Multi-Center Phase III Clinical Trials and NCI Cooperative Groups: Workshop Summary (2009)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Session 4: Costs of Cooperative Group Clinical Trials The last session of the workshop addressed the costs linked to coop- erative group clinical trials, how to document those costs, and how they are covered. During this session, speakers expanded on some of the cost issues suggested by earlier speakers, such as a lack of insurance reim- bursement for routine patient care costs in clinical trials, the increased costs of running registration trials or complex trials that include pharma- cogenomic tests, and the decline in government funding for cooperative group trials. Cooperative groups cost analysis The first speaker in this session, Dr. Alan Benson, professor of medi- cine at the Northwestern University Feinberg School of Medicine and associate director for clinical investigations at the Robert H. Lurie Com- prehensive Cancer Center, provided a cost analysis of cooperative groups. He reiterated that there has been a decline in NCI funding for the coopera- tive groups, and he noted that because of NCI funding limitations, coop- erative groups usually receive 30 to 50 percent less than the total grant money asked for on their applications. Efforts to increase NCI support include a new “trial complexity” payment mechanism that CTEP is in the final stages of developing. But it is expected that the additional support for complex trials will be limited in scope and will not sufficiently relieve the cost burdens of these trials. In the future, NCI support of cooperative group trials may decline 75 

76 MULTI-CENTER PHASE III CLINICAL TRIALS further, Dr. Benson said, given that preliminary data between 2005 and 2008 show that the number of new Phase III trial concepts submitted to the NCI by cooperative groups decreased by 75 percent and that the num- ber of concepts approved for development of protocols decreased by more than 90 percent.28 Until this trend is substantially reversed, the declining numbers of trials and accruals will mean that sites will receive lower total capitation payments than was previously the case. It may, however, be possible for cooperative groups and U10-funded sites to reallocate capita- tion funds to other needs, Dr. Benson added. Furthermore, he added, as trials become more complex and require more pharmacogenomic tests, there will be more fiscal pressures that are not likely to be relieved by the NCI’s trial complexity payment mechanism. Industry provides supplemental funding for cooperative group tri- als that can support the cooperative group’s centralized activities, such as the effort involved in meeting the data reporting requirements for registration trials, or for data management and quality assurance efforts. Much of such industry support is for correlative studies. Industry also provides financial support to individual sites, such as reimbursement for patient tests and procedures beyond the standard of care, or for additional required data submissions or submissions needed for an independent post-study review of imaging findings. But the level of patient reimburse- ments is often insufficient, Dr. Benson said, “and there is a continuing battle between the pharmaceutical industry and the insurance providers as to who pays for the standard-of-care costs.” He explained that indus- try support is garnered by a drug manufacturer negotiating a contract with and providing payments to the lead cooperative group, which then disburses funds directly to its sites. Industry recognizes that cooperative group trials offer a significant bargain, Dr. Benson said, and it is often willing to collaborate with the cooperative groups for that reason. Dr. Benson pointed out, as other speakers had, that lack of reimburse- ment for the standard-of-care costs of patients in clinical trials leads to another major cost that cooperative groups have to bear. He noted, how- ever, that recently Medicare has been working with the NCI and the FDA to broaden support for clinical trials, which has led to the announcement of high-priority trials that should be reimbursed (NCI, 2008a). It is too soon to determine the effects of this announcement, he said, “but certainly such interaction among our federal agencies should be encouraged to 28  The submissions decreased from 70 in calendar year 2005 to 17 through May 1 in 2008; the approvals declined from 48 in calendar year 2005 to 4 through May 1 in 2008. However, the numbers from 2008 are not final, so these decreases may not be as dramatic. Niederhuber, J. 2008. Presentation to the Cooperative Group Chairs. Chicago, IL, and personal communication, S. John, National Cancer Institute, October 17, 2008. 

COSTS OF COOPERATIVE GROUP CLINICAL TRIALS 77 offer levels of support.” He added that payment reluctance on the part of insurers is reinforced by the high costs of newly developed drugs and biologics. Dr. Benson noted that there is a growing need for correlative and translational studies that use biospecimen banks. The NCI provides some financial support for specimen submissions, but waiting for such supple- mental support from the NCI or from industry can substantially delay activation of trials. “Many academic labs will no longer subsidize some of these efforts,” he said. “We have constrained resources, both for our concurrent studies and our retroactive studies for bank specimens, and it is often hard to convince pharmaceutical companies of the importance of these studies.” Dr. Benson summed up his talk by saying, “The assault on clinical research is obvious. Clinical research is clearly undervalued in terms of the time, effort, cost, and importance by all involved, including academ- ics, government, media, and [health care payors]. We feel that the public- sector trials are at grave risk.” In the discussion following Dr. Benson’s presentation, Dr. Michaele Christian, former director of DCTD at the NCI, suggested that many of the financial pressures Dr. Benson summarized could be alleviated by having fewer cooperative group clinical trials that are done much faster. “That’s actually something within our control,” she said. “We need to further strengthen internal review and prioritization so that we can choose the most important studies and pay more for them.” Dr. Johnson said that greater involvement of patients receiving care from the government-funded Department of Veterans Affairs (VA) patients should be encouraged when conducting cooperative group trials. Dr. Benson agreed but noted obstacles to this approach, including a VA IRB not accepting a review of a clinical study produced by a central IRB or a local institution’s IRB. “We have seen VA accrual steadily drop because it has become harder and harder to conduct research in the VA system,” Dr. Benson said. “It’s another example of governmental agencies working at cross purposes—no one is cooperating, and an important population of patients at the VA are being denied access to the latest research efforts,” he said. Clinical trial cost Management The next speaker, Ms. Marcy Waldinger, chief administrative officer at the University of Michigan Comprehensive Cancer Center, showed how the costs of a clinical trial can be determined and used to negotiate better 

78 MULTI-CENTER PHASE III CLINICAL TRIALS financial support from trial sponsors. She noted that many of the costs of clinical trials are overlooked or understated, such as pre- and post-award costs and the costs of specimen collection, processing, and shipping. As trials have become more complex, she noted, there can be more expenses linked to shipping specimens, especially if they have to be shipped indi- vidually and are time-sensitive. Also often overlooked or understated are the costs of pathology evaluations, multi-center coordination, and long- term follow-up. The effects of inflation should also be taken into account when performing a cost assessment. The standard approach to determining clinical costs is to base them on the number of patients accrued—for example, to assume that one full-time employee for data management is needed for every 30 patients. “This is truly anachronistic in today’s age, when trials have become so complicated,” said Ms. Waldinger. “We need a new approach with mul- tiple variables that account for the study complexity,” she said, suggest- ing that the NCI should facilitate the development of such a complexity model. In addition to the issue of whether the trial is Phase I, II, or III, variables that should be considered include the amount of sampling, the age and degree of morbidity of the patients, how easily the patients can be accrued, and whether they will be treated as inpatients or outpatients. To help assess clinical trial costs more accurately, the University of Michigan developed an “effort tracker” that provides real-time reporting of research support staff costs. Data managers, research nurses, regulatory staff, and staff involved in specimen processing spend about five minutes each day logging how they are spending their time for various clinical trials. For example, research nurses note how much time they spend on consenting patients, on patient education, and on screening. “This has been very helpful to us to validate our budgets and to actu- ally document any increased effort, as appropriate,” Ms. Waldinger said. For example, the increased effort involved in acquainting new industry monitoring personnel with a clinical trial incurs a cost that is documented with the effort tracker system, and this documentation is used to petition the sponsor to increase payment. The effort tracker also provides a metric to justify staffing and workload assignments. “If staff is spending much more time on a study than had been budgeted, they can document how they spent that time, and then we can evaluate if maybe there are non- value-added steps, or if, in fact, that time was legitimately spent,” Ms. Waldinger said. In the discussion following the presentation, Ms. Linda Beekman, administrative director of clinical research at the University of Michigan Comprehensive Cancer Center, noted that an initial analysis of 67 trials whose budgets were determined using an effort tracker found that only 5 of them went over budget. But those studies had just recently opened 

COSTS OF COOPERATIVE GROUP CLINICAL TRIALS 79 and had incurred significant start-up costs, so the over-budget costs were potentially overstated. If those nascent studies were taken out of the analysis, the budgets for the remaining studies were more than 95 percent accurate, she said. In response to a question raised after her presentation, Ms. Waldinger noted that there was initial resistance from staff to having to continu- ally document their time with the effort tracker, and some staff found the activity demeaning. But, she said, “They are our employees, so they really don’t have a choice in the matter. Our vantage point is that lawyers, accountants, and very high-priced consultants do this,” she said. In fact, many employees now appreciate the effort tracker, because it documents efforts on clinical trials that were previously underestimated, thereby vali- dating employee complaints that more time is needed to do the work. She suggested that administrators implement effort trackers or simi- lar data systems that can assess budgeted versus actual costs and provide real cost data that can aid negotiations with sponsors. Those negotia- tions should be done by experienced professionals who are assertive and knowledgeable about all the costs linked to a clinical trial. Ms. Waldinger noted that most physicians are not trained in budgeting and budget negotiations, and many times their interest in conducting a trial trumps their attention to related costs, especially since most of those costs are not obvious. The University of Michigan Comprehensive Cancer Center saw an increase in its clinical trials budgets when it changed personnel three years ago to hire staff who were experienced in clinical trials finance and negotiation. Another helpful tool that is increasingly being used to assess clini- cal trial costs, according to Ms. Waldinger, is “value-stream mapping,” which is similar to the process map analysis Dr. Dilts presented earlier. In value-stream mapping all the steps involved in a clinical trial are mapped, including the steps of regulatory agencies, and the “value-added” and “non-value-added” steps are identified and quantified. Another useful graphic is one that depicts a “circle of influence.” Used as a tool to reach consensus and prioritize where to devote energy and resources, the circle of influence graphic outlines what is within the control or influence of participating parties in a clinical trial process (Figure 3). 29 In the remainder of her talk Ms. Waldinger discussed tools and strate- gies for improving the efficiency of clinical trials as well as what she called “cost-out” solutions. One tool that has greatly improved the efficiency of University of Michigan clinical trials is a homegrown electronic medical record search engine with the acronym EMERSE. This search engine is the equivalent of a Google search engine for electronic medical records, 29  The “circles of influence” concept was adapted from Covey (1989). 

80 MULTI-CENTER PHASE III CLINICAL TRIALS Out of our Control/Influence (40%) PI protocol submission Co-PIs accepting eResearch role PRC follow-up Within our Influence (30%) Program director Ancillary committee accepting roles approvals Company sending the CDA PRC letters to PI Routing to PI review and different areas approval in eResearch (medical school, MTA work with Within our Control (30%) DRDA) Hematology/Oncology CTO Internal communication IRB corrections and Workload and data management protocol revisions PI interviews eResearch completion (3 sections) Visible information flow and tracking Budget negotiation Coordination of information flow Financial process Protocol dissection CTRAC application (paper and electronic) Portions of CTRAC process completion Determining/Collecting routing documents Billing calendar development PIs sign CDA and Site questionnaire send to company Budget revisions IT effort estimates Collecting routing documents for PAF IRB review and Data manager associated processes communications with Waiting for sponsor sponsor to respond to budget Lack of IT systems synchronization Protocol presentations to sections PI interviews Sponsor drug delays PI and sponsor protocol finalization FIGURE 3  The circles of influence. The graphic demonstrates that only 30 percent of clinical trial activities are within the control of the Clinical Trials Office, while another 30 percent of activities are within the influence. Forty percent of clinical trial activities are outside the Clinical Trials Office’s control or influence. App = application, CDA = confidential disclosure agreement, Co-Is = Co-investigators, Figure 3new broadside CTO = Clinical Trials Office, CTRAC = Clinical/Translational Resource Alloca- tion Committee, DRDA = Division of Research Development and Administra- tion, Hem/Onc = Hemotology/Oncology, IRB = institutional review board, IT = information technology, Mgr = manager, MTA = materials transfer agreements, PAF = proposal approval form, PI = principal investigator, PRC = protocol review committee, syncing = synchronization. SOURCE: Waldinger presentation (July 2, 2008) and adapted from Montague, S. 2000. Circles of influence: an approach to structured, succinct strategy. http://pmn. net/library/Circles_of_Influence_An_Approach.htm (accessed January 6, 2009). Reprinted, with permission, from Circles of influence: An approach to structured, suc- cinct strategy. Copyright 2000 by Performance Management Network, Inc. Ms. Waldinger said, and it can help staff to quickly determine patient eligibility for clinical trials and manage the data of those trials. “This tool has been really helpful for our data managers to rapidly access salient information,” she said. “Instead of having to pore through the electronic medical record, they can just type the word ‘stage’ or ‘disease progression’ 

COSTS OF COOPERATIVE GROUP CLINICAL TRIALS 81 or ‘tamoxifen,’ and every case where that term exists pops up.” She added that data management staff at the University of Michigan Comprehensive Cancer Center reported in a recent EMERSE user satisfaction survey that the search engine saved them as much as two hours a day. Ms. Waldinger suggested using an effort tracker or similar system to determine the hidden costs related to screening activities and screen failures. Echoing earlier suggestions, she also stressed the importance of reducing the number of open trials with no or minimal accruals so that staff time can be devoted to more productive trials. Using effort tracker data, Ms. Waldinger created a chart of the annual costs incurred by clini- cal trials that accrue two or fewer patients (Figure 4). But, as Dr. Comis later noted, this chart did not include the $5,000 to $8,000 of start-up costs that have been documented in other studies, and that amount should be added in to her calculations when assessing the costs of low-accruing studies. In addition, Ms. Waldinger suggested using the lowest-cost provider to conduct the various tasks in a clinical trial. “We have quite a lot of circumstances in all of our cancer centers where physicians are doing work that nurses could do and nurses are doing work that data managers should be doing,” she noted. Clinical trials could also be more efficient if the forms required by industry sponsors were simplified and standard- ized, she suggested. $4,500 $4,028.72 $4,000 $3,500 Cost per year $3,000 $2,562.76 $2,500 $2,000 $1,500 $1,000 $728.28 $500 $0 Cooperative Industry Institutional–Industry Sponsor Type FIGURE 4  Average cost per active year for low-accruing studies. Annual costs incurred by clinical trials that accrue two or fewer patients, including data man- agement and regulatory expenses. Figure does not include research nursing, start- up, or long-term follow-up costs. SOURCE: Waldinger presentation (July 2, 2008). Figure 4.eps 

82 MULTI-CENTER PHASE III CLINICAL TRIALS Regulatory costs The next speaker, Dr. Maurie Markman, vice president for clinical research at the M. D. Anderson Cancer Center, addressed the impact of regulatory compliance on conducting clinical research. Much of the regulation of clinical trials is aimed at patient safety concerns, and Dr. Markman raised the issue of whether clinical trials have therapeutic intent. The supposed lack of therapeutic intent of clinical trials prevents insurance reimbursements and underlies the lengthy consent forms and IRB reviews, he said, yet most patients opt to participate in clinical trials and most physicians place them in those trials because they hope it will benefit them. “There is an extensive body of ethical literature that says that patients, through their doctors, are somehow misinformed or do not understand because there is no therapeutic intent in much of what we do, certainly in the Phase I area,” he said (Glannon, 2006; Henderson et al., 2007; Joffe et al., 2001). “I would reject that argument absolutely.” He sug- gested that this issue be explored further by the IOM. Later in discussion, a participant pointed out that FDA guidelines specify that Phase I trials merely determine the toxicity and dosing of an experimental drug and that the main difference between Phase I and II trials is the therapeutic intent of the latter. Dr. Markman also discussed the enormous costs linked to carrying out recent regulatory decisions. The OHRP recently ruled that if, during the course of a clinical trial, substantial new toxicity comes up, it is no longer sufficient for an investigator or other provider on the research team to inform new patients verbally about this toxicity. Instead, the clinical trial has to be reviewed again by the IRB at the local institution and the written consent form modified accordingly (Abrams and Mooney, 2008; Goldberg, 2008b). “This could take months,” Dr. Markman said. In addition, Medicare now requires Healthcare Common Procedure Coding System (HCPCS) modifiers to distinguish between routine care costs and those associated with an investigational clinical service when billing for patients involved in clinical trial research (CMS, 2008b). “Effective April 2008, every line item, every bill, every single test that you send on any Medicare patient has to have a modifier to say if the routine care cost is part of the trial or not part of the trial. Why are they doing this?” Dr. Markman asked. “Is this how we should really be spending our health care dollars?” In a subsequent presentation, Dr. Fitterman answered that question by saying that changes to the HCPCS modifier requirements were insti- tuted, in part, to enable CMS to assess the likely costs of its beneficiaries participating in clinical research so that in the future those costs might be incorporated into the prospective payment that is provided to the Medi- care Advantage organizations. This would prevent beneficiaries of these 

COSTS OF COOPERATIVE GROUP CLINICAL TRIALS 83 plans from having to make 20 percent copayments of the high costs that are often linked to clinical trials and, in theory, should foster more of their participation in those trials. Dr. Markman said that research-for-hire regulations complicate research contracts and funding. Internal Revenue Service (IRS) tax laws prevent academic institutions with tax-exempt bonds from doing research for hire, he explained. To ensure that a cooperative group is not a research- for-hire institution used by industry to run its clinical trials, it must retain the intellectual property rights that result from the trials it conducts. Consequently, there can be months of negotiations between cooperative groups and industry sponsors, Dr. Markman said, so that “if someone audited those contracts at the level of the IRS, they would not conclude that you were doing research for hire. If that were the conclusion, all tax- exempt bonds would be lost.” Negotiations can also be lengthy, because both industry and coop- erative groups want to retain the ownership and use of biological speci- mens collected during an industry-sponsored trial, because they might be useful for future studies. “This can add months and can potentially kill studies from ever being activated,” Dr. Markman said. He suggested that stakeholders under the sponsorship of the NCI or the IOM should work together to craft a common agreement on the use and ownership of biological specimens and intellectual property and to resolve research-for- hire concerns. “If there can be national agreement on that, this could save an enormous amount of time and effort and cost,” he said. A final issue Dr. Markman raised is the new federal law that requires all clinical trials, including industry trials, to be registered on a publicly accessible site.30 The law also mandates timely reporting of study results, but it is not clear how this will be implemented and who will have responsibility for reporting data from multi-center trials or from industry- sponsored trials. “The IOM could play a role in trying to help establish what the criteria will be for getting this information out to the public,” he suggested. He also reiterated the concern that industry sponsors are increasingly requiring cooperative group institutions to assume the major regulatory responsibility for registration trials. This is a costly responsibil- ity that should be recognized and reimbursed, he said. In the discussion that followed Dr. Markman’s talk, a participant suggested that to ease compliance with the new labeling mandate from Medicare, cooperative groups should provide a guide to their sites that specifies for each protocol what tests and procedures are considered part of routine care within or outside of the clinical trial. But Dr. Comis responded that cooperative groups do not have the resources to do this. 30  See http://clinicaltrials.gov. 

84 MULTI-CENTER PHASE III CLINICAL TRIALS He added that his studies estimate that regulatory costs make up 35 per- cent of the total costs of a clinical trial. CMS perspective The next speaker, Dr. Leslye Fitterman, epidemiologist in the Office of Clinical Standards and Quality at CMS, explained the intricacies of the agency’s coverage of clinical trial costs for its beneficiaries. Prior to 2000, she said, CMS beneficiaries were not even allowed to participate in clini- cal trials. Recognizing the need for this elderly population to have access to cutting-edge treatments, Medicare made a national coverage determi- nation that took effect in 2001, which mandated that CMS cover the rou- tine clinical care costs of its beneficiaries who participate in clinical trials, assuming that benefit categories for those costs already were established. Non-covered items and services include the specific technology being investigated, completion of case report forms, and any other services provided free of charge to other participants in the trial (CMS, 2008a). In this coverage determination CMS defined routine clinical care costs as those falling within three categories: •  onventional care that would be provided to the patient whether C or not the patient was participating in a clinical trial; •  are involved in administering a treatment and monitoring that C treatment for adverse events; and •  are linked to diagnosing and treating complications from adverse C events. There are still inconsistencies in coverage, Dr. Fitterman said, because each CMS contractor is allowed to determine whether or not an item or service is considered standard of care in its area and whether tests or procedures called for by the study protocol are necessary and reasonable. There is no national consensus on this issue, she said. Another problem with the CMS clinical trial policy is that it did not account for the disincentive toward participating in clinical research that beneficiaries of Medicare Advantage Plans face. These plans reimburse providers on a fee-for-service basis and require 20 percent copayments by beneficiaries. Many beneficiaries cannot afford the copayments for the expensive drugs often used in clinical trials.31 In addition, the masking of participants and providers in clinical trials is hampered by the obvious 31Although experimental drugs are often provided free of charge to participants, many trials combine these drugs with already licensed drugs, for which Medicare Advantage Plan beneficiaries must pay 20 percent. 

COSTS OF COOPERATIVE GROUP CLINICAL TRIALS 85 differences in copayments for reimbursements between the investiga- tional items and services among Medicare Advantage beneficiaries. In 2005 Medicare made a national coverage determination that cov- ered the off-label use of four anti-cancer drugs (NCI, 2008a). The coverage, however, was restricted to nine trials that were sponsored by the NCI. For those trials CMS began a pilot project that it is currently conducting jointly with the NCI aimed at developing tools to facilitate enrollment of Medicare beneficiaries in clinical trials. “We are trying to make things a bit easier and consistent as we can across the country,” Dr. Fitterman said. For this project the two agencies developed billing instructions that could be sent to all providers that detailed how to bill for the study drug and other covered drugs and how to bill for tests and procedures. There was explicit information on what costs Medicare would cover, what costs the trial would cover, and what costs the sponsor would cover. In addi- tion, investigators were encouraged to meet with CMS contractors to make sure that they were on the same page as to what Medicare would reimburse. “These were guidelines to cut down on the confusion and disparity about how people were billing so that claims would get processed,” Dr. Fitterman said. In addition, CMS wrote claims-processing instructions for all the contractors. Outreach to beneficiaries and their physicians was also part of the pilot project, which was done with the aid of cooperative groups and national patient advocacy groups in the gastrointestinal/ colorectal areas, as the trials were focused on those diseases. Dr. Fitterman said she was encouraged by an initial analysis that found that, for most of the trials, Medicare-eligible subjects comprised between one-fifth and one-third of the participants currently enrolled (Table 3). Later, during the discussion, Ms. Andrea Denicoff, a nurse consultant in the clinical investigations branch of CTEP in the NCI, said that traditionally only about 13 percent of people enrolled in clinical trials have been aged 65 and older. A little more than three-quarters of Medicare-eligible participants are using Medicare to cover their clinical trial costs in the pilot project, Dr. Fitterman noted. She suspects that the remaining Medicare-eligible patients are instead using some other supple- mental insurance. In 2007, the CMS clinical trial policy was modified to clearly state that if an investigational drug or other item is covered outside of the trial, it will also be covered in the trial. In addition, the policy was expanded to include Coverage with Evidence Development, which enables CMS to cover a medical intervention conditional on the agency’s concurrent col- lection of data on the intervention while reimbursing it.32 Beneficiaries 32  See http://www.cms.hhs.gov/clinicaltrialpolicies. 

86 MULTI-CENTER PHASE III CLINICAL TRIALS TABLE 3  Enrollment by Age and Use of Medicare Total Medicare Total Accrual Total Subjects Subjects Using Study ID (as of 3/31/08) Medicare Eligible (%) Medicare C80405 1,279 444 (34.7) 371 E2204 130   45 (34.6) 34 E4203 108   38 (35.2) 24 E5202 1,041 387 (37.2) 277 E5204 184   24 (13) 20 NSABP-R-04 851 246 (28.9) 175 RTOG-0522 503 109 (21.7) 98 S0502 Opened 4/15/08 — — 7325      8    2 (25) 0 SOURCE: Fitterman presentation (July 2, 2008) and Cancer Therapy Evaluation Program Clinical Data Update System, http://ctep.cancer.gov/reporting/cdus.html. can receive the item or service only if they are participating in a clinical research study or if they are in a specified registry. In the discussion that followed Dr. Fitterman’s presentation, a partici- pant said that he was unable to accrue 200 Medicare patients to a clinical trial, because the patients had Medicare Advantage Plans. “A lot of these trials use a combination of an investigational agent that would be sup- plied by the sponsor and a licensed drug that is not being supplied by the sponsor,” the participant said. “The latter becomes the deal breaker, because those are very expensive also.” Dr. Fitterman said that she rec- ognizes this is a problem and suggested a legislative movement to have Congress address the issue, along with the Office of the Actuary and the Part C program reviewing the issue internally. Private health insurer perspective The last speaker at the workshop was Dr. Lee Newcomer, senior vice president of oncology at UnitedHealthcare. He provided the private insurance company’s perspective on covering the costs linked to clinical trials. He began his talk by noting that, historically, insurance companies were reluctant to cover experimental treatments, because they were trying to prevent coverage of quackery. That reluctance evolved into a general exclusion of therapy lacking experimental evidence of effectiveness, he said. “The paradox here, of course, is how do you generate evidence without a clinical trial?” Dr. Newcomer said that he is convinced of the value of clinical tri- als in providing evidence for a highly specific treatment plan. “There is an awful lot of variation in clinical practice, and a clinical trial narrows 

COSTS OF COOPERATIVE GROUP CLINICAL TRIALS 87 that variation to a very discrete practice,” he said, noting that a clinical trial provides a very standardized protocol—a defined treatment plan— for patients. This definition of good care that clinical trials provide aids UnitedHealthcare’s coverage decisions. However, there are still looming policy concerns related to coverage of clinical trial costs, Dr. Newcomer said. These policy concerns include: • defining experimental treatment and its scope; • the need for cost-effective analyses in clinical trials; • an excessive number of “me too” trials; • payment for off-label drugs used in trials; and •  hether observational trials can substitute for randomized con- w trolled trials. The term “experimental” is rather vague, Dr. Newcomer said, and it is difficult to define what should be covered within the clinical trial space. Particularly problematic is scope creep, whereby an insurance company will cover an experimental drug offered within a trial and then be expected to also cover it when the same drug is used to treat patients outside of the trial before the drug is approved for the specific indication being tested. “I’m hearing an increasing number of complaints about this from practicing oncologists,” Dr. Newcomer said. “It has been a major friction point and barrier to getting more coverage for clinical trials.” Clinical trials not only can indicate what treatments are effective but also can show what treatments are cost-effective. For example, Dr. Scott Ramsey utilized data from a cooperative group clinical trial to show that a recommended therapy for lung cancer was about $12,000 more expensive than an alternative therapy that the trial showed was just as effective (Ramsey and Kessler, 2002). “This is going to be very important information as we move forward, because we have to address cost,” Dr. Newcomer said. “Not only should we know which therapy is more effec- tive, but we should know what it costs to gain that extra benefit or value so that we can make decisions about what we should cover.” He was critical of what he called “me too” clinical trials that duplicate other studies excessively. He noted that at one time there were 42 open trials testing high-dose chemotherapy with bone-marrow transplantation for breast cancer, but only three of those were published—a tremendous waste of resources, time, and effort, he said. Often those “me too” stud- ies are done by academics hoping that they will boost their tenure status more than large multi-center trials would, because little credit is given for team research, Dr. Newcomer said. “We have to figure out how to change that so that young investigators can still advance their careers without creating one of these ‘me too’ trials that really does not add anything to the knowledge base,” he said. 

88 MULTI-CENTER PHASE III CLINICAL TRIALS Dr. Newcomer also explained why insurers are reluctant to pay for an already approved drug that is used within a clinical trial, as often happens when it is a combination of cancer drugs and not the individual drugs themselves that is being tested for effectiveness. Referring to the industrial drug sponsor in such a trial, he said “Here is a company that is already making a profit on this drug in the 20 to 25 percent range. Why should we be providing support for its next off-label indication? That’s the cost of doing business, and they should be providing the drug free in the trial scenario.” Industry sponsors rarely pay this cost, however, Dr. Newcomer said. The increased cost of covering clinical trials is another issue that Dr. Newcomer discussed. Actuaries at UnitedHealthcare estimate that only about 93 cents out of its monthly $200 premium is for clinical trial cover- age. But studies suggest that each premium increase of only 1 percent prompts hundreds of thousands of people to suspend their insurance coverage.33 “Costs are getting unaffordable, and the way ERISA recipients will think about this is, ‘Okay, you just added another buck a month for every employee I have,’ and some people will drop out of the insurance system because of that,” Dr. Newcomer said. “You have to have a very good argument to help them understand why it’s worth paying for.” Dr. Newcomer expressed skepticism about clinical trials that use peo- ple with few health complications to test experimental treatment, because these people do not represent what is seen in the “real world.” He also questioned what he called the “tyranny of the randomized controlled trial.” The randomized, controlled trial is considered the gold standard, but given the problems with sufficient patient accrual—which are caused, in part, by patients not willing to be put into a control arm of a study—Dr. Newcomer suggested using well-run observational studies instead. These studies tend to be less expensive to run, quicker to accrue, and often pro- vide results comparable to a randomized, controlled trial. He showed a graph from one study that supported this point (Figure 5) but added that “who you enter in those observational trials has to be carefully controlled in order to get legitimate results.” Dr. Newcomer also discussed a few operational issues, such as the dif- ficulties his company has in identifying trial enrollees and the experimen- tal component of trials for payment purposes as well as the difficulties in identifying who should pay for clinical trial costs when individuals have coverage determined by a mosaic of insurance carriers, state regulations, and employer plans. Insurers must also determine which types of trials qualify to be covered. UnitedHealthcare covers all clinical trials sponsored 33  Estimates for the drop in insurance coverage range from 164,000 (Chernew et al., 2005) to 300,000 (Sheils et al., 1998) per 1 percent increase in premium. 

COSTS OF COOPERATIVE GROUP CLINICAL TRIALS 89 FIGURE 5 A comparison of randomized, controlled trials and observational studies. Solid circles represent randomized, controlled trials; open circles indicate observa- tional studies. SOURCE: Newcomer presentation (July 2, 2008) and Concato, J., N. Shah, and R. I. Horowitz. Randomized, controlled trials, observational studies, and the hierarchy of research designs. New England Journal of Medicine 342(25):1887–1892. Copyright © 2000 Massachusetts Medical Society. All rights reserved. Figure 5 bitmap by the NIH, CDC, Agency for Healthcare Research and Quality, CMS, Department of Defense, and the VA. “That’s what we use for scope,” Dr. Newcomer said, “but I think we could use any help possible in figuring out what would be the right kinds of studies to cover.” Dr. Newcomer concluded his talk with a number of suggestions: •  liminate duplication in studies and pursue novel approaches that E answer important questions; • Penalize sites for failure to accrue patients to their trials; and • Have mandatory publication of negative trials. Finally, he suggested excluding small “local” trials from coverage. New Jersey has a state law, for example, that mandates coverage for participation in any clinical trial that occurs within the state. “We know we are paying for a lot of one- and two-person trials that never gain any knowledge and are a waste of resources,” Dr. Newcomer said.