Further discussion centered on standardizing biopsy procedures so they are useful for biomarker analyses and whether health insurers will reimburse for predictive marker tests. Dr. Herbst of MD Anderson Cancer Center pointed out the need for reliability in the tumor tissue used to determine the predictive biomarker. He raised the point that biopsy tissue should be prepared in a way that is compatible with biomarker assays, and he questioned whether image guidance is needed to biopsy a “hot spot” in a tumor or to consider tumor heterogeneity. Dr. Heath responded that, ideally, measurements are made immediately on fresh tissue and that those measurements are available later to researchers who use the stored tissue. “We found that when we looked at stored tissues, including blood, one of the largest fluctuations was the protocol that was used to store it,” he said.

As for addressing tumor heterogeneity in sampling and testing tumors, Dr. Von Hoff noted that one his colleagues, Michael Barrett, has developed a technique in which he separates cancer cells into those that have the normal number of chromosomes and those that do not. He then conducts biomarker assays on both populations as a way of addressing tumor heterogeneity. Dr. Sullivan added that he endorsed the idea of image-guided biopsies and suggested this could be done with a variety of imaging modalities. He added that “it is possible to reach any place in the body with a needle under image guidance now.” He called for more collaboration among radiologists conducting image-guided biopsies, and oncologists and pathologists.

Dr. Mendelsohn raised the issue of the costs of doing biomarker assaying as part of clinical trials and who will pay those costs. Dr. Von Hoff reported that the cost of doing the biomarker assay in his Bisgrove clinical trial is $6,800 per patient and BlueCross BlueShield of Arizona agreed to reimburse that cost because they were interested in the results of the trial.

COSTS OF CLINICAL TRIALS

Clinical trials of new cancer drugs are expensive, and with implementation of better, more sophisticated studies, costs are projected to rise. Although the use of predictive biomarkers to enrich study populations might make smaller trials more likely to succeed, they pose the added costs of the biomarker tests and raise the issue of who will pay for those added costs. Currently, costs are only partially borne by NIH grants or by contracts with pharmaceutical companies. In some situations, third-party payers contribute for basic hospital or clinic services. However, considerable costs fall on the academic institutions at which studies of this type are being piloted.



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 67
 WORKSHOP SUMMARY Further discussion centered on standardizing biopsy procedures so they are useful for biomarker analyses and whether health insurers will reimburse for predictive marker tests. Dr. Herbst of MD Anderson Cancer Center pointed out the need for reliability in the tumor tissue used to determine the predictive biomarker. He raised the point that biopsy tissue should be pre- pared in a way that is compatible with biomarker assays, and he questioned whether image guidance is needed to biopsy a “hot spot” in a tumor or to consider tumor heterogeneity. Dr. Heath responded that, ideally, measure- ments are made immediately on fresh tissue and that those measurements are available later to researchers who use the stored tissue. “We found that when we looked at stored tissues, including blood, one of the largest fluctua- tions was the protocol that was used to store it,” he said. As for addressing tumor heterogeneity in sampling and testing tumors, Dr. Von Hoff noted that one his colleagues, Michael Barrett, has developed a technique in which he separates cancer cells into those that have the nor- mal number of chromosomes and those that do not. He then conducts bio- marker assays on both populations as a way of addressing tumor heterogene- ity. Dr. Sullivan added that he endorsed the idea of image-guided biopsies and suggested this could be done with a variety of imaging modalities. He added that “it is possible to reach any place in the body with a needle under image guidance now.” He called for more collaboration among radiologists conducting image-guided biopsies, and oncologists and pathologists. Dr. Mendelsohn raised the issue of the costs of doing biomarker assay- ing as part of clinical trials and who will pay those costs. Dr. Von Hoff reported that the cost of doing the biomarker assay in his Bisgrove clinical trial is $6,800 per patient and BlueCross BlueShield of Arizona agreed to reimburse that cost because they were interested in the results of the trial. COSTS OF CLINICAL TRIALS Clinical trials of new cancer drugs are expensive, and with imple- mentation of better, more sophisticated studies, costs are projected to rise. Although the use of predictive biomarkers to enrich study populations might make smaller trials more likely to succeed, they pose the added costs of the biomarker tests and raise the issue of who will pay for those added costs. Currently, costs are only partially borne by NIH grants or by contracts with pharmaceutical companies. In some situations, third-party payers con- tribute for basic hospital or clinic services. However, considerable costs fall on the academic institutions at which studies of this type are being piloted.

OCR for page 67
 IMPROVING THE QUALITY OF CANCER CLINICAL TRIALS There is a need to better understand the average cost of these sophisticated and multifaceted clinical trials, the factors driving such costs, and the drug development costs potentially saved by obtaining answers (positive or nega- tive) more rapidly. More exploration of alternative funding approaches is needed, including more public–private collaborations among academia, pharmaceutical and biotechnology companies, and government agencies. The fourth session of the conference explored these issues with presenta- tions by Robert Comis of the Coalition of National Cancer Cooperative Groups, Kevin Schulman of Duke University Medical School, and Gwen Fyfe of Genentech. Dr. Comis began by noting that for Cooperative Group-funded oncology studies,12 the overall per-patient cost is about $6,000, of which only $2,000 is reimbursed by government grants (The Lewin Group, 2005). In contrast, one study of four companies found that the per-patient costs for industry-spon- sored studies ranged from $60,000 to $85,000 for Phase III studies (of which about $15,000 to $18,000 is reimbursed), and from $46,000 to $85,000 for Phase II studies (of which $20,000 to $25,000 is reimbursed). It is generally much less expensive to conduct a clinical trial in a foreign country, with the costs of certain cancer clinical trials in Western Europe being nearly half the cost of the same trial conducted in the United States.13 Regulatory Costs Prepatient start-up costs account for a significant portion of clinical trial costs, Dr. Comis said. For publicly sponsored Phase II or III studies, these costs are about $5,000, and about $8,000 for privately sponsored studies (The Lewin Group, 2005). Much of this cost is due to addressing regulatory requirements of various institutional review boards (IRBs) and government agencies, Dr. Comis and Dr. Schulman pointed out. Of nine functional steps identified for the conduct of high-quality trials, six include elements 12The Clinical Trials Cooperative Group Program, which is sponsored by the NCI, is designed to promote and support clinical trials of new cancer treatments, explore methods of cancer prevention and early detection, and study quality-of-life issues and rehabilitation during and after treatment. Cooperative groups include researchers, cancer centers, and community physicians throughout the United States, Canada, and Europe. They work with the NCI to identify important questions in cancer research and to design clinical trials to answer these questions. 13TrialSpace Grants Manager (see http://www.fasttracksystems.net/grantsmanager_ products.php).

OCR for page 67
 WORKSHOP SUMMARY Box 2 Regulations Govern Most Functional Steps Required for Conducting Studies • Of nine functional steps identified for the conduct of high-quality trials, six include elements related to federal regulations: – Institutional Review Board Submission – Site Approval – Preparation for Study execution – Study execution – Data Review – Study Closeout • An average of 35% of clinical research costs is spent on compliance. SOURCeS: Comis presentation (October 5, 2007) and The Lewin Group (2005). related to federal regulations (Box 2). “There are numerous regulatory func- tions that are included in bringing a study up, including the IRB costs, the approvals with the FDA and with the NCI, etc., and our estimate from working with those sites is that about 35 percent of the costs that accrue for a clinical trial relate to regulatory issues and regulatory compliance,” Dr. Comis said. Dr. Schulman added that there is often “protocol creep”—after a protocol has undergone the regulatory review process with various agen- cies and internal boards, so many additional research steps are required that the actual cost for the study far exceeds the amount budgeted. In the discussion following Dr. Comis’s talk, conferee Dr. Birch pointed out that much of the paperwork required for regulatory approvals has been done in previous trials, and both time and money would be saved through a national database for this type of information. Dr. Schilsky added that there is unnecessary redundancy and a lack of harmonization among the multiple organizations that review clinical trials. For example, in addition to being reviewed by the NCI’s central IRB, trials have to be reviewed by their own institution’s review boards, in part because the Department of Health and Human Services Office for Human Research Protections also

OCR for page 67
0 IMPROVING THE QUALITY OF CANCER CLINICAL TRIALS allows institutions to exercise local control over trials. Dr. Schulman agreed, adding, “Having huge burdens related to the regulatory process that aren’t adding value is costing us.” Dr. Comis suggested that “We have to eliminate the bureaucracies in order to get advances for cancer patients.” Because the fixed start-up costs are independent of the number of subjects enrolled in a clinical trial and are so large, having higher patient accrual in fewer studies is more economically efficient than having lower accrual in more studies, Dr. Comis noted. Yet only 56 percent and 63 percent, respectively, of open government trials and open industry trials had subjects enrolled, one study found (The Lewin Group, 2005). Later in the discussion, Dr. Doroshow reported on a recent NCI study of four larger NCI-funded Comprehensive Cancer Centers. In those centers 25 percent of their trials accrued no patients, and 26 percent accrued four or fewer patients. A review of those four cancer centers along with two large cooperative groups and CTEP revealed that the amount of time it takes to start up a study is nearly 3 years, Dr. Doroshow noted. Start-up costs are not separately covered by government-sponsored studies. “There is a lot of inefficiency in the system,” Dr. Comis said. “No site should ever bring up a study for which they don’t have the patient or research resources. By the time you enter one patient in a study, you have already spent 5 or 6 thousand dollars that neither government nor industry reimbursement will probably make up.” But Dr. Doroshow voiced some optimism by noting that in 2005, NCI-supported cancer centers accrued 41,000 patients to treatment trials, with only a quarter of those supported by pharmaceutical companies. “That is an enormous amount of accrual—of those 50 centers, that is roughly about 20 percent of their patient population,” he said. “Comprehensive cancer care centers do a very good job accruing patients to clinical trials.” However, he added, a recent evaluation for the cooperative group system between 2000 and 2005 showed that while 463 sites accrued only 1 to 5 patients in 5 years, at 12 sites accrual was more than 500 patients for the same time period. “It is very clear that we do extraordinarily well at certain places. I think we can compete very effectively with Europe if we focus on enhancing the efficiency of our system,” Dr. Doroshow said. Patient Accrual Much of the rest of this discussion focused on the problems associated with patient accrual and how to address them. Dr. Fyfe pointed out that

OCR for page 67
 WORKSHOP SUMMARY Genentech has had enormous difficulties recruiting patients for their cancer drug clinical trials because most patients in this country are not willing to agree to randomization for drugs. This forces many companies to move their studies to other countries where it is easier to accrue patients. “The recruitment percentages for patients in the United States is only about 2 or 3 percent, which is unfortunate because we are going to lose the opportunity to understand how drugs that will be marketed in the United States actually perform in the milieu of care in the United States as opposed to Ukraine or Russia or places where the standard of care is so decidedly different,” Dr. Fyfe said. Dr. Comis agreed and added that more than half of patient accruals in the United States come from community-based practices, which are increasingly being endangered by financial pressures. Margo Michaels from the Education Network to Advance Clinical Tri- als asked what the cooperative groups are doing at the local level in terms of working with community and advocacy groups to inform people about the types of trials offered and to increase accrual. Dr. Comis responded that the Coalition of Cancer Cooperative Groups14 has been working closely with the American Society of Clinical Oncology, the American Cancer Society (ACS), and other major cancer-related organizations to inform physicians and patients about cancer clinical trials currently accruing patients. For example, the Coalition helped the ACS establish a search engine on its website, called Trial Check, which can be used to search for cancer clinical trials. Dr. Von Hoff suggested using the “Just-in-Time” approach to improve patient accrual. With this approach, rather than having sites activated prior to screening for patients, a centralized IRB-approved protocol is activated only at sites that have enough potential patients to participate in the trial. One recent study found this approach improved patient accrual and reduced trial-related costs in a pancreatic cancer trial (Wiener et al., 2007). Dr. Comis noted that complicated trials tend not to accrue as well as simpler studies, and the biggest draws for patient volunteers are clinical trials that offer a new drug as opposed to a new use of an established drug. This is in contrast to many cancer trials, which are done on “me too” drugs or on approved drugs being studied for new indications. Discussant Dr. 14The Coalition of Cancer Cooperative Groups is a nonprofit organization dedicated to improving patient awareness of clinical trials, facilitating access, and promoting participation. It is composed of members from 10 NCI-sponsored Cooperative groups, the country’s leading patient advocacy organizations, and thousands of oncology and cancer research specialists.

OCR for page 67
 IMPROVING THE QUALITY OF CANCER CLINICAL TRIALS Schilsky said studies show that the main influence on patient accrual is the physician, and that a minority of physicians in any practice setting refer the majority of patients to clinical trials. “It is anywhere between 20 to 30 percent of the doctors who enroll 70 to 80 percent of the patients who are in clinical studies,” he said. “There is a core group of committed physicians out there who do most of the accrual and there are a lot of physicians who give lip service to it and do nothing.” Dr. Schilsky stressed that there is no incentive for physicians to enter their patients into clinical trials, but rather disincentives because such entry requires more time and effort on the physician’s part. He estimated that for the average oncologist, the time needed to discuss a clinical trial with a patient and gain consent is probably at least three or four times the amount of time needed to discuss the standard chemotherapy the patient will be receiving. Dr. Schilsky suggested prompting more physicians to refer patients to clinical trials through a reimbursement mechanism, or by mak- ing referral a requirement for maintaining their credentials. “If there was a billing code that allowed a doctor to bill at a higher rate for managing a patient on a clinical trial than for giving off-protocol care, that might give an incentive to doctors to actually put patients on trials,” Dr. Schilsky said. Conferee Dr. Patricia Ganz concurred and added that enrolling patients in clinical trials “should be the norm and it is only if you don’t qualify or the patient doesn’t want to participate that this doesn’t occur. This would enable a much more efficient drug evaluation process as well as high-quality care,” she said. Global Outsourcing Dr. Fyfe described the cost savings that could be gained if patient recruitment occurred more quickly and more robustly, as the length of a trial is often extended because of delays in patient accrual. If 25 to 50 per- cent of all eligible cancer patients enrolled in clinical trials, she said, “the cost of trials would decrease dramatically because you would have fewer sites, trials would take less [time] to recruit, and you would have your answer much more quickly. The reality is a lot of drug costs come from the fact that 20 percent of the physicians give you 70 percent of the patients and at this point in cancer, people have to go to the rest of the world because they can’t get the patients here.” Clinical trials are increasingly being conducted outside the United States. Dr. Comis noted with a graph (Figure 18) showing that between

OCR for page 67
 WORKSHOP SUMMARY 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% 1997 1999 2001 2003 2005 United States Western Europe Rest of World FIGURE 18 Clinical investigations are going global, as shown by the distribution of 1572 forms by location of investigative site. NOTE: A 1572 form must be submitted to the FDA by a clinical investigator prior to initiating a study in human subjects. SOURCE: Comis presentation (October 5, 2007), reprinted, with permission, from Outlook 00, 2007. Copyright 2007 by Tufts Center for the Study of Drug figure 18 Development. 1997 and 2005, the percentage of clinical trials being conducted in this country fell from 85 to 65 percent (Tufts Center for the Study of Drug Development, 2007). The United States ranked sixth in the countries conducting the most clinical trials, Dr. Comis said. Discussant Dr. Samir Khleif from the NCI pointed out that the majority of the countries that rank above the United States, such as France and Canada, have socialized medicine in which physicians are paid salaries. Therefore, spending more time enrolling patients in clinical trials does not have a negative impact on their income. But Dr. Schulman urged caution in conducting clinical trials overseas. “It is cheaper to have a patient enrolled in your study in Eastern Europe at a per-patient level, but if you are not there, at the study site, making sure that they are not fabricating the data, it may not be less expensive at the end of the day, and there have been plenty of trials where an entire country’s worth of data became lost as a result of that [data fabrication],” Dr. Schulman said.

OCR for page 67
 IMPROVING THE QUALITY OF CANCER CLINICAL TRIALS Dr. Doroshow suggested that the electronic data capture system that is currently being finalized at the FDA and the NCI will improve the effi- ciency of clinical trials and prevent industry from globally outsourcing such studies. “It will enable the most efficient system for cancer clinical trials in the world, and studies will get done at a level of quality that just are not going to be comparable because we are spending so much time develop- ing the infrastructure that is going to be unique,” he said. Dr. Schulman countered that electronic data capture will not sufficiently speed up the time it takes to start up a clinical trial because much of that time is due to “onerous” regulations. In his presentation, Dr. Schulman delineated the other costs in addi- tion to start-up and patient accrual costs that explain why clinical trials are so expensive. These include the costs of patients, procedures, trial materi- als and distribution, as well as the costs of site and data management and statistical analyses. Obviously, the more patients a trial requires, and the more procedures done on those patients, the more costly the trial will be. Although the use of predictive biomarkers might reduce the number of patients required for a trial, they add the costs of conducting those tests on every patient who enters the trial, Dr. Schulman pointed out. They also raise the possibility that findings from studies that use them to enrich study volunteers may not be relevant to lower risk populations, effectively reduc- ing the potential market size for the drug. Time Is Money Time is also a big cost driver, Dr. Schulman noted. The less time a new drug takes to make it to the market, the more time that is left on the drug’s patent and the more likely investors will financially support the drug’s development. A short time frame is also needed to ensure timely cash flow from sales of the drug that can offset its development costs. “Time is critical and is a huge cost that actually is probably larger than the difference in the cost per patient. Everyone says that every day’s delay in making it to market is about a million dollars in the life cycle of a product,” Dr. Schulman said. A trial with long patient accrual or follow-up times consequently is going to be costly to run. Biomarkers that can predict survival and lessen follow-up time may reduce the duration of clinical trials, but as Dr. Schulman noted, “The ques- tion is whether the information is going to be as predictive as the survival data.” Such biomarker tests also are an added cost in the study that may

OCR for page 67
 WORKSHOP SUMMARY not be trivial. “The more different measures we have in the study, the more expensive the study is,” he said. “The more information we collect, the more we have to validate that information, and this adds costly complexity.” Biomarker tests also increase the data analysis time, so they are not used as effectively as they could be. “We are rushing forward without actually using all the information we have, which makes it harder to put in more complex analyses in terms of some early-stage biomarkers,” Dr. Schulman said. Data management and analysis for large biomarker/genomic studies will be increasingly difficult and complex. “There are obviously not enough statisticians in the world to figure out how 23,000 genes interact with each other in cancer,” he said. Electronic data capture and health records might help to alleviate some of the current costs linked to data management and analysis, he added; but the more complex the data are, the more costly it is not only to collect the data but also to make sure the dataset is accurate and to sift through in response to specific queries. The cost for each resolved data query may average between $50 and $90, Dr. Schulman noted. Site management also has costs, which are associated with ensuring that site investigators are following the protocol. “Every different sponsor has a different way they want things done and that increases some of the complexity and costs,” Dr. Schulman said. One hidden cost is the cost of “loser” drugs. The estimate that each drug approved costs about a billion dollars to develop includes the $4 or $5 million cost of failed drugs, Dr. Schulman noted. To deal with those losses, the drug industry is increasingly relying on biotechnology companies to take on the financial risks of early drug development so “private investors are the ones that soak up a large number of losers right now,” he said. Later in discussion, Dr. Mendelsohn pointed out that “if we could figure out ways to ditch products that are not going to work better, we could save a whole lot of money.” An alternative, Dr. Schulman added, would be to devise dif- ferent schemes for paying for failed drugs. Public funding for Phase I trials would reduce much of that cost, he said. Public–Private Collaborations Other suggestions were made on how to alleviate the costs of cancer clinical trials. Such trials are funded by either industry or the NCI. Industry tends to do early-phase testing and the NCI tends to fund mainly Phase III trials, most of which are to extend the indications of already approved drugs. “There is a symbiosis in the country between the public side of the system

OCR for page 67
 IMPROVING THE QUALITY OF CANCER CLINICAL TRIALS and the private side of the system, which, in the end, benefits all patients and all cancer patients,” Dr. Comis said. But an average of 29 percent of a clinical trial site’s clinical research revenue originates from nontrial sources, he said (The Lewin Group, 2005). “This takes a tremendous amount of institutional commitment, which is true for both academic and community practices,” Dr. Comis added. Dr. Schilsky pointed out that there is a tremendous lack of public resources for cancer clinical trials. The budget for the cooperative groups is about $150 million annually, and that is used to support about 500 active clinical trials, including about 80 or 90 Phase III trials, he said. “Tell me any pharmaceutical company that can operate 80 or 90 Phase III clinical trials on a $150 million budget. It doesn’t happen and so we are relying enormously on the contribution of our investigator population and the institutions that participate in the cooperative groups to make up the dif- ference. It is getting more and more difficult for them to be able to do that,” Dr. Schilsky said. Dr. Comis warned that the U.S. public system that girds cancer clinical trials is grossly underfunded and endangered, and the U.S. private system is challenged by foreign competitors. A serious effort to combine resources, increase efficiency, and decrease regulatory burden will be required for the United States to continue to be on the forefront of cancer clinical research, Dr. Comis said. A few discussants and speakers suggested that CMS take a more active role in funding cancer clinical research because the large major- ity of people who develop cancer are Medicare recipients who would benefit from such studies. “I would be happy to propose no cancer patient would get paid for therapy within Medicare unless they were in a registry, unless we had some access to their tissue because this is all an experiment. We don’t know all the answers and the quicker we can get some resolution, then the less money it will cost Medicare,” said Dr. Schulman. When a clinical trial done by an NCI-funded cooperative group has regulatory implications (e.g., if it will be a registration trial for a drug), the additional costs linked to that registration increasingly are paid for by the drug’s sponsor, Dr. Comis said. Using this as a model, with judicious nego- tiating and planning with industry, the cooperative groups might be able to double their budget so that half comes from drug companies and half comes from the NCI, Dr. Mendelsohn pointed out. Dr. Comis agreed and noted that a similar model is already in use in Canada for funding clinical trials.