SEARCH

SEARCH BY CITATION

Keywords:

  • benefit assessment;
  • clinical trials;
  • cost-effectiveness;
  • evidence-based medicine;
  • risk assessment;
  • therapeutics

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Treatment effects are modest
  5. Quantitative interactions are common, qualitative interactions are rare (so far)
  6. Unintended effects are common because surrogates are treacherous
  7. Long-term and short-term effects are not necessarily the same
  8. Combinations are unpredictable
  9. Class effect cannot be taken for granted
  10. Effective therapies produce a combination of benefits and risks
  11. Most beneficial therapies do not save money, but produce an incremental benefit for incremental cost
  12. Applying the results of clinical trials is beneficial
  13. Some areas of medicine are underserved
  14. Participation is critical
  15. Conflict of interest statement
  16. References
  17. Appendices

Abstract. Califf RM (Duke University Medical Center, Durham, NC, USA). Issues facing clinical trials of the future (Clinical Trials). J Intern Med 2003; 254: 426–433.

Diagnostic and therapeutic technology continues to advance rapidly, as does our knowledge of therapeutics and of clinical research methods. Unfortunately, these advances have been only poorly coupled with our knowledge of therapeutic principles, leading to increasing uncertainty about which technologies are truly effective and, amongst those that are effective, which are most effective for the cost. This article presents general principles derived from several investigators’ experiences with clinical trials, and uses them to suggest how future clinical trials may differ from current approaches. A proposed organization for future trials also is elucidated.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Treatment effects are modest
  5. Quantitative interactions are common, qualitative interactions are rare (so far)
  6. Unintended effects are common because surrogates are treacherous
  7. Long-term and short-term effects are not necessarily the same
  8. Combinations are unpredictable
  9. Class effect cannot be taken for granted
  10. Effective therapies produce a combination of benefits and risks
  11. Most beneficial therapies do not save money, but produce an incremental benefit for incremental cost
  12. Applying the results of clinical trials is beneficial
  13. Some areas of medicine are underserved
  14. Participation is critical
  15. Conflict of interest statement
  16. References
  17. Appendices

Diagnostic and therapeutic technology is continuing to advance at a rapid pace. At the same time, our knowledge of therapeutics and of clinical research methods also is advancing. Unfortunately, the tremendous technological success of the past several decades has been only poorly coupled with our knowledge of therapeutic principles. The result has been increasing uncertainty about which technologies are truly effective and, amongst those that are effective, which are most effective for the cost. Therefore, as we experience unparalleled medical advances, we also face a future in which the dissociation between our increasing understanding of complex biology and an only rudimentary understanding of clinical evidence and therapeutics impedes our ability to deliver benefits to patients.

The fruits of modern technology include rapid improvements in function and longevity in the population with the concomitant burden of chronic disease. The growing elderly population is more highly educated than past elderly populations and expects a higher level of medical care. At the same time, the dramatic change in the ratio of workers producing resources to pay for medical care to those consuming such resources has not been anticipated. In the United States alone, this ratio will decline from 4 : 1 to 2 : 1 over the next 30 years [1], with even a larger decline anticipated for Europe.

We also have continued expansion of biological and technological targets for diagnostic and therapeutic technologies, fuelled by our investment in basic biomedical research. These technologies will work, and people will live even longer and have a better quality of life. Along the way, however, the power of new technology will carry increasing risk during the period of early development.

As the elderly population grows, we also will have increasingly sophisticated technology that can improve the plight of children and young people with complex disease. These technologies will be able to produce dramatic benefit, but they also will carry significant risk if used inappropriately.

All of these factors conspire to demand a different approach to clinical research in the future. A recent series of articles reviewed general principles derived from several investigators’ experiences with clinical trials (see Appendix) [2, 3]. This article will use those principles to make suggestions about how clinical trials of the future may differ from current approaches. Finally, a view of the organization of the future for clinical trials will be elucidated.

Treatment effects are modest

  1. Top of page
  2. Abstract
  3. Introduction
  4. Treatment effects are modest
  5. Quantitative interactions are common, qualitative interactions are rare (so far)
  6. Unintended effects are common because surrogates are treacherous
  7. Long-term and short-term effects are not necessarily the same
  8. Combinations are unpredictable
  9. Class effect cannot be taken for granted
  10. Effective therapies produce a combination of benefits and risks
  11. Most beneficial therapies do not save money, but produce an incremental benefit for incremental cost
  12. Applying the results of clinical trials is beneficial
  13. Some areas of medicine are underserved
  14. Participation is critical
  15. Conflict of interest statement
  16. References
  17. Appendices

Most therapies are not curative. Rather, if used appropriately, each adds incrementally to benefit. In most areas of therapeutics, relative improvements in the range of 25% are considered outstanding, and for survival, benefits of 10–15% on a relative scale are considered worthwhile.

If treatment effects are modest, then the size of trials must be large to protect against type II error. Type II errors occur when a trial is not large enough to detect the true effect of treatment, and the study erroneously concludes that there is no difference between the treatment and control groups. In the past [2, 3], most studies suffered from the type II error, and even though this issue has been effectively addressed for randomized trials, many ‘negative’ observational studies continue to have this problem.

To conduct large trials, extensive networks of investigators are needed, and they must extend well beyond academic centres into clinical practice. Accordingly, it will be imperative in the future for clinical trials to include vast numbers of practitioners, organized into networks so that trials can be performed rapidly, hastening advancement of biological and clinical knowledge.

Hopefully, these networks can be expanded to include patients or people at risk of developing particular diseases. Several diseases characterized by activist patient-advocacy organizations have already developed systems to channel patients into clinical trials [4, 5]. In the future, the Internet will offer an efficient mechanism to allow people with particular diseases to volunteer for participation in clinical trials.

Quantitative interactions are common, qualitative interactions are rare (so far)

  1. Top of page
  2. Abstract
  3. Introduction
  4. Treatment effects are modest
  5. Quantitative interactions are common, qualitative interactions are rare (so far)
  6. Unintended effects are common because surrogates are treacherous
  7. Long-term and short-term effects are not necessarily the same
  8. Combinations are unpredictable
  9. Class effect cannot be taken for granted
  10. Effective therapies produce a combination of benefits and risks
  11. Most beneficial therapies do not save money, but produce an incremental benefit for incremental cost
  12. Applying the results of clinical trials is beneficial
  13. Some areas of medicine are underserved
  14. Participation is critical
  15. Conflict of interest statement
  16. References
  17. Appendices

When the mass of clinical trials is examined, it is unusual to find a subgroup for whom the treatment is beneficial if the overall treatment effect is negative, or for whom the treatment is harmful if the overall effect is beneficial (qualitative interaction). However, quantitative interactions are common; the magnitude of the benefit or harm usually varies as a function of baseline characteristics.

These principles point to the importance of measuring basic patient characteristics using standard nomenclature. The need for standard nomenclature and data standards is even more pressing in clinical practice. The failure of the medical professions to develop and implement data standards has left medicine far behind most other major enterprises. Anyone with a bank account can withdraw cash almost anywhere in the world; yet for a critically ill patient it is difficult to gain access to the medical record. Even if access is gained, the lack of standardization creates a medical ‘Tower of Babel’ in which patients cannot be sure that the descriptions of illness in the medical record are interpreted consistently from one physician to another.

Quantitative interactions can be critical in the consideration of cost-effectiveness of therapeutics. Particularly as the therapeutic arena becomes more crowded, and as therapy becomes more long-term, the relative value of a therapy compared with other options will be an increasingly critical consideration. Consider the cost to society of taking statin drugs for a lifetime, in a future in which children are identified as being at risk for atherosclerosis.

Rapid advances in our understanding of genomic, proteomic and metabolic information likely will revolutionize our understanding of the biological activity of therapies. A small but increasing number of biological modifiers of therapeutic response have been identified, and the number should grow rapidly over the next decade. Although most modifiers are likely to produce more refined quantitative interactions (more or less benefit or harm), a certain proportion truly will segregate populations into those with a favourable or unfavourable balance of risk and benefit.

Much of the ‘hype’ about pharmacogenomics has reflected the belief that clinical trials could be performed in only a small number of patients. This approach is unlikely to succeed, however. The primary reason is the complexity of biology, such that any single measure of a therapeutic effect on a biological pathway is unlikely to yield a definitive picture of the ultimate benefit and risk of treatment. Additionally, the mathematics of ensuring that a treatment interaction is true are daunting, especially considering that many of these interactions will involve multiple genes or proteins with a graded effect on the biological measures of interest.

To understand this issue, consider a genetic variation (polymorphism) determined to produce aspirin resistance. If the polymorphism reduces the protective effect of aspirin on vascular events by 50% (that is, from a 25% reduction to a 12.5% reduction), it would take over 30 000 patients to detect this effect; then, to ensure that the finding was not a false positive, the study would need to be replicated.

Unintended effects are common because surrogates are treacherous

  1. Top of page
  2. Abstract
  3. Introduction
  4. Treatment effects are modest
  5. Quantitative interactions are common, qualitative interactions are rare (so far)
  6. Unintended effects are common because surrogates are treacherous
  7. Long-term and short-term effects are not necessarily the same
  8. Combinations are unpredictable
  9. Class effect cannot be taken for granted
  10. Effective therapies produce a combination of benefits and risks
  11. Most beneficial therapies do not save money, but produce an incremental benefit for incremental cost
  12. Applying the results of clinical trials is beneficial
  13. Some areas of medicine are underserved
  14. Participation is critical
  15. Conflict of interest statement
  16. References
  17. Appendices

Biomarkers are biological measurements that reflect the disease process, and usually they correlate with ultimate clinical outcomes. Many individual biomarkers have mistakenly been positioned as surrogate end-points, or ‘stand-ins’ for the true outcomes of interest. For a putative surrogate to meet the criteria to allow its acceptance with confidence, the treatment effect on the biomarker not only must correlate with the effect on the true clinical outcome but also must explain a large portion of the treatment effect [6]. To make matters even more complex, the modern concept of therapeutics as representing a positive balance of risk and benefit in multiple dimensions complicates any discussion of surrogates.

Future trials will need to use biomarkers more effectively to sort through the myriad of available biological targets. The appropriate use of biomarkers almost certainly will not be a single measure but instead will require the integration of an array of biomarkers. Even then, promising therapies must be subjected to the true test of comparative clinical end-points because, ultimately, there is no substitute for measuring such outcomes.

Long-term and short-term effects are not necessarily the same

  1. Top of page
  2. Abstract
  3. Introduction
  4. Treatment effects are modest
  5. Quantitative interactions are common, qualitative interactions are rare (so far)
  6. Unintended effects are common because surrogates are treacherous
  7. Long-term and short-term effects are not necessarily the same
  8. Combinations are unpredictable
  9. Class effect cannot be taken for granted
  10. Effective therapies produce a combination of benefits and risks
  11. Most beneficial therapies do not save money, but produce an incremental benefit for incremental cost
  12. Applying the results of clinical trials is beneficial
  13. Some areas of medicine are underserved
  14. Participation is critical
  15. Conflict of interest statement
  16. References
  17. Appendices

Most clinical trials have been based on the measurement of intermediate outcomes in a short time frame. The concept of early hazard permeates medicine, as most surgical procedures are associated with a short-term risk of death or complications in return for a later benefit. As we face earlier identification of disease risk, however, we also face many years of treatment with preventive therapies with unknown long-term risk (or benefit). Future clinical trials will need to measure outcomes for the relevant period of treatment.

To deal with this complexity, new methods of long-term follow-up will need to be devised. In many countries, national health registries offer an opportunity to capitalize on long-term outcome data at a very low cost. The enactment of the Health Insurance Portability and Accountability Act (HIPAA) in the United States [7] has deterred efforts to develop long-term follow-up systems because of extreme concern about patient privacy. Violation of the rules of this act can result in criminal penalties for the health care provider, and this has produced a chilling effect on excitement about participating in clinical research.

The Internet again is likely to offer the most accurate and least expensive means of long-term follow-up for patients in clinical trials. Although many groups now are experimenting with Internet systems that bypass the patient's doctor, a more effective system likely will include both the doctor and the patient so that the clinical research effort is not seen as an impediment to the delivery of personalized medical care.

Combinations are unpredictable

  1. Top of page
  2. Abstract
  3. Introduction
  4. Treatment effects are modest
  5. Quantitative interactions are common, qualitative interactions are rare (so far)
  6. Unintended effects are common because surrogates are treacherous
  7. Long-term and short-term effects are not necessarily the same
  8. Combinations are unpredictable
  9. Class effect cannot be taken for granted
  10. Effective therapies produce a combination of benefits and risks
  11. Most beneficial therapies do not save money, but produce an incremental benefit for incremental cost
  12. Applying the results of clinical trials is beneficial
  13. Some areas of medicine are underserved
  14. Participation is critical
  15. Conflict of interest statement
  16. References
  17. Appendices

When several therapies are combined, each of which may be beneficial alone, the result is not easily predictable. The combination could be synergistic (more than additive), additive, subadditive (better than either alone, but not additive), equal to either alone, or worse than either single drug. This effect could pertain to a biomarker or any element of benefit or risk. When amlodipine and benazepril were combined into a single pill, for example, the side-effect profile looked better than with either pill alone, and the effect on blood pressure was roughly additive [8]. On the other hand, toxicity when combined with other drugs forced mibefradil to be withdrawn from the market, although it was a proven antihypertensive with a good risk profile in well-controlled (not real-life) clinical trials [9].

Clinical trials of the future must develop an approach to sorting through these complex combinations. Again, the only imaginable approach will require large numbers of patients to sort through the factorial combinations. The Antihypertensive and Lipid-Lowering treatment to prevent Heart Attack Trial (ALLHAT) [10] gives some insight into the complexity that will be encountered. It took more than 40 000 patients to determine that a thiazide diuretic was a better choice for the first antihypertensive agent compared with an angiotensin-converting enzyme (ACE) inhibitor, an α-receptor antagonist, or a calcium-channel blocker. Critics of the trial point out that when an ACE inhibitor is used first, the second drug used most often is a diuretic; this combination was prohibited by the trial, given that the goal in step 2 of treatment was to add a drug not used in step 1. Thus, it is entirely plausible that if one is to use only one drug, a diuretic is the best choice, but if two drugs are needed, an ACE inhibitor combined with the diuretic would be the best choice. Perhaps the failure to use this combination made the ACE inhibitor look bad overall.

Class effect cannot be taken for granted

  1. Top of page
  2. Abstract
  3. Introduction
  4. Treatment effects are modest
  5. Quantitative interactions are common, qualitative interactions are rare (so far)
  6. Unintended effects are common because surrogates are treacherous
  7. Long-term and short-term effects are not necessarily the same
  8. Combinations are unpredictable
  9. Class effect cannot be taken for granted
  10. Effective therapies produce a combination of benefits and risks
  11. Most beneficial therapies do not save money, but produce an incremental benefit for incremental cost
  12. Applying the results of clinical trials is beneficial
  13. Some areas of medicine are underserved
  14. Participation is critical
  15. Conflict of interest statement
  16. References
  17. Appendices

The term ‘class effect’ has been poorly characterized. Industry often has used it to ‘piggyback’ a second compound on top of a successful similar compound attacking a common biological target. This has been convenient for the clinical community, because it avoided the ‘messy’ issue of distinguishing effects of one compound from another.

The equivalent of the class effect in the device field is modification of a ‘template’ device. The drug-eluting stent with rapamycin as the drug, once proven, will undergo numerous minor modifications to improve the polymer and stent. It is reasonable to accept such modifications without requiring a new set of definitive clinical trials. At some point, however, the modification exceeds a reasonable degree of certainty that the clinical results would be the same. Similarly, in behavioural therapeutics, multiple versions of the same basic intervention may yield different results depending on their delivery methods.

Unfortunately, the only way to determine which of several similar therapies is superior is to compare them in head-to-head clinical trials with objective measurement of end-points. This type of trial is not feasible, except in rare circumstances, because of the large costs required to detect small differences. These trials will be possible only if routine clinical data can be used, rather than creating an expensive new mechanism for data collection every time a new trial is done.

Effective therapies produce a combination of benefits and risks

  1. Top of page
  2. Abstract
  3. Introduction
  4. Treatment effects are modest
  5. Quantitative interactions are common, qualitative interactions are rare (so far)
  6. Unintended effects are common because surrogates are treacherous
  7. Long-term and short-term effects are not necessarily the same
  8. Combinations are unpredictable
  9. Class effect cannot be taken for granted
  10. Effective therapies produce a combination of benefits and risks
  11. Most beneficial therapies do not save money, but produce an incremental benefit for incremental cost
  12. Applying the results of clinical trials is beneficial
  13. Some areas of medicine are underserved
  14. Participation is critical
  15. Conflict of interest statement
  16. References
  17. Appendices

As described above, a therapy is desired because of the complex balance of risk and benefit. Because this balance depends not only on the intrinsic characteristics of the therapy but also on (i) other therapies being ingested, (ii) patient characteristics, and (iii) how the therapy is delivered in the healthcare system, it is critical to measure real clinical outcomes of interest in the actual population at risk in as realistic an environment as possible.

Because of the increasingly recognized complexity of therapeutics, the balance of risk and benefit can be improved either by improving efficacy or by decreasing toxicity or costs. Thus, trials of therapies with similar benefit and noninferiority trials will become more important. Such trials create several difficult and complex issues about the determination of meaningful clinical differences and quantitative concepts that are unfamiliar to most clinicians.

Most beneficial therapies do not save money, but produce an incremental benefit for incremental cost

  1. Top of page
  2. Abstract
  3. Introduction
  4. Treatment effects are modest
  5. Quantitative interactions are common, qualitative interactions are rare (so far)
  6. Unintended effects are common because surrogates are treacherous
  7. Long-term and short-term effects are not necessarily the same
  8. Combinations are unpredictable
  9. Class effect cannot be taken for granted
  10. Effective therapies produce a combination of benefits and risks
  11. Most beneficial therapies do not save money, but produce an incremental benefit for incremental cost
  12. Applying the results of clinical trials is beneficial
  13. Some areas of medicine are underserved
  14. Participation is critical
  15. Conflict of interest statement
  16. References
  17. Appendices

The idea that we pay more for a greater benefit is basic to most national economies. Yet in healthcare, there has been a tendency to expect that more effective therapies should save money. In fact, this is rarely the case. Even when preventive therapies reduce short-term costs, long-term costs may increase as people live longer and use more healthcare resources overall.

The importance of comparative economics has been identified as a major need by the Clinical Research Roundtable [11]. This need can be met only by large trials that measure clinical outcomes over relevant time periods in addition to measuring healthcare resource use.

Applying the results of clinical trials is beneficial

  1. Top of page
  2. Abstract
  3. Introduction
  4. Treatment effects are modest
  5. Quantitative interactions are common, qualitative interactions are rare (so far)
  6. Unintended effects are common because surrogates are treacherous
  7. Long-term and short-term effects are not necessarily the same
  8. Combinations are unpredictable
  9. Class effect cannot be taken for granted
  10. Effective therapies produce a combination of benefits and risks
  11. Most beneficial therapies do not save money, but produce an incremental benefit for incremental cost
  12. Applying the results of clinical trials is beneficial
  13. Some areas of medicine are underserved
  14. Participation is critical
  15. Conflict of interest statement
  16. References
  17. Appendices

Until recently, we had little proof that the practice of ‘evidence-based medicine’ was beneficial. We are now seeing the linkage between adherence to evidence-based professional standards and better patient outcomes [12]. To enhance this cycle of quality, broad-based practice networks will be needed. As more evidence for the link develops, agencies that pay for medical care likely will become more interested in funding research aimed at determining the comparative effectiveness of therapies. As practitioners strive to develop approaches that improve clinical performance, new types of trials that examine methods of delivering clinical care will become even more important.

One approach to integrating clinical-trial findings with clinical practice is the ‘bidirectional’ clinical network, in which clinical practice is characterized in a registry format in all patients seen in the practice, or in a representative sample. This ‘biosensor’ function allows gaps in practice and poor outcomes to be identified. Correction of problems can be accomplished through clinical trials to develop new therapies or audit and feedback systems. When clinical trials are begun, the network allows ascertainment of how representative the trial population is and how much the findings are incorporated into practice. The Society of Thoracic Surgeons has implemented such a systematic approach [13].

Some areas of medicine are underserved

  1. Top of page
  2. Abstract
  3. Introduction
  4. Treatment effects are modest
  5. Quantitative interactions are common, qualitative interactions are rare (so far)
  6. Unintended effects are common because surrogates are treacherous
  7. Long-term and short-term effects are not necessarily the same
  8. Combinations are unpredictable
  9. Class effect cannot be taken for granted
  10. Effective therapies produce a combination of benefits and risks
  11. Most beneficial therapies do not save money, but produce an incremental benefit for incremental cost
  12. Applying the results of clinical trials is beneficial
  13. Some areas of medicine are underserved
  14. Participation is critical
  15. Conflict of interest statement
  16. References
  17. Appendices

Although many areas of medicine are replete with major clinical-outcome studies, others essentially have no dependable evidence upon which to base practice. It is difficult to discern any particular reason for this. For example, although the mortality rates of acute myocardial infarction (MI) and acute heart failure are similar and account for similar numbers of admissions from the emergency department, more than 500 000 patients with acute MI have been entered into clinical trials but fewer than 50 000 patients with acute heart failure have.

For historical reasons, it has been assumed that disease entities with low volumes cannot be studied in multicentre trials. In fact, as potential therapies become more potent, clinical evidence may become even more important for low-volume diseases, given that many of them carry serious clinical consequences. A rational approach to public support of research networks will be needed to ensure that small-volume diseases are not left out of the effort to develop evidence to guide standards of medical care.

Participation is critical

  1. Top of page
  2. Abstract
  3. Introduction
  4. Treatment effects are modest
  5. Quantitative interactions are common, qualitative interactions are rare (so far)
  6. Unintended effects are common because surrogates are treacherous
  7. Long-term and short-term effects are not necessarily the same
  8. Combinations are unpredictable
  9. Class effect cannot be taken for granted
  10. Effective therapies produce a combination of benefits and risks
  11. Most beneficial therapies do not save money, but produce an incremental benefit for incremental cost
  12. Applying the results of clinical trials is beneficial
  13. Some areas of medicine are underserved
  14. Participation is critical
  15. Conflict of interest statement
  16. References
  17. Appendices

We have been labouring under a false dichotomy between clinical research and delivery of medical care. Given that clinical practice not based on evidence is unplanned experimentation, from which learning is unlikely to occur, it is more sensible to include the conduct of research as a professional standard of healthcare practitioners. Uncertainty about this issue has created a tension between points of view of clinical research and how it should be conducted. One point of view is that research is a different activity from the practice of medicine. Some of those who advocate this perspective have expressed the view that there is an inherent conflict when a patient's doctor is also an investigator. The alternative perspective is that clinical research is simply an element of modern clinical practice. Given our current understanding of therapeutics, it is appropriate for therapists and patients (increasingly viewed as consumers) to understand that we often simply do not know the best treatment, and that the highest-order relationship would recognize the benefit to society of efficient participation in clinical research.

If these principles are correct, the needs of clinical trials of the future can be considered from the perspectives of patients afflicted with diseases or at risk of them, those who must deliver healthcare (practitioners), and those who must pay for it (payers and governments). From the perspectives of these ‘stakeholders,’ the goal should be to create knowledge that informs clinical practice in the most efficient and effective manner. To meet the demands of the future, the clinical-research enterprise must move into a new era of maturity in which the approach to the prevention, diagnosis, and treatment of disease is seen as a partnership that is not separate from the practice of medicine, but represents the ‘state of the art.’

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Treatment effects are modest
  5. Quantitative interactions are common, qualitative interactions are rare (so far)
  6. Unintended effects are common because surrogates are treacherous
  7. Long-term and short-term effects are not necessarily the same
  8. Combinations are unpredictable
  9. Class effect cannot be taken for granted
  10. Effective therapies produce a combination of benefits and risks
  11. Most beneficial therapies do not save money, but produce an incremental benefit for incremental cost
  12. Applying the results of clinical trials is beneficial
  13. Some areas of medicine are underserved
  14. Participation is critical
  15. Conflict of interest statement
  16. References
  17. Appendices
  • 1
    Styring W, Jonas DK. Health Care 2020: The Coming Collapse of Employer-Provided Health Care. Indianapolis, IN: Hudson Institute, 1999.
  • 2
    DeMets DL, Califf RM. Lessons learned from recent cardiovascular clinical trials. Parts 1 and 2. Circulation 2002; 106: 74651, 8806.
  • 3
    Califf RM, DeMets DL. Principles from clinical trials relevant to clinical practice: Parts I and II. Circulation 2002; 106: 101521, 11725.
  • 4
    American Foundation for AIDS Research (AmFAR). Global Link Treatment Directory Search. Available at http://199.105.91.6/treatment/Home.asp (accessed 10 February 2003).
  • 5
    National Cancer Institute. PDQ Clinical Trials Search. Available at http://www.cancer.gov/cancer_information/doc.aspx?viewid=F2AFAEA4-64BD-4E44-B421-56026E252389 (accessed 10 February 2003).
  • 6
    Fleming TR, DeMets DL. Surrogate end points in clinical trials: are we being misled? Ann Intern Med 1996; 125: 60513.
  • 7
    Muhlbaier L. HIPAA in Clinical Trials: a Practical Guide for Research Compliance. Marblehead, MA: HCPro, Inc., 2000.
  • 8
    Novartis Pharmaceuticals Corporation. Lotrel Package Insert, May 2002. Available at http://www.pharma.us.novartis.com/product/pi/pdf/lotrel.pdf (accessed 10 February 2003).
  • 9
    Hoffman-LaRoche, Inc. Roche announces voluntary withdrawal of Posicor, June 8, 1998. Available at http://www.fda.gov/medwatch/safety/1998/safety98.htmposico (accessed 10 February 2003).
  • 10
    The ALLHAT Officers and Coordinators for the ALLHAT Collaborative Research Group. Major outcomes in high-risk hypertensive patients randomized to angiotensin-converting enzyme inhibitor or calcium channel blocker vs. diuretic: the Antihypertensive and Lipid-Lowering treatment to prevent Heart Attack Trial (ALLHAT). JAMA 2002; 288: 298197.
  • 11
    Clinical Research Roundtable. The Role of Purchasers and Payers in the Clinical Research Enterprise. Washington, DC: National Academy Press, 2002.
  • 12
    Peterson ED, Pollack CV, Roe MT et al. Early use of glycoprotein IIbIIIa inhibitors and outcomes in non-ST-elevation acute myocardial infarction: observations from the National Registry of Myocardial Infarction 4. J Am Coll Cardiol 2002; 39 : 303A (Abstract).
  • 13
    Society of Thoracic Surgeons. Home page. Available at http://www.sts.org/ (accessed 3 March 2003).

Appendices

  1. Top of page
  2. Abstract
  3. Introduction
  4. Treatment effects are modest
  5. Quantitative interactions are common, qualitative interactions are rare (so far)
  6. Unintended effects are common because surrogates are treacherous
  7. Long-term and short-term effects are not necessarily the same
  8. Combinations are unpredictable
  9. Class effect cannot be taken for granted
  10. Effective therapies produce a combination of benefits and risks
  11. Most beneficial therapies do not save money, but produce an incremental benefit for incremental cost
  12. Applying the results of clinical trials is beneficial
  13. Some areas of medicine are underserved
  14. Participation is critical
  15. Conflict of interest statement
  16. References
  17. Appendices

Appendix: Issues in future clinical trials

Problem no. 1

The cost of clinical research is excessive; if the cost could be cut by 50%, twice as many research questions could be answered.

Solutions:

  • Reduce redundant data collection in the healthcare process
  • Develop data standards and nomenclature to allow seamless sharing of research data
  • Develop public–private partnerships to improve congruity between research carried out for commercial purposes and that done to address public health needs.

Impediments:

  • Too many vested interests in particular nomenclature and data standards
  • Regulatory departments often ensconced in their own regulations, making them recalcitrant to change.

Problem no. 2

Much of the clinical research energy and funding today is spent attempting to collect data rather than design the best possible studies and analyse the data to provide solutions to health problems.

Solutions:

  • Nomenclature: Each area of medicine should develop a standard nomenclature that would be integrated across domains to produce a national standard set of terms to describe medicine. This systematic approach must allow frequent updates through a public process and experimental use of alternative or more detailed nomenclature for specific purposes, and it must be common to all of medicine (practice, research and regulatory). Essentially, the US governmental juggernaut of the National Institutes of Health (NIH), Food and Drug Administration (FDA), Center for Medicare and Medicaid Services (CMS), Department of Defense (DOD), and Veterans Affairs (VA) produces a standard that cannot be defeated (within the US). Practitioners will conform to CMS and DOD requirements to be paid. Researchers will comply with NIH requirements to get grants and contracts. Industry complies with FDA requirements to gain marketing approval.
  • Data standards: In addition to nomenclature, standards for sharing data must be completed. As with nomenclature, this effort will be most effective if driven by agreement amongst the key federal agencies that affect healthcare decisions by providers.
  • Data management hubs: Given universal access to data with common nomenclature and data standards, the focus can be on analytic capability of using the aggregated data to produce knowledge.

Impediments:

  • Getting common purpose within different federal organizations has been difficult
  • Harmonization with global standards will remain necessary to allow as much data sharing as needed
  • HIPAA regulations are prohibitive.

Problem no. 3

Investigators are not adequately trained to conduct clinical research to the standards that are needed.

Solutions:

  • A national training programme that establishes a minimum standard of knowledge for clinical investigators. This should be taught in related but distinct versions in medical, nursing, pharmacy, and physician-extender schools. This program should be available on the Internet and should provide a record of basic proficiency in clinical research.
  • Each NIH Institute, in conjunction with relevant professional societies, should develop a ‘specialty certification,’ including demonstration of knowledge concerning the basic nomenclature of the disease area.
  • Specific research networks or private organizations should feel free to develop even more specific standards for subspecialty areas that might be disease- or discipline-oriented.

Impediments:

  • The natural opposition to additional bureaucratic standards must be overcome. This will happen only if the standard is compelling and the case for it is endorsed by both industry and government.

Problem no. 4

Ethical and regulatory structures are complex, poorly organized, and of uncertain value.

Solutions:

  • Develop a set of ‘common rules’ that define standards for Data Monitoring Committee (DMC), Institutional Review Board (IRB) and FDA regulations
  • Include the Office for Human Research Protections (OHRP) in these ‘common rules’ so that investigators can have confidence that the chosen approaches will have universal agreement
  • Research needs to be funded to evaluate regulatory structures and systems to define which approaches are most effective.

Impediments:

  • These structures have become entrenched in their own purposes
  • Aggressively evaluating the ethics of research may raise more questions.

Problem no. 5

The questions asked by clinical trials do not answer the questions of greatest concern to patients, practicing doctors, and those who pay for medical care.

Solutions:

  • Include patients, payers and practitioners in peer review and prioritization.

Impediments:

  • Including nonresearchers in peer review would require substantial change compared with current standards.

Problem no. 6

There is tremendous difficulty in translating research results into practice.

Solutions:

  • Broad-based practitioner networks are needed that are bidirectional: doing large, simple clinical trials and providing a ‘biosensor’ function for what is happening in practice
  • More research is needed on systems to provide point of care decision support for practitioners and patients.

Impediments:

  • No one has yet defined how to maintain such networks at a reasonable cost
  • It is unclear whether practitioners will be willing to participate.

Problem no. 7

There is a gap between the current science of clinical research and the knowledge of the practicing community and the public (including its representatives).

Solutions:

  • Educating experts through larger and more effective training programmes
  • Educating practitioners, beginning with a different approach to medical school and continuing through enhanced Continuing Medical Education (CME), probably using professional societies
  • Educating the public.

Impediments:

  • Traditional medical education does not emphasize quantitative concepts
  • The public agenda with regard to education is crowded with other priorities.