Idea Transcript
STATISTICS IN MEDICINE Statist. Med. 2002; 21:2779–2787 (DOI: 10.1002/sim.1281)
Clinical trials in the new millennium David L. DeMets∗ Department of Biostatistics; University of Wisconsin-Madison; 600 Highland Avenue; K6=446A CSC; Madison; Wisconsin 53792-4675; U.S.A.
SUMMARY Since the introduction of the randomized clinical trial (RCT) over 50 years ago, this method has become the corner stone for evaluating new pharmacologic or biologic agents with many disease areas beneting. There are numerous examples demonstrating benecial interventions as well as others either not benecial or harmful. Statistical methodology for clinical trials has grown rapidly. Advances in information technology for data collection have allowed trials to be conducted around the world. Academia, industry and government have worked in partnership to conduct RCTs. Despite these many RCT achievements, the most interesting and challenging era of clinical trials lies ahead of us. With the human genome now sequenced, we face a new set of challenges to transform vast amounts of data into useful information. The post-genomics era will better identify the disease mechanism and thus help design better treatments, and be more selective in screening patients, yielding more ecient clinical trials. For some areas of medicine, such as medical devices, standards of acceptance and regulatory approval are changing. Other areas, such as medical procedures and alternative medicines, are generally not well evaluated and could benet greatly from a wider use of RCT methodology. As the ICH guidelines facilitate and encourage international clinical trials, the scientic and ethical dimension of conducting trials in Third World countries are raised. For example, Western society’s best standard of care may not be available or aordable to these countries as the control Western investigators should not exploit patients in the Third World. Those and many other challenges face us in the decade ahead. It is truly an exciting time with new opportunities for the RCT to contribute to medicine and health care or prevention. Copyright ? 2002 John Wiley & Sons, Ltd. KEY WORDS:
randomized clinical trial; future trials; trial model; genomics in trials; developing countries
INTRODUCTION As we begin the new millennium, the randomized clinical trial (RCT) will become an even more important research tool as we evaluate new treatments and interventions in new areas of medicine at an ever-increasing pace. Furthermore, since the human genome is now sequenced, we are faced with megabytes of data that must be turned into useful information. The fact that the randomized clinical trial has an extremely bright future is in part due to the success it has had over the past three decades. This methodology has been utilized successfully in most major ∗ Correspondence
to: David L. DeMets, Department of Biostatistics; University of Wisconsin-Madison, 600 Highland Avenue, K6=446A CSC, Madison, Wisconsin 53792-4675, U.S.A.
Copyright ? 2002 John Wiley & Sons, Ltd.
2780
D. L. DEMETS
disease areas. Owing to regulatory agencies, most newly approved drugs have been tested with the RCT. In cardiology, for example, the RCT has been used to prove the benets of betablocker drugs in the treatment of patients suering a heart attack [1–3] or congestive heart failure [4–6], and statins for the reduction of high cholesterol [7]. Tamoxifen has been shown to be a possible prevention strategy for women at high risk for breast cancer [8] as well as an adjuvant therapy for women with breast cancer. Diabetic patients have benetted from RCTs demonstrating the clinical benets of tight control of glucose levels [9]. Ophthalmologists have steadily added new procedures or treatments such as photocoagulation and laser therapy to improve vision or slow the loss of visual acuity [10]. AZT was demonstrated to be eective through RCTs in AIDs patients [11]. In some cases, RCTs have shown that new promising treatments were, in fact, not benecial and in some cases even harmful [12–14]. Regardless of the outcome, a well conducted RCT adds valuable information and contributes to progress in health care. While we can describe an incredible list of successes and important contributions due to the RCT, we should never take the RCT for granted. The RCT competes continually with other research strategies which are used to promote the benets of new interventions. Nonrandomized control studies still abound, observational studies are numerous and databases are becoming more prominent, at least in the U.S., as a means of evaluating health care delivery. We are all, of course, aware of the limitations and pitfalls of these methods as the ultimate or denitive test [15, 16]. An instructive example of these limitations is provided by the role of beta-carotene in the prevention of cancer. Several epidemiologic or observational studies indicated an increased risk of lung cancer with lower levels of serum beta-carotene. The nutritional world began to promote the benets of beta-carotene, often as food supplements, especially in breakfast cereal. In the late 1980s, the National Institutes of Health (NIH) supported three major clinical trials evaluating the benets of beta-carotene supplements, at fairly high doses, in reducing lung cancer incidence and total cancer mortality. The rst of these, the Alpha-tocopherol Beta Carotene (ATBC) trial was a randomized, placebo-controlled, factorial trial of beta-carotene and alpha-tocopherol in 26000 Finnish male smokers [17]. When the trial was completed, those individuals receiving beta-carotene had a statistically signicant (P = 0:01) increased incidence of lung cancer and lung cancer mortality, contrary to the original hypothesis. The same results, essentially, were observed in a similar U.S. trial referred to as CARET [18]. What is even more remarkable is that in the placebo arm of each trial, the observational studies are veried. That is, individuals with low serum beta-carotene had increased risk of cancer, yet increasing serum beta-carotene made the risk worse. The third trial, the U.S. Physician’s Health Study (PHS) [19] which randomized 22000 male physicians with a low percentage of smokers, indicated no treatment eect, either positive or negative, Table I. New challenges. 1. 2. 3. 4. 5. 6. 7.
International clinical trials External forces New areas and standards Academic–industry partnership Genomics and surrogates RCTs in developing countries Training
Copyright ? 2002 John Wiley & Sons, Ltd.
Statist. Med. 2002; 21:2779–2787
CLINICAL TRIALS IN THE NEW MILLENNIUM
2781
with a relative risk of almost unity. Without these three well-conducted trials, we might never have appreciated the lack of benet of this popular nutritional supplement. As we begin the new millennium, we must build on our successes and failures, to meet new challenges and opportunities. Some of these challenges are listed in Table I. Even this short list presents a daunting challenge but yet an exciting opportunity to have an impact and make a dierence in health care worldwide. INTERNATIONAL TRIALS For many years, each country developed its own standards for regulatory approval of newly discovered drugs. An RCT used successfully in one country for new drug approval may not have been judged adequate in another, a frustrating and terribly inecient process in the drug development process. In the early 1990s the International Conference on Harmonization (ICH) began to explore making regulations for product approval more uniform across Europe, North America and Japan. Within a few years, by 1997, ICH guidelines began to emerge, covering topics in clinical trial design, management and analysis [20]. Documents focused on general clinical issues as well as statistical aspects. The process is still ongoing as additional guidelines are drafted and rened. This eort has resulted in more uniform standards for drug approval. The net result is that trials should be completed more rapidly, regulatory review and approval should be achieved in a more timely fashion, and patients will have faster access to well-proven eective treatments. However, the ICH concept needs to expand beyond the Europe, North America and Japan regulatory agencies. EXTERNAL FORCES Recent innovation in computer information technology will further accelerate the ability to conduct clinical trials worldwide. Fax and internet web-based data entry allow even remote clinical sites to participate. Protocols and manual of operations can be available on the web. Electronic signatures allow regulatory agencies to track data validity and accountability. The power of this information technology will dramatically change how trials are conducted operationally. As we have gained momentum on international standardization, RCTs are continually facing external forces who have a stake in the research enterprise. With recent problems in the safety of gene therapy, hearings in the U.S. Congress led to new policy statements by the U.S. Department of Health and Human Services (HHS), which released statements that all trials under the jurisdiction of the National Institutes of Health (NIH) and the Food and Drug Administration (FDA) must have a data monitoring plan [21]. Congressional interest in clinical trials was also stirred by the discovery of fraud in one clinic of a large breast cancer trial [22]. The U.S. Congress intervened in approving payment for mammography in women aged 40–50 years, despite the lack of benecial evidence [23]. We should expect continued review by political process. With the dramatic increase in RCT activity by the pharmaceutical, biotechnology and device industry, the investment community and special interest groups follow more closely the progress of individual trials. Clearly, the success of a new product in improving health care or risk will ultimately lead to a nancial prot for those who invested in the sponsor. Based Copyright ? 2002 John Wiley & Sons, Ltd.
Statist. Med. 2002; 21:2779–2787
2782
D. L. DEMETS
on personal observation, investment companies seek advice on how the progress of a trial, its recruitment and duration, might predict ultimate success. In fact, studies have been computer simulated using trial specic monitoring plans to gauge during a trial what failure to terminate early might predict overall success. Clearly, this activity, if taken to an extreme, could have a serious impact on a trial. Special interest groups can have a dramatic impact on the scientic process [24]. AIDS activities and women’s health groups have produced a new constituency that clinical trials must address, the patient advocacy groups. The AIDS epidemic aected, among others, a population who had the intellectual training, motivation and resources to learn and understand the disease process, the availability of treatments, the regulatory approval process and the fundamentals of the RCT. While at rst the AIDS activist groups were sceptical and critical of the RCT, they eventually became champions of the RCT as the best, fastest and most reliable process to discover new, safe and eective treatments for AIDS. Women’s groups have taken a very active interest in results of RCTs especially in breast cancer. We should expect more patient advocacy groups and in fact welcome them. While it may cause extra work, and perhaps some constraints, the clinical research process and RCTs will be better o and far more eective if our ultimate consumers are more fully involved.
NEW AREAS OF APPLICATION By the turn of the millennium, the RCT was well established as the method for evaluating new drugs and biologics. There are still major areas where much is yet to be done, or even started. For example, medical devices are a major enterprise in health care. Controversy surrounding the use of breast implants and other devices led to an FDA task force review and report chaired by Dr Robert Temple of the FDA’s drug review [25]. The report was critical of FDA standards for device approval. In the recent few years, the FDA has worked to steadily increase medical device standards to include RCTs and clinically relevant endpoints. The process will take several years for the impact to be fully realized. One immediate challenge is that there are over 8000 device companies registered in the U.S., most with less than 100 employees. Many of these small device companies do not have the nancial resources to conduct denitive trials. Rapid change in standards could force these companies to stop innovation and slow progress. Medical procedures, unless device related, come under no U.S. regulatory review. Thus, there is no uniform standard for a new procedure to become part of a medical practice. Recently, with the attention to health costs, U.S. heath care providers are challenging some procedures for evidence for eectiveness. One such example is lung volume reduction for patients with emphysema. The concept is that if non-functioning lung tissue is removed, the lung will more eciently use the remaining healthy tissue to promote gas exchange. However, before federal providers routinely paid for this procedure, they asked for a denitive trial. The National Emphysema Treatment Trial (NETT) is underway to test this surgical procedure [26]. A 1993 article [27] described that the U.S. population spends more money on alternative medicines that on traditional drugs and devices. Most of these have not been evaluated in any rigorous fashion. The NIH has established an Oce on Alternative Medicines to evaluate those popular approaches. Given the large number of those elements, strategic choices will have to be made since federal funding is still limited [28]. Copyright ? 2002 John Wiley & Sons, Ltd.
Statist. Med. 2002; 21:2779–2787
CLINICAL TRIALS IN THE NEW MILLENNIUM
Funding Agency
Policy Advisory Board Data Monitoring Committee
Data Center
2783
Steering Committee
Clinics
Central Labs
Working Committees
Figure 1. NHLBI CT model.
Executive Committee
Sponsor
DSMC
Statistical Data Analysis Center
CRO
Central Labs
Central Review Committees
Clinical Centers
Figure 2. Industry modied NIH model.
ACADEMIA–INDUSTRY PARTNERSHIPS While funding for NIH supported clinical trials has increased signicantly overall during the past three decades, the past ten years have been a period of relatively stable funding. The only major exception in the U.S. is for AIDS research and women’s health. Fortunately, the pharmaceutical industry has expanded its activity dramatically during this same period. For example, in cardiology many of the leading clinical trials presented at national or international meetings are industry sponsored. For three decades, academia has developed a research partnership with federal sponsors but now over the past ten years a new partnership between academia and industry has also emerged [29]. One RCT model successfully used by NIH is illustrated in Figure 1 where the major components of a trial are shown. Key components, besides the sponsor, are the Steering Committee, the independent Data and Safety Monitoring Board and the independent Data Co-ordinating Centre. This model has served NIH and academia well and continues to be successful. A critical factor is academic independence and access to data after trial completion for scholarly activity. As models for academic partnering with industry are explored, as much of the successful NIH-type model should be preserved as possible. One such model is shown in Figure 2. The key modication is that the traditional data co-ordinating centre has been split into two pieces, data management and statistical analysis [30]. Data management may be carried on Copyright ? 2002 John Wiley & Sons, Ltd.
Statist. Med. 2002; 21:2779–2787
2784
D. L. DEMETS
by the sponsor internally or externally through a contract research organization (CRO) of their choice. The second component, statistical analysis, can in principle be done internally by the sponsor or externally with an independent group such as an academic statistical unit. There are several reasons for an external data analysis centre. One major reason is to support the independent external Data and Safety Monitoring Board (DSMB), a component common to both models and critical in each. This independent statistical centre allows the DSMB to better protect the integrity of the trial and provides investigators with better access to the data once a trial is completed. If a trial is negative or neutral, industry tends to quickly reassign sta to another project, with a higher priority. For positive trials, industry statisticians become committed to preparing the regulatory application. While this is a strategic realignment from the sponsor’s perspective, it often makes investigator access to data dicult if not impossible. While this model, and variations of it, may need further modication, it does preserve the best of the NIH-type model.
APPROPRIATE OUTCOME MEASURES There are a few major challenges ahead of us that have not yet been adequately addressed. The issues surrounding surrogate outcome measures have been discussed, including many examples where their use has failed [31, 32]. However, genomics will challenge this issue even further. A surrogate, a laboratory substitute for a clinically relevant outcome, must not only predict the clinical outcome but also capture the full treatment eect [33]. A classic example of surrogate failure is the Cardiac Arrhythmia Suppression Trial (CAST). Ventricular arrhythmias are established predictors of sudden death. Drugs known to suppress these arrhythmias were tested in CAST and were found to increase, not decrease, the risk of death and sudden death [34]. CAST was designed to randomize 1455 patients with three years of follow-up to detect a 30 per cent reduction in mortality. However, when less than 15 per cent of the observed deaths had occurred, the trial was terminated with results as shown in Table II. Many other such examples have been previously described where surrogates failed [31]. With genomics and gene chip technology, we will soon have unparalleled volumes of data. The rst goal is to transform this data into useful information. Current gene chip technology can generate 160 MB of data per visit. If a baseline plus 1-2 follow-up ‘samples’ were taken to track gene therapy, we could generate 480 MB of data per patient. The temptation would be to rely on a genomic surrogate to evaluate whether the gene therapy corrected the genetic aw. However, as before, we must also know that the gene therapy has not altered anything else. We will need to be careful in our evaluation and use of this potentially powerful method. Initially it seems unlikely that we should use only genomic data to evaluate new therapies Table II. Cardiac Arrhythmia Suppression Trial (CAST): early termination in two drug arms.
Sudden death Total mortality *Encanide=ecanide
Copyright ? 2002 John Wiley & Sons, Ltd.
Drugs
Placebo
33 56
9 22
z¿3:2
Statist. Med. 2002; 21:2779–2787
CLINICAL TRIALS IN THE NEW MILLENNIUM
2785
with no clinical follow-up. Rather, it is more likely that we will be able to use this new information to better target drug and gene therapy, screen new potential treatments faster and be more selective in patients entering a trial.
TRIALS IN DEVELOPING COUNTRIES In parallel with ICH, many multinational clinical trials are now entering patients from eastern Europe, India, China and Third World countries. These countries represent a large patient population pool. CRO networks are forming in many of these regions to gain access to this population. In addition to obvious logistical challenges in data collection, there is wide variation in the population risk, background therapy and culture. Several ethical issues present themselves. The rst is the issue of informed consent. Some countries do not have Institutional Review Boards and consent of patients is not part of the culture. Rather, patients defer to their physician. A second issue has to do with investigator conict. If a trial is asking a question that is truly multinational, then a broad ethnic representation might be highly desirable. On the other hand, if the trial is being done by a western investigator or company in a foreign country because they cannot recruit patients otherwise, there may be an ethical conict. Two papers challenged the use of placebo controls for AIDS treatment in Third World countries and the use of best local therapy as a control if it is inferior to best existing therapy [35, 36]. Some of these issues were debated in a trial conducted by an investigator at the University of Wisconsin. A breast cancer trial in Vietnam was designed and funded by the NIH to ask if a simple adjuvant therapy of oophorectomy plus tamoxifen was superior to the local standard of care following observation. Best existing U.S. standard of care including radiation plus complex chemotherapy was simply not aordable or even available in Vietnam [37]. For Vietnam, the investigator argued that this simple aordable adjuvant therapy would be an improvement in their standard of clinical care. In addition, a relevant scientic question on the role of tamoxifen in an Asian population could also be addressed. These types of issues must be addressed before a trial can proceed.
TRAINING THE NEXT GENERATION A nal major issue to be discussed is the issue of training the next generation of clinical triallists, both statisticians and clinicians. The rate of new PhD trained statisticians has not increased for the past two decades while the number of positions has increased dramatically [38]. The need to mine the genomics data will increase the demand even more, far beyond the current supply. Academic clinicians in the U.S. are facing tremendous pressures to see more patients and generate additional clinical revenue. Few young academic clinicians have the time to conduct clinical trials, and even fewer have the formal training. Given the pace of clinical trials and their importance, a systematic national and international plan is needed with adequate funding over the next decade so that the next generation of clinical triallists will be available to meet the challenges. Currently, there is not an adequate number of courses, curriculum or even incentive to deliver such training. Copyright ? 2002 John Wiley & Sons, Ltd.
Statist. Med. 2002; 21:2779–2787
2786
D. L. DEMETS
SUMMARY AND CONCLUSIONS While there are numerous challenges to face in the next decade, they are largely addressable. If adequate resources and planning are made available, the RCT can have an even larger impact on health care than it already has enjoyed. As the new millennium begins, the many successes and numerous lessons in clinical trials from the past three decades provide substantial momentum to carry us forward. While there are numerous challenges to face just in the next decade, they are largely addressable with adequate training, planning and resources. The best and most exciting years for the randomized clinical trial lay ahead of us. With proper vision, the randomized clinical trial methodology can have an even larger impact on health care than it already has. Clinical trial methodology can in fact be expanded beyond health care to aect areas as diverse as education, manufacturing and correctional strategies or programmes. It is indeed a powerful methodology for evaluation. REFERENCES 1. Norwegian Multicenter Study Group. Timolol-induced reduction in mortality and reinfarction in patients surviving acute myocardial infarction. New England Journal of Medicine 1981; 304:801– 807. 2. Hjalmarson A, Herlitz J, Malek I, Ryden L, Vedin A, Waldenstrom A, Wedel H, Elmfeldt D, Holmberg S, Nyberg G, Swedberg K, Waagstein F, Waldenstrom J, Wilhelmsen L, Wilhelmson C. Eect on mortality of metoprolol in acute myocardial infarction: a double-blind randomized trial. Lancet 1981; II:823– 827. 3. Beta-Blocker Heart Attack Trial Research Group. A randomized trial of propranolol in patients with acute myocardial infarction. I. Mortality results. Journal of the American Medical Association 1982; 247(12): 1707– 1714. 4. CIBIS-II Investigators and Committees. The cardiac insuciency Bisoprolol study II (CIBIS-II): a randomised trial. Lancet 1999; 353:9–13. 5. MERIT-HF Study Group. Eect of metoprolol CR=XL in chronic heart failure: Metoprolol CR=XL randomised intervention trial in congestive heart failure. Lancet 1999; 353:2001– 2007. 6. Packer M, Bristow MR, Cohn JN, Colucci WS, Fowler MB, Gilbert EM, Shusterman NH, for the U.S. Carvedilol Heart Failure Study Group. The eect of carvedilol on morbidity and mortality in patients with chronic heart failure. New England Journal of Medicine 1996; 334(21):1349–1355. 7. Scandinavian Simvastatin Survival Study Group. Randomized trial of cholesterol lowering in 4444 patients with coronary heart disease: The Scandinavian Simvastatin Survival Study (4S). Lancet 1994; 344:1383–1389. 8. Fisher B, Costantino J, Wickerham DL, Redmond CK, Kavanah M, Cronin W, Vogel V, Robidoux A, Dimitrov N, Atkins J, Daly M, Wieand S, Tan-chiu E, Ford L, Wolmark N and other National Surgical Adjuvant Breast and Bowel Project investigators. Tamoxifen for prevention of breast cancer: report of the National Surgical Adjuvant Breast and Bowel Project P-1 study. Journal of the National Cancer Institute 1998; 90(18): 1371–1388. 9. Diabetes Control and Complications Trial Research Group. The eect of intensive treatment of diabetes on the development and progression of long-term complications in insulin-dependent diabetes mellitus. New England Journal of Medicine 1993; 329(14):977– 986. 10. Diabetic Retinopathy Study Research Group. Photocoagulation treatment of proliferative diabetic retinopathy: clinical application of Diabetic Retinopathy Study (DRS) ndings, DRS report number 8. Opthalmology 1981; 88(7):583– 600. 11. Fischl MA, Richman DD, Grieco MH, Gottlieb MS, Volberding PA, Laskin OL, Leedom JM, Groopman JE, Mildvan D, Schooley RT, Jackson GG, Durack DT, King D and the AZT Collaborative Working Group. The ecacy of azidothymidine (AZT) in the treatment of patients with AIDS and AIDS-related complex: a doubleblind, placebo-controlled trial. New England Journal of Medicine 1987; 317(4):185 –191. 12. Echt DS, Liebson PR, Mitchell LB, Peters RW, Obias-Manno D, Barker AH, Arensberg D, Baker A, Richardson DW, and the CAST Investigators. Mortality and morbidity in patients receiving encainide, ecanide, or placebo. New England Journal of Medicine 1991; 324:781– 788. 13. Cohn JN, Goldstein SO, Greenberg BH, Lorell BH, Bourge RC, Jaski BE, Gottlieb SO, McGrew F, DeMets DL, White BG, for the Vesnarinone Trial Investigators. A dose-dependent increase in mortality with vesnarinone among patients with severe heart failure. New England Journal of Medicine 1998; 339(25):1810 –1816. 14. Packer M, Carver JR, Rodeheer RJ, Ivanhoe RJ, DiBianco R, Zeldis SM, Hendrix GH, Bommer WJ, Elkayam U, Kukin ML, Mallis GI, Sollano JA, Shannon J, Tandon PK, and DeMets DL for the PROMISE Copyright ? 2002 John Wiley & Sons, Ltd.
Statist. Med. 2002; 21:2779–2787
CLINICAL TRIALS IN THE NEW MILLENNIUM
15. 16. 17. 18. 19.
20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38.
2787
Study Research Group. Eect of oral milrinone on mortality in severe chronic heart failure. New England Journal of Medicine 1991; 325(21):1468 –1475. Taubes G. Epidemiology faces its limits. Science 1995; 269:164 –169. Byar DP, Simon RM, Friedewald WT, Schlesselman JJ, DeMeis DL, Ellenberg JH, Gail MH and Ware JH. Randomized clinical trials: perspectives on some recent ideas. New England Journal of Medicine 1976; 295: 74 – 80. Alpha-tocopherol, Beta Carotene Cancer Prevention Study Group. The eect of vitamin E and beta carotene on the incidence of lung cancer and other cancers in male smokers. New England Journal of Medicine 1994; 330(15):1029 –1035. Omenn GS, Goodman GE, Thornquist MD, Balmes J, Cullen MR, Glass A, Keogh JP, Meyskens FL, Valanis B, Williams JH, Barnhart S, Hammar S. Eects of a combination of beta carotene and vitamin A on lung cancer and cardiovascular disease. New England Journal of Medicine 1996; 334(18):1150 –1155. Hennekens CH, Buring JE, Manson JE, Stampfer M, Rosner B, Cook N, Belanger C, LaMotte F, Gaziano JM, Ridker PM, Willett W, Peto R. Lack of eect of long term supplementation with beta carotene on the incidence of malignant neoplasms and cardiovascular disease. New England Journal of Medicine 1996; 334(18): 1145 –1149. ICH Expert Working Group. International Conference on Harmonisation of Technical Requirements for Registration of Pharmaceuticals for Human Use. ICH Harmonised Tripartite Guideline. Statistical principles for clinical trials. Statistics in Medicine 1999; 18:1905 –1942. U.S. Department of Health and Human Services. Secretary Shalala bolsters protections for human research subjects. May 23, 2000 at www.os.dhhs.gov=news=press=. Kassirer JP. Practicing medicine without a license – the new intrusions by Congress. New England Journal of Medicine 1997; 336(24):1747. Fletcher SW. Whither scientic deliberation in health policy recommendations? Alice in the Wonderland of breast-cancer screening. New England Journal of Medicine 1997; 336(16):1180 –1183. Deyo RA, Psaty BM, Simon G, Wagner EH, Omenn GS. The messenger under attack-intimidation of researchers by special-interest groups. New England Journal of Medicine 1997; 336(16):1176 –1180. Final Report of the Committee for Clinical Review (based on a review of selected medical device applications). Food and Drug Administration, March 1993 [Temple Report]. The National Emphysema Treatment Trial Research Group. Rationale and design of the National Emphysema Treatment Trial (NETT): a prospective randomized trial of lung volume reduction surgery. Journal of Thoracic Cardiovascular Surgery 1999; 118:518 – 528. Eisenberg DM, Kessler RC, Foster C, Norlock FE, Calkins DR, Delbanco TL. Unconventional medicine in the United States – Prevalence, costs and patters of use. New England Journal of Medicine 1993; 328(4): 246 – 252. Angell M, Kassirer JP. Alternative Medicine – The risks of untested and unregulated remedies. Editorial. New England Journal of Medicine 1998; 339(12):839– 841. Blumenthal D, Campbell EG, Causina N, Louis KS. Participation of life-science faculty in research relationships with industry. New England Journal of Medicine 1996; 335(23):1734 –1739. Fisher M, Roecker EB, DeMets D. The role of an independent statistical analysis center in the industry-modied National Insititutes of Health model. Drug Information Journal 2001; 35(1):115–129. Fleming TR, DeMets DL. Surrogate endpoints in clinical trials: are we being misled? Annals of Internal Medicine 1996; 125(7):605 – 613. Temple RJ. A regulatory authority’s opinion about surrogate endpoints. Drug Information Agency meeting, Hilton Head, 1993. Prentice RL. Surrogate endpoints in clinical trials: denition and operational criteria. Statistics in Medicine 1989; 8:431– 440. Cardiac Arrhythmia Suppression Trial (CAST) Investigators. Preliminary report: Eect of encainide and ecainide on mortality in a randomized trial of arrhythmia suppression after myocardial infarction. New England Journal of Medicine 1989; 321(6):406– 412. Lurie P, Wolfe SM. Unethical trials of interventions to reduce perinatal transmission of the human immunodeciency virus in developing countries. New England Journal of Medicine 1997; 337(12):853– 856. Angell M. The ethics of clinical research in the Third World. Editorial. New England Journal of Medicine 1997; 337(12):847– 849. Love RR, Fost CN. Ethical and regulatory challenges in a randomized control trial of adjuvant treatment for breast cancer in Vietnam. [Review] Journal of Investigative Medicine 1997; 45(8):423– 431. DeMets DL, Woolson R, Brooks C, Qu R. Where the jobs are: a study of AMSTAT news job advertisements. American Statistician 1998; 52(4):303–307.
Copyright ? 2002 John Wiley & Sons, Ltd.
Statist. Med. 2002; 21:2779–2787