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 68
Transforming Clinical Research in the United States: Challenges and Opportunities BOX 6-1 Current Initiatives of NCI’s Clinical Trials Working Group to Optimize the Clinical Trials System Coordination through information sharing and collaboration. In addition to increasing collaboration within cooperative groups, NCI is working to enhance coordination with other NIH institutes (e.g., the National Heart, Lung, and Blood Institute [NHLBI] through the Bone Marrow Transplant Clinical Trials Network) and international networks (both industry-sponsored and those sponsored by the country of origin). A successful international collaboration is the ongoing trial in osteosarcoma, a rare tumor in children, between the U.S. Children’s Oncology Group and several pediatric oncology clinical research networks in Europe. Each country adheres to its own regulatory and human subjects protection requirements, but the trial has a single, central coordinating center and one data safety monitoring database. Prioritization of clinical trials for funding. Scientific Steering Committees have been created to oversee particular disease areas within oncology (gastrointestinal, gynecologic, head and neck, genitourinary, breast and lung, and hematologic malignancies, as well as symptom management and health-related quality of life). The committees include broad representation from practicing physicians with clinical trial experience, translational scientists, biostatisticians, community oncologists, and patient advocates. All phase III treatment trials sponsored by NCI’s Clinical Trials Cooperative Group Program are evaluated, prioritized, and approved by those committees. Standardization and promotion of common tools. Through CaBIG at NCI, a number of technological tools will be standardized. The most recent of these is the Oncology Patient Enrollment Network (OPEN), a Web-based portal that provides a centralized enrollment system for registration of patients in NCI-sponsored cooperative group clinical trials and patient randomization across the cooperative group system. OPEN clinical research personnel—investigational staff and the infrastructure to support the clinical trial; clinical supplies made by industry—procuring of comparators (drugs used in the control arm of a clinical trial), which often involves relabeling and approval to use the comparator as an experimental agent; processing of trial-related specimens—acquiring and banking tumors and biological fluids; negotiating of research grants;
OCR for page 69
Transforming Clinical Research in the United States: Challenges and Opportunities should improve site activation of clinical trials and monitoring of patient recruitment to trial protocols. Operational efficiency of trials. NCI is reviewing strategies for improving the speed with which trials are initiated and conducted and is focusing on two key areas: Bottlenecks in trial development—A recent study (Dilts et al., 2009) found that for NCI-sponsored phase III clinical trials, the time from concept submission to trial activation by a cooperative group required a median of 602 days. The study also revealed that at least 296 distinct processes are required for activating a phase III trial, including 238 working steps, 52 major decision points, 20 processing loops, and 11 stopping points. NCI hopes to streamline this process and has set a goal of reducing the time from concept approval to trial activation by 50 percent. Possible areas of improvement in the process include Institutional Review Board (IRB) approval and industry contracting. Interaction with the U.S. Food and Drug Administration (FDA)—If an NCI-sponsored phase III trial involves a treatment with a potential licensing indication, a company may express interest in the trial and become a barrier to its activation. Thus, it is crucial to obtain FDA review of and comments on a potential licensing indication at an early stage. To eliminate this source of delay in conducting trials, NCI has developed a process for obtaining FDA input on a phase III trial at the concept stage so that a company’s licensing of a potential indication can be completed as early as possible. Enterprise wide restructuring and oversight. NCI-wide advisory committees and coordinating groups have been created to oversee the entire clinical trials enterprise and progress on the four initiatives described above. Mooney is hopeful that the next several years will yield improvements in efficiency across the NCI cooperative group system. adjudication committees’ fees—more of an issue outside of the United States (adjudication committees are necessary when endpoints of time-to-event are used, such as progression-free survival in cancer); and fees associated with IRBs and Data Monitoring Committees (DMCs). At each phase of clinical research (phases I, II, and III), the cost increases. Canetta explained that the later the stage of development in which a compound fails, the higher the cost of that failure will be. In cancer, the
OCR for page 70
Transforming Clinical Research in the United States: Challenges and Opportunities rate of success for bringing a compound through the drug development process to patients is less than stellar. Thus, there is significant interest in reducing the cost of clinical research and thereby the cost of drug development failures. Canetta mentioned three aspects of clinical research that have the potential for cost reduction: Data collection—Standardized case report forms (CRFs) would help investigators conduct a trial more efficiently. Also, reducing the number of data points that require monitoring for each patient in a clinical trial (i.e., selective monitoring) could make it possible to reduce cost while maintaining quality. Comparator and experimental drug charging—Acquiring and relabeling expensive comparator drugs for a clinical trial is a significant cost driver. Canetta suggested that comparator drugs being used in a clinical trial for an approved indication could be paid for by the insurance industry as a way to induce more patients to enroll. Time cost—As discussed throughout the workshop, activating clinical trials has become a lengthy process. Canetta identified four aspects of the clinical trial initiation process that could benefit from increased efficiency: (1) internal review by the sponsor, (2) contract negotiations with institutions and investigators, (3) local regulations (IRBs), and (4) special protocol assessments (from the FDA in the United States) or scientific advice (outside the United States). Canetta reported that historically the internal review process at Bristol-Myers Squibb involves 34 review cycles for each individual trial protocol, totaling 8 months for the company to produce/activate a trial protocol. Efforts are currently under way to bring the company’s timeline for internal review down to 5 months by aligning review cycles with various internal functions. The time to activate a clinical trial protocol varies across institutions and clinical trial sponsors. In the United States, for example, the Eastern Cooperative Oncology Group (ECOG) requires a median of 808 days to complete the steps necessary to activate a clinical trial protocol (Dilts et al., 2008). Canetta presented data from individual institutions revealing shorter times to activation. At the University of Arkansas, for example, the median is 70 days. Canetta suggested that this shorter time is due to the fact that the university is a small operation and thus can streamline internal processes more easily. Outside the United States, the time required for approval of clinical trial protocols are very similar to those in the United States—that is, lengthy.
OCR for page 71
Transforming Clinical Research in the United States: Challenges and Opportunities 7 Clinical Trials in Diabetes An estimated 7.8 percent of the U.S. population has diabetes, a chronic disorder affecting the body’s metabolism. The most common form is type 2 diabetes, affecting approximately 90–95 percent of those with the disease. Type 2 diabetes is most often associated with older age, obesity, a family history of diabetes, physical inactivity, and certain ethnicities (NDIC, 2008). In addition, new research has also improved our understanding of the genetic underpinnings of type 2 diabetes. The diagnosis of type 2 diabetes includes the identification of insulin resistance, or the body’s inability to process insulin, which ultimately results in a build up of glucose in the body. In contrast to type 1 diabetes, the symptoms of type 2 diabetes develop slowly over time. Recent research focuses on preventing or delaying type 2 diabetes in at-risk populations and has revealed that lifestyle interventions and some medications can reduce the development of diabetes. The origin and progression of type 1 diabetes is notably different from that of type 2. Jay Skyler, Chairman of the Type 1 Diabetes TrialNet, explained that individuals are born with a genetic predisposition to type 1. Autoimmune, genetic, and environmental factors are believed to play a role in the immune system’s attack on insulin-producing beta cells and the development of this form of the disease (NDIC, 2008). At some point during early life, perhaps even in utero, an environmental trigger initiates such an attack. Prior to the clinical appearance of type 1 diabetes through an oral glucose test, a number of stages in the development of the disease are amenable to intervention. Intervention studies can be conducted in an attempt to develop methods or therapies that can interrupt the process of developing
OCR for page 72
Transforming Clinical Research in the United States: Challenges and Opportunities the disease. Because type 1 diabetes affects such a small proportion of the diabetic population, however, there has been less investment in the development of new therapies for the disease relative to the more prevalent type 2. This chapter begins with a discussion of government-sponsored diabetes clinical trials. Next, a clinical research network—TrialNet—is described, along with the ways in which it conducts trials in diabetes. A case study that illuminates some of the strengths and weaknesses of government versus industry-sponsored clinical trials in diabetes is then presented. Finally, the chapter turns to innovative ways in which regulatory challenges to conducting clinical trials can be overcome. GOVERNMENT-SPONSORED TRIALS IN DIABETES Judith Fradkin, Director of the Division of Diabetes, Endocrinology, and Metabolic Diseases in the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) at the National Institutes of Health, discussed NIDDK’s research efforts and the process of conducting government-sponsored clinical trials in diabetes. She explained that NIDDK seeks to conduct diabetes trials that are typically not pursued by drug companies. Because NIDDK’s focus has been on conducting clinical trials evaluating various approaches to diabetes therapy as opposed to analyzing particular drugs, large-scale trials with a fairly long timeline are necessary. The NIDDK Clinical Trial Development Process Fradkin discussed the advantages and disadvantages of NIDDK’s process for developing and implementing clinical trials. Advisory groups first identify an important research question. Once the agency has determined that funding exists to support a new research initiative, a competitive Request for Applications (RFA) is issued. After reviewing applications submitted by investigators, NIDDK makes awards to a data coordinating center and clinical sites to design and implement the trial. At this point, the process can move quite slowly as a diverse group of diabetes and clinical trial experts have many different ideas about how the trial should be designed and conducted. In addition, a significant amount of money is flowing during the trial design process, and contracts and negotiations surrounding these financial transfers can affect general progress on trial development. Fradkin explained that the RFA process for identifying diabetes trial investigators has advantages that include the diverse expertise that is brought to bear in the trial design phase and the rigorous, iterative process of site selection, which has resulted in highly robust multicenter clinical trials that have transformed diabetes therapy. When NIDDK issues an RFA for a new study, considerable uncertainty exists regarding such issues as the primary
OCR for page 73
Transforming Clinical Research in the United States: Challenges and Opportunities outcome of the trial, the sample size, the effect size, and the retention rate of trial participants. Because of these factors, the duration of the trial and its total budget are unknown throughout the RFA process. Thus, disadvantages of the RFA process are its length, uncertainty regarding the feasibility of the trial design, and the undetermined budget. NIDDK recently tried a new approach to conducting diabetes clinical trials—the investigator-initiated planning grant. In this process, a principal investigator (PI) assembles a team of investigators and receives a planning grant to develop a trial protocol and a manual of procedures. Compared with the RFA process, the planning grant has the potential to be more efficient (shorter process, cost savings in the trial design phase, and a budget that is largely determined prior to initiation). On the other hand, because investigators are chosen by the PI, diverse viewpoints may be minimized in the planning grant process. Developing Informative Clinical Trials for a Chronic, Heterogeneous Disease Since NIDDK-sponsored clinical trials seek to examine approaches to diabetes therapy rather than particular drugs, subjects are often followed after the trial has ended or after its primary outcome measures have been assessed. Because diabetes is a chronic disease with complications that develop over a long period of time, lengthy follow-up increases understanding of the disease. Fradkin stressed that critical scientific findings have resulted from this follow-up after the completion of a trial. For example, the Diabetes Control and Complications Trial studied more than 1,000 type 1 diabetes patients, comparing intensive glucose control with standard glucose control over a 6.5-year study period. The trial revealed that intensive glucose control dramatically reduced the rate of development of complications associated with type 1 diabetes. By continuing to follow these patients for an additional 10 years, however, it was discovered that the benefits of the finite period of intensive glucose control were prolonged well into the future; a “metabolic memory” was created, even once the glycemic control of the two groups was similar. Trials in type 2 diabetes revealed similar findings. Fradkin noted that these findings from extended patient follow-up provided information on the importance of good glucose control early in the course of diabetes, before complications develop. In addition, significant pathophysiologic information gained from the prolonged follow-up provided a greater understanding of the etiology of the disease and how complications develop over time. At the same time, the resources expended on follow-up can limit the ability of NIDDK to initiate new studies, a consideration that becomes increasingly important as financial resources are limited and NIH funding remains flat.
OCR for page 74
Transforming Clinical Research in the United States: Challenges and Opportunities A prolonged study time to evaluate the progression of diabetes also yields benefits for payers in the health care system. For instance, some outcomes in a randomized controlled trial (RCT) may be insufficient for payers. In the case of evaluating the effects of a lifestyle intervention to delay or prevent the onset of diabetes, payers seek information on the effects in preventing diabetes and on how durable those effects are. A study that evaluates the extent to which patients cross over the diabetes continuum (i.e., from pre-diabetes to diabetes) is generally less informative than one that evaluates the extent to which preventing diabetes stalls the complications of the disease. To provide more informative trial results on the long-term effects of diabetes prevention, NIDDK has invested significant resources in a follow-up study to the Diabetes Prevention Program (DPP). Studying type 2 diabetes also introduces some challenges to the design of informative trials. For instance, the pathophysiologic heterogeneity of type 2 diabetes (i.e., the wide-ranging combinations of symptoms exhibited by patients) can make it difficult to identify which subsets of patients actually respond better to a certain drug or approach to therapy. For example, different type 2 diabetes patients have different combinations of insulin resistance and decreased beta cell function. When these heterogeneous manifestations of the disease are combined into a single group based on a given glycemic level for the purposes of a clinical trial, the opportunity to identify patients who benefit from a particular therapeutic approach can be lost. Fradkin mentioned that the heterogeneity of type 2 diabetes has assumed a larger role as the number of different classes of drugs to treat the disease has increased over the years, making it especially important to identify the subset of patients who respond better to a certain drug. Fradkin also highlighted the success of NIH/NIDDK multicenter trials in recruiting racially and ethnically heterogeneous populations, suggesting that NIH studies have the advantage over industry-funded trials in this regard. For example, the government-sponsored DPP included 45 percent minority populations. The study looked at diabetes incidence rates for three study arms—lifestyle intervention, metformin, and placebo. The benefits of lifestyle interventions and metformin in reducing the incidence of type 2 diabetes appeared to be manifest across all of the ethnic and racial groups studied (see the further discussion of these results below). That is, the progression rate of type 2 diabetes in the placebo group did not differ by race/ethnicity. This is an especially critical finding given that type 2 diabetes affects minority populations disproportionately. In addition to the inclusion of ethnically diverse populations, government-sponsored trials have excelled in characterizing patients with diabetes by phenotype. Careful phenotyping in the Diabetes Control and Complications Trial included measures of C-peptide, an indicator of how much insulin beta cells are producing, and resulted in the striking finding that patients
OCR for page 75
Transforming Clinical Research in the United States: Challenges and Opportunities with some residual C-peptide did better in terms of glycemic control and decreased hypoglycemia. As a result of this finding, the FDA has agreed to allow C-peptide to serve as a clinical endpoint for type 1 diabetes new-onset trials. This unanticipated finding resulting from phenotyping in the Diabetes Control and Complications Trial has had important effects on the development of type 1 diabetes trials. In the DPP, phenotyping by means of pharmacogenomics analyses revealed that even for those individuals at increased genetic risk for type 2 diabetes, lifestyle interventions were effective in decreasing their risk for the disease. Fradkin noted that this was a powerful result—genetics is not destiny in terms of developing type 2 diabetes. NIDDK: A Model for Clinical Trial Collaborations NIDDK’s research is highly collaborative—most of its studies involve working with other NIH institutes, according to Fradkin. These collaborations have resulted in unanticipated yet important findings. For example, in the collaboration with the National Institute on Aging (NIA) on the DPP, it was a prerequisite that at least 20 percent of the clinical trial participants be older patients. Initially, investigators were concerned that older patients would be unwilling to participate in the lifestyle intervention aspect of the study. NIA countered that the highest prevalence of diabetes is in older populations, so they should be included in the study. As it turned out, the study revealed that older patients were more sensitive than other age groups to the lifestyle change. Given the prevalence of diabetes in older adults, the protracted time course of the disease, and the fact that diabetes is a major driver of Medicare costs, Fradkin believes diabetes is a good candidate for collaboration between NIH and the Centers for Medicare and Medicaid Services (CMS). Very few of the practices Medicare pays for have been rigorously examined in RCTs. Given NIDDK’s track record in conducting paradigm-shifting diabetes trials in diverse populations, conducting clinical trials in the Medicare population could offer an opportunity for cost savings. NIH would pay the research costs, CMS would pay the costs of providing clinical care, and the trial results would have the benefit of being conducted in a real-life health care setting. Tracking clinical trial results via Medicare beneficiary claims would generate meaningful, long-term outcomes with potentially compelling economic cases. TRIALNET: A NETWORK APPROACH TO TYPE 1 DIABETES TRIALS Funded jointly by NIH, the Juvenile Diabetes Research Foundation International (JDRF), and the American Diabetes Association (ADA), as
OCR for page 76
Transforming Clinical Research in the United States: Challenges and Opportunities well as a special appropriation from Congress, TrialNet is an international network of researchers exploring ways to prevent, delay, and reverse the progression of type 1 diabetes.1 TrialNet researchers are drawn from 18 clinical centers in Australia, Canada, Finland, Germany, Italy, New Zealand, the United Kingdom, and the United States. More than 150 medical centers and physician offices participate in the TrialNet network. Skyler described the TrialNet protocol development process and the network’s efforts to end type 1 diabetes. TrialNet receives protocols from investigators within the network, external academic investigators, and industry. The network conducts trials in a range of type 1 diabetes areas, including natural history, prevention (including vaccines), treatment for early onset, and mechanisms of action. Four TrialNet committees initially review protocols: The Scientific Review Committee examines the scientific validity of the study’s approach. The Clinical Feasibility Committee examines whether the study protocol can reasonably be implemented. The Ethics Review Committee weighs ethical considerations of the study design and its practical implementation. The Infectious Disease Safety Review Committee ensures that immunomodulatory agents are being used properly. Based on the recommendation of an Institute of Medicine (IOM) report that the scientific review, ethical review, and subject safety functions be carried out separately, these four independent committees review the proposed study protocol before sending their results to the Intervention Strategies and Prioritization Committee. That committee includes members from both TrialNet and outside organizations, such as JDRF and the Immune Tolerance Network, as well as international experts. Once the protocol has been approved, the Protocol Development Team uses its standardized tools (e.g., case report forms) to translate protocol procedures into practice. Simultaneously, the Protocol Committee, consisting of the person who originally proposed the study and others with expertise in the study area, collaborates to further develop the study protocol and finalize its use. To complete the process, the TrialNet Chairman’s Office, the Coordinating Center, center directors, and trial coordinators implement the protocol and carry out the study. According to Skyler, his experience in conducting clinical trials in type 1 diabetes suggests, first, that clinical decisions should be based not on 1 Additional information on TrialNet can be found at http://www.diabetestrialnet.org/index.htm.
OCR for page 77
Transforming Clinical Research in the United States: Challenges and Opportunities small pilot studies but on adequately powered RCTs. Second, clinical trial designs should not be changed in the middle of the trial. If a design change is necessary, the analysis of trial data should account for the impact of that change. And third, study subjects in type 1 diabetes trials should be within the age range of 9–15 as this is the age of peak onset of the disease. CASE STUDY: GOVERNMENT- VS. INDUSTRY-SPONSORED TRIALS IN TYPE 2 DIABETES Steven Kahn, Professor of Medicine in the Division of Metabolism, Endocrinology, and Nutrition at the University of Washington and VA Puget Sound Health Care System in Seattle, compared two RCTs in the area of type 2 diabetes. One was a government (NIH)-sponsored prevention trial and the other an industry-sponsored intervention trial. The primary aim of the NIH-sponsored trial, DPP, was to examine whether type 2 diabetes can be prevented in people with impaired glucose intolerance. The three intervention arms of the trial were (1) metformin, the commonly used first-line therapy for type 2 diabetes; (2) lifestyle changes aimed at weight loss and increased exercise duration; and (3) placebo. After 4 years, it was found that metformin reduced the risk of developing diabetes by 31 percent in study subjects. Lifestyle changes had an even greater impact—a 58 percent reduction in the risk of developing diabetes compared with placebo (no treatment). The benefits of metformin and lifestyle changes were so dramatic that the data safety monitoring board (DSMB) stopped the study early because continuing the placebo study arm was considered unethical. The second trial, A Diabetes Outcome Progression Trial (ADOPT), sponsored by GlaxoSmithKline, was a head-to-head comparison of three different marketed drugs (rosiglitazone, metformin, and glyburide) for people recently diagnosed with type 2 diabetes. ADOPT was a large, multicenter, international clinical trial. After 4 years of follow-up, it was found that glyburide was the least effective of the three drugs in maintaining glucose control, rosiglitazone was the most effective, and metformin was intermediate. ADOPT was a landmark clinical trial that changed first-line treatment decisions in favor of drugs that maintain glucose control to a greater degree. It was also unique for the pharmaceutical industry to engage in a comparative effectiveness study that explored issues beyond whether a drug can lower glucose. Through the ADOPT results, broader areas of diabetes management were explored, including durability, beta cell function, and a number of issues related to the link between diabetes and cardiovascular disease. Kahn highlighted the differences in recruitment and retention of subjects for the DPP and ADOPT studies. After screening 30,996 individuals by
OCR for page 78
Transforming Clinical Research in the United States: Challenges and Opportunities means of oral glucose tolerance testing, the DPP study randomized 3,234 to the three study arms. The remarkable feature of the DPP recruitment efforts was that 97.4 percent of the 3,234 participants completed the study. Even after 12 years in the DPP study, the retention rate was 88 percent. Kahn argued that this success was due in large part to the structure and design of the DPP study and the use of designated staff and budgeted resources for specific recruitment and retention efforts (see Figure 7-1). The investigators and staff at each of the 27 centers implementing the protocol felt invested in the study, according to Kahn, and the presence of a formal Recruitment and Retention Committee kept study monitors in constant contact with the centers. Any problems could be dealt with quickly through the network of committees overseeing the trial. Managing a relatively complex organizational structure by means of a 25-member Steering Committee and the Protocol Oversight Program, NIDDK was able to maintain tight control over the conduct of the trial and ensure compliance with the trial protocol. In contrast to the DPP, the ADOPT trial employed a relatively simple study management design. Of 6,676 individuals screened for the ADOPT trial 4,360 were randomized to the three study arms. Of the 4,360 who were recruited, only 60.3 percent completed the trial. The simple management structure (Figure 7-1) included the sponsor (GlaxoSmithKline), which worked with the DSMB; a nine-person Steering Committee; and an independent Adjudication Committee. Kahn suggested that, with no committees to oversee clinical operations, the investigators involved in the study may have been slightly less committed to the study than those involved in the DPP study. Moreover, the fact that ADOPT had 488 centers across 17 countries made it impossible to bring the 488 principal investigators and study coordinators together on a regular basis to discuss study progress. The number of subjects per research site also differed significantly between the DPP trial and ADOPT. The largest ADOPT center had 48 subjects enrolled in the trial, whereas the largest DPP site had 200 individuals enrolled (see Table 7-1). The reimbursement process for clinical trial staff is another key distinction between government- and industry-sponsored trials that can affect the quality of the research. The NIH approach to reimbursement provides financial support to full-time equivalent (FTE) trial staff. Kahn commented that this approach has contributed to the success of government-sponsored trials because it allows for the retention of trial staff and the appropriate number of study participant visits, even in long-term trials with two to three patient visits per year. The large dropout rate in ADOPT (40 percent) introduced potential bias into the study and could cast doubt on the significance of the differences among the three treatments. A rigorous sensitivity analysis by study staff, as well as statisticians independent of the study, determined that the
OCR for page 79
Transforming Clinical Research in the United States: Challenges and Opportunities FIGURE 7-1 Study management structure for a government-sponsored randomized controlled trial (Diabetes Prevention Program [DPP]) and an industry-sponsored randomized controlled trial (A Diabetes Progression Outcomes Trial [ADOPT]). SOURCE: Kahn, 2009. Reprinted with permission from Steven Kahn 2009.
OCR for page 80
Transforming Clinical Research in the United States: Challenges and Opportunities TABLE 7-1 Structure of Study Centers for a Government-Sponsored Randomized Controlled Trial (Diabetes Prevention Program [DPP]) and an Industry-Sponsored Randomized Controlled Trial (A Diabetes Progression Outcomes Trial [ADOPT]) DPP ADOPT No. of Countries 1 (USA) 17 (USA, Canada, Europe) No. of Centers 27 488 No. of Subjects 3,819 4,360 Subjects per Center 62–193 (Nat. American: 20–80) 1–48 Caucasian 55% 88% Reimbursement FTE based Visit based NOTE: FTE = full-time equivalent. SOURCE: Kahn, 2009. difference between the best drug (rosiglitazone) and the worst drug (glyburide) in the study was not attributable to bias and therefore still reliable. However, bias could not be ruled out as the cause of the observed difference between the best (rosiglitazone) and intermediate (metformin) drugs. In the DPP trial, the designation of specific staff and budgeting of resources for retaining participants were successful in achieving a 97 percent retention rate, thus avoiding bias in the study results. The DPP was likely more expensive than ADOPT in terms of cost per patient, according to Kahn. However, the 97 percent patient retention rate in the DPP was perhaps worth the additional cost given that large dropout rates can call into question the legitimacy of the results of any trial. The DPP could have been completed and found the same reduction in the risk of developing type 2 diabetes with a less intensive, less costly level of lifestyle intervention. In general, however, the results of NIH-sponsored, long-term studies such as the DPP, which have high rates of participant follow up, are often more valuable than those of industry-sponsored studies conducted over a short period of time and with dropout rates in the range of 20–25 percent. OVERCOMING REGULATORY CHALLENGES In addition to the challenges discussed in Chapter 3, Carla Greenbaum, Director of the Benaroya Research Institute Diabetes Program and Clinical Research Center, reflected on her experience conducting clinical trials