who are at high risk because of genetic characteristics or other clinical and demographic factors.
The cooperative groups' experiences with CP/C research suggest that, with adequate resource investments, existing clinical research networks can expand into new areas of scientific investigation. Through partnerships with the pharmaceutical industry and the biotechnology sector, NCI/DCP may be able to generate the resources needed to expand the capacity of these networks to test new preventive strategies and to accelerate the transfer of proven interventions into clinical practice.
The authors thank the 65 cooperative group and CCOP representatives who participated in interviews for their time and valuable insights. We also thank Cynthia Whitman, NCI/DCP (National Cancer Institute, Division of Cancer Prevention), for providing the data needed to assess each cooperative group's CP/C research activity.