National Academies Press: OpenBook

Modern Methods of Clinical Investigation (1990)

Chapter: Appendix A: Comparing the Development of Drugs, Devices, and Clinical Procedures

« Previous: 11. Attitudinal Factors That Influence the Utilization of Modern Evaluative Methods
Suggested Citation:"Appendix A: Comparing the Development of Drugs, Devices, and Clinical Procedures." Institute of Medicine. 1990. Modern Methods of Clinical Investigation. Washington, DC: The National Academies Press. doi: 10.17226/1550.
×
Page 147
Suggested Citation:"Appendix A: Comparing the Development of Drugs, Devices, and Clinical Procedures." Institute of Medicine. 1990. Modern Methods of Clinical Investigation. Washington, DC: The National Academies Press. doi: 10.17226/1550.
×
Page 148
Suggested Citation:"Appendix A: Comparing the Development of Drugs, Devices, and Clinical Procedures." Institute of Medicine. 1990. Modern Methods of Clinical Investigation. Washington, DC: The National Academies Press. doi: 10.17226/1550.
×
Page 149
Suggested Citation:"Appendix A: Comparing the Development of Drugs, Devices, and Clinical Procedures." Institute of Medicine. 1990. Modern Methods of Clinical Investigation. Washington, DC: The National Academies Press. doi: 10.17226/1550.
×
Page 150
Suggested Citation:"Appendix A: Comparing the Development of Drugs, Devices, and Clinical Procedures." Institute of Medicine. 1990. Modern Methods of Clinical Investigation. Washington, DC: The National Academies Press. doi: 10.17226/1550.
×
Page 151
Suggested Citation:"Appendix A: Comparing the Development of Drugs, Devices, and Clinical Procedures." Institute of Medicine. 1990. Modern Methods of Clinical Investigation. Washington, DC: The National Academies Press. doi: 10.17226/1550.
×
Page 152
Suggested Citation:"Appendix A: Comparing the Development of Drugs, Devices, and Clinical Procedures." Institute of Medicine. 1990. Modern Methods of Clinical Investigation. Washington, DC: The National Academies Press. doi: 10.17226/1550.
×
Page 153
Suggested Citation:"Appendix A: Comparing the Development of Drugs, Devices, and Clinical Procedures." Institute of Medicine. 1990. Modern Methods of Clinical Investigation. Washington, DC: The National Academies Press. doi: 10.17226/1550.
×
Page 154
Suggested Citation:"Appendix A: Comparing the Development of Drugs, Devices, and Clinical Procedures." Institute of Medicine. 1990. Modern Methods of Clinical Investigation. Washington, DC: The National Academies Press. doi: 10.17226/1550.
×
Page 155
Suggested Citation:"Appendix A: Comparing the Development of Drugs, Devices, and Clinical Procedures." Institute of Medicine. 1990. Modern Methods of Clinical Investigation. Washington, DC: The National Academies Press. doi: 10.17226/1550.
×
Page 156
Suggested Citation:"Appendix A: Comparing the Development of Drugs, Devices, and Clinical Procedures." Institute of Medicine. 1990. Modern Methods of Clinical Investigation. Washington, DC: The National Academies Press. doi: 10.17226/1550.
×
Page 157
Suggested Citation:"Appendix A: Comparing the Development of Drugs, Devices, and Clinical Procedures." Institute of Medicine. 1990. Modern Methods of Clinical Investigation. Washington, DC: The National Academies Press. doi: 10.17226/1550.
×
Page 158
Suggested Citation:"Appendix A: Comparing the Development of Drugs, Devices, and Clinical Procedures." Institute of Medicine. 1990. Modern Methods of Clinical Investigation. Washington, DC: The National Academies Press. doi: 10.17226/1550.
×
Page 159
Suggested Citation:"Appendix A: Comparing the Development of Drugs, Devices, and Clinical Procedures." Institute of Medicine. 1990. Modern Methods of Clinical Investigation. Washington, DC: The National Academies Press. doi: 10.17226/1550.
×
Page 160
Suggested Citation:"Appendix A: Comparing the Development of Drugs, Devices, and Clinical Procedures." Institute of Medicine. 1990. Modern Methods of Clinical Investigation. Washington, DC: The National Academies Press. doi: 10.17226/1550.
×
Page 161
Suggested Citation:"Appendix A: Comparing the Development of Drugs, Devices, and Clinical Procedures." Institute of Medicine. 1990. Modern Methods of Clinical Investigation. Washington, DC: The National Academies Press. doi: 10.17226/1550.
×
Page 162
Suggested Citation:"Appendix A: Comparing the Development of Drugs, Devices, and Clinical Procedures." Institute of Medicine. 1990. Modern Methods of Clinical Investigation. Washington, DC: The National Academies Press. doi: 10.17226/1550.
×
Page 163
Suggested Citation:"Appendix A: Comparing the Development of Drugs, Devices, and Clinical Procedures." Institute of Medicine. 1990. Modern Methods of Clinical Investigation. Washington, DC: The National Academies Press. doi: 10.17226/1550.
×
Page 164
Suggested Citation:"Appendix A: Comparing the Development of Drugs, Devices, and Clinical Procedures." Institute of Medicine. 1990. Modern Methods of Clinical Investigation. Washington, DC: The National Academies Press. doi: 10.17226/1550.
×
Page 165
Suggested Citation:"Appendix A: Comparing the Development of Drugs, Devices, and Clinical Procedures." Institute of Medicine. 1990. Modern Methods of Clinical Investigation. Washington, DC: The National Academies Press. doi: 10.17226/1550.
×
Page 166
Suggested Citation:"Appendix A: Comparing the Development of Drugs, Devices, and Clinical Procedures." Institute of Medicine. 1990. Modern Methods of Clinical Investigation. Washington, DC: The National Academies Press. doi: 10.17226/1550.
×
Page 167
Suggested Citation:"Appendix A: Comparing the Development of Drugs, Devices, and Clinical Procedures." Institute of Medicine. 1990. Modern Methods of Clinical Investigation. Washington, DC: The National Academies Press. doi: 10.17226/1550.
×
Page 168
Suggested Citation:"Appendix A: Comparing the Development of Drugs, Devices, and Clinical Procedures." Institute of Medicine. 1990. Modern Methods of Clinical Investigation. Washington, DC: The National Academies Press. doi: 10.17226/1550.
×
Page 169
Suggested Citation:"Appendix A: Comparing the Development of Drugs, Devices, and Clinical Procedures." Institute of Medicine. 1990. Modern Methods of Clinical Investigation. Washington, DC: The National Academies Press. doi: 10.17226/1550.
×
Page 170
Suggested Citation:"Appendix A: Comparing the Development of Drugs, Devices, and Clinical Procedures." Institute of Medicine. 1990. Modern Methods of Clinical Investigation. Washington, DC: The National Academies Press. doi: 10.17226/1550.
×
Page 171
Suggested Citation:"Appendix A: Comparing the Development of Drugs, Devices, and Clinical Procedures." Institute of Medicine. 1990. Modern Methods of Clinical Investigation. Washington, DC: The National Academies Press. doi: 10.17226/1550.
×
Page 172
Suggested Citation:"Appendix A: Comparing the Development of Drugs, Devices, and Clinical Procedures." Institute of Medicine. 1990. Modern Methods of Clinical Investigation. Washington, DC: The National Academies Press. doi: 10.17226/1550.
×
Page 173
Suggested Citation:"Appendix A: Comparing the Development of Drugs, Devices, and Clinical Procedures." Institute of Medicine. 1990. Modern Methods of Clinical Investigation. Washington, DC: The National Academies Press. doi: 10.17226/1550.
×
Page 174
Suggested Citation:"Appendix A: Comparing the Development of Drugs, Devices, and Clinical Procedures." Institute of Medicine. 1990. Modern Methods of Clinical Investigation. Washington, DC: The National Academies Press. doi: 10.17226/1550.
×
Page 175
Suggested Citation:"Appendix A: Comparing the Development of Drugs, Devices, and Clinical Procedures." Institute of Medicine. 1990. Modern Methods of Clinical Investigation. Washington, DC: The National Academies Press. doi: 10.17226/1550.
×
Page 176
Suggested Citation:"Appendix A: Comparing the Development of Drugs, Devices, and Clinical Procedures." Institute of Medicine. 1990. Modern Methods of Clinical Investigation. Washington, DC: The National Academies Press. doi: 10.17226/1550.
×
Page 177
Suggested Citation:"Appendix A: Comparing the Development of Drugs, Devices, and Clinical Procedures." Institute of Medicine. 1990. Modern Methods of Clinical Investigation. Washington, DC: The National Academies Press. doi: 10.17226/1550.
×
Page 178
Suggested Citation:"Appendix A: Comparing the Development of Drugs, Devices, and Clinical Procedures." Institute of Medicine. 1990. Modern Methods of Clinical Investigation. Washington, DC: The National Academies Press. doi: 10.17226/1550.
×
Page 179
Suggested Citation:"Appendix A: Comparing the Development of Drugs, Devices, and Clinical Procedures." Institute of Medicine. 1990. Modern Methods of Clinical Investigation. Washington, DC: The National Academies Press. doi: 10.17226/1550.
×
Page 180
Suggested Citation:"Appendix A: Comparing the Development of Drugs, Devices, and Clinical Procedures." Institute of Medicine. 1990. Modern Methods of Clinical Investigation. Washington, DC: The National Academies Press. doi: 10.17226/1550.
×
Page 181
Suggested Citation:"Appendix A: Comparing the Development of Drugs, Devices, and Clinical Procedures." Institute of Medicine. 1990. Modern Methods of Clinical Investigation. Washington, DC: The National Academies Press. doi: 10.17226/1550.
×
Page 182
Suggested Citation:"Appendix A: Comparing the Development of Drugs, Devices, and Clinical Procedures." Institute of Medicine. 1990. Modern Methods of Clinical Investigation. Washington, DC: The National Academies Press. doi: 10.17226/1550.
×
Page 183
Suggested Citation:"Appendix A: Comparing the Development of Drugs, Devices, and Clinical Procedures." Institute of Medicine. 1990. Modern Methods of Clinical Investigation. Washington, DC: The National Academies Press. doi: 10.17226/1550.
×
Page 184
Suggested Citation:"Appendix A: Comparing the Development of Drugs, Devices, and Clinical Procedures." Institute of Medicine. 1990. Modern Methods of Clinical Investigation. Washington, DC: The National Academies Press. doi: 10.17226/1550.
×
Page 185
Suggested Citation:"Appendix A: Comparing the Development of Drugs, Devices, and Clinical Procedures." Institute of Medicine. 1990. Modern Methods of Clinical Investigation. Washington, DC: The National Academies Press. doi: 10.17226/1550.
×
Page 186
Suggested Citation:"Appendix A: Comparing the Development of Drugs, Devices, and Clinical Procedures." Institute of Medicine. 1990. Modern Methods of Clinical Investigation. Washington, DC: The National Academies Press. doi: 10.17226/1550.
×
Page 187
Suggested Citation:"Appendix A: Comparing the Development of Drugs, Devices, and Clinical Procedures." Institute of Medicine. 1990. Modern Methods of Clinical Investigation. Washington, DC: The National Academies Press. doi: 10.17226/1550.
×
Page 188
Suggested Citation:"Appendix A: Comparing the Development of Drugs, Devices, and Clinical Procedures." Institute of Medicine. 1990. Modern Methods of Clinical Investigation. Washington, DC: The National Academies Press. doi: 10.17226/1550.
×
Page 189
Suggested Citation:"Appendix A: Comparing the Development of Drugs, Devices, and Clinical Procedures." Institute of Medicine. 1990. Modern Methods of Clinical Investigation. Washington, DC: The National Academies Press. doi: 10.17226/1550.
×
Page 190
Suggested Citation:"Appendix A: Comparing the Development of Drugs, Devices, and Clinical Procedures." Institute of Medicine. 1990. Modern Methods of Clinical Investigation. Washington, DC: The National Academies Press. doi: 10.17226/1550.
×
Page 191
Suggested Citation:"Appendix A: Comparing the Development of Drugs, Devices, and Clinical Procedures." Institute of Medicine. 1990. Modern Methods of Clinical Investigation. Washington, DC: The National Academies Press. doi: 10.17226/1550.
×
Page 192
Suggested Citation:"Appendix A: Comparing the Development of Drugs, Devices, and Clinical Procedures." Institute of Medicine. 1990. Modern Methods of Clinical Investigation. Washington, DC: The National Academies Press. doi: 10.17226/1550.
×
Page 193
Suggested Citation:"Appendix A: Comparing the Development of Drugs, Devices, and Clinical Procedures." Institute of Medicine. 1990. Modern Methods of Clinical Investigation. Washington, DC: The National Academies Press. doi: 10.17226/1550.
×
Page 194
Suggested Citation:"Appendix A: Comparing the Development of Drugs, Devices, and Clinical Procedures." Institute of Medicine. 1990. Modern Methods of Clinical Investigation. Washington, DC: The National Academies Press. doi: 10.17226/1550.
×
Page 195
Suggested Citation:"Appendix A: Comparing the Development of Drugs, Devices, and Clinical Procedures." Institute of Medicine. 1990. Modern Methods of Clinical Investigation. Washington, DC: The National Academies Press. doi: 10.17226/1550.
×
Page 196
Suggested Citation:"Appendix A: Comparing the Development of Drugs, Devices, and Clinical Procedures." Institute of Medicine. 1990. Modern Methods of Clinical Investigation. Washington, DC: The National Academies Press. doi: 10.17226/1550.
×
Page 197
Suggested Citation:"Appendix A: Comparing the Development of Drugs, Devices, and Clinical Procedures." Institute of Medicine. 1990. Modern Methods of Clinical Investigation. Washington, DC: The National Academies Press. doi: 10.17226/1550.
×
Page 198
Suggested Citation:"Appendix A: Comparing the Development of Drugs, Devices, and Clinical Procedures." Institute of Medicine. 1990. Modern Methods of Clinical Investigation. Washington, DC: The National Academies Press. doi: 10.17226/1550.
×
Page 199
Suggested Citation:"Appendix A: Comparing the Development of Drugs, Devices, and Clinical Procedures." Institute of Medicine. 1990. Modern Methods of Clinical Investigation. Washington, DC: The National Academies Press. doi: 10.17226/1550.
×
Page 200
Suggested Citation:"Appendix A: Comparing the Development of Drugs, Devices, and Clinical Procedures." Institute of Medicine. 1990. Modern Methods of Clinical Investigation. Washington, DC: The National Academies Press. doi: 10.17226/1550.
×
Page 201

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

Appendix A Comparing the Development of Drugs, Devices, and Clinical Procedures ANNETINE C. GELIJNS This chapter, initially published as a background document for the workshop discussions that underlie this volume, has three objectives. It provides an initial conceptualization of the medical technology development process within the broader innovation spectrum. It subsequently compares the evaluative strate- gies currently used in the development of new drugs, medical devices, and clini- cal procedures. Finally, it considers the implications of these strategies for the rationality and efficiency by which biomedical research findings are translated into clinical practice, and identifies some opportunities for change. AN INITIAL CONCEPTUALIZATION OF THE DEVELOPMENT PROCESS One of the essential and perhaps defining characteristics of Homo sapiens has always been the development and use of tools, often in response to environ- mental demands and challenges.! In this respect, the development and use of instruments to catch, collect, transport, and prepare food and to make clothing can be traced back to the very origins of human societies. Whereas environ- mental conditions have influenced the development of specific technologies, it can equally be observed that technology has influenced the human environment, thereby changing its underlying conditions. For example, it has been argued that the efficiency of late Paleolithic hunting technology may have caused the disappearance of large animals; the resulting difficulties in finding food stimu- lated development of the technologies of agriculture (1~. *This paper was partially supported by the Querido Award from the Netherlands Praeventiefonds (Dutch Fund for Disease Prevention). 147

148 ANNETINE C. GELIJNS Throughout history, one can observe this complex interrelationship between the development of technology and the physical, social, and economic environ- ment. For example, making a quantum leap through time from Paleolithic tools to the emergence of modern technology during the Enlightenment, the develop- ment and large-scale introduction of John Kay's shuttled transformed the textile industry fundamentally and, together with James Watt's steam engine some decades later, was one of the major forces shaping the industrial revolution (2~. Since then, technological change has had enormous economic consequences; in modern industrialized societies it has become the critical factor in long-term economic growth (3~. In addition, it has also contributed to the transformation of social relations, such as patterns of work and leisure, procreation, and com- munication. But, as Landau and Rosenberg observe, technological change "functions successfully only within a larger social and economic environment that provides incentives and complementary inputs into the innovation process" (4~. Both cultural and economic forces (a society's intellectual baggage and tol- erance for new ideas, investment in capital fo~ation, savings quotas, etc.), and the government policies reflecting them, have greatly influenced technological development. In comparison to the cybernetical relationship between techno- logical change and environmental factors (going back all the way to the origin of human societies3), the relationship between "science" and "the development of technology" is much younger. For many centuries the development of tech- nology was largely based on empirical knowledge arrived at by trial and error and was essentially independent of scientific understanding. However, the nature of technology development has changed considerably over time. A cru- cial period in the relation between science and technology occurred in the sev- enteenth and eighteenth centuries, when through the work of such scientists as Rene Descartes, Francis Bacon, Isaac Newton, and in medicine, Claude Bernard, the concept of nature was changed and the basis of a mechanistic worldview was laid. This new paradigm of the existence of mankind and its world based on the objectification of nature and the establishment of the experimental investigational method fueled scientific advances and increased the pace of technological change. In the nineteenth and twentieth centuries, sci- ence and technology became truly interdependent, as illustrated by the growth in industrial technology related to scientific advances in such fields as mechan- ics, electrodynamics, and chemistry (5) and more recently by the rapid expan- sion in professionally managed institutions for research and development. This change in the science-technology relationship gave rise to the so-called linear model of technological innovation (see Figure A.1), i.e., results were per- ceived to flow from basic research to applied research, targeted development, . ~ Basic ~ Applied ~ Targeted Research Research Development I [IGURE A.1 A linear model of the innovation chain. ~ Manufactunng ~ Ad option r Use

COMPARING DEVELOPMENT OF TECHNOLOGIES 149 manufacturing and marketing, adoption, and use. With the rapid expansion in biomedical research since World War II, this model has also become the popular representation of the process by which biomedical research findings are trans- lated into clinical practice. In medicine, this translation process can be catego- rized into three components: the development of new drugs and biologicals, that of medical devices, and that of clinical procedures. In other sectors of the economy, this linear-sequential representation has been found to impose a number of important conceptual limitations for the pur- pose of analyzing the development process. First, it implies that technological innovation is much more systematic than it really is. The stages of the innova- tive process are highly interactive with many feedback loops. For example, a strong reciprocal relationship exists between research and development: although both scientific and engineering research findings stimulate technology development, the availability of highly advanced technological products and processes stimulates and facilitates research. With regard to medical devices, for instance, the introduction of non-invasive imaging techniques made the cen- tral nervous system accessible to direct investigation of the anatomical corre- lates of function, opening up new vistas for research in neurophysiology. Furthermore, the linkage between research and development exists not only at the beginning of development, but also continues throughout the development process. In principle, the research and development stages are concurrent; for example, to solve problems encountered in the development of a new technolo- gy one may revert to the existing body of knowledge as accumulated in research or one may initiate new research (6~. The second limitation of the linear model is that not only research but also the broader environment as expressed through market forces influences each stage of the development process. For years the literature on technological innovation could be divided into "technology or science-push" theories (empha- sizing the importance of advances in research and technology as the main impe- tus to innovation) or "demand-pull" theories (stressing the importance of market demand as the main force in innovation). Mowery and Rosenberg, however, have demonstrated that technological development is an iterative process, in which both an underlying and evolving scientific and engineering knowledge base and market demand interact to achieve a particular innovation (7~. In a general sense, this observation also holds for innovation in medicine, and Figure A.2 depicts the medical technology development process as influ- enced by both supply and demand factors. Health care technology development can then be defined as a multi-stage process through which a new biological or chemical agent, medical device prototype, or clinical procedure is modified and tested until it is ready for regular production and utilization in the health care market. This development process can be divided into two closely related series of activities: technical modification and refinement (with pharmaceuticals and devices this includes scaling-up for production) and clinical evaluation of a potential innovations (see Figure A.2~.

150 ANNETINE C. GELIJNS Flow 1: Research, Discovery, and Invention 1 \ Flow 111: Health Care Market Diffusion ~ r FIGURE A.2 An interactive model of research, development, and diffusion streams. Whereas it is fairly obvious that current scientific and engineering knowl- edge (and its accessibility) determines the overall feasibility of specific techno- logical developments, the influence of market demand factors is more difficult to determine. The notion of a "market" in health care is different from the mar- ket concept in other sectors of the economy, where in principle the consumer determines what product he or she wants and then subsequently purchases it. The following major differences can be discerned: The market demand concept implies autonomous choice and a knowledge of available alternatives by consumers and patients. However, both autonomous choice and a realistic knowledge of the alternatives are often severely limited, and therefore health care professionals usually decide the kind and volume of technological interventions needed (8~. In a sense, these professionals are the consumers in the health care market, although their demand is derived from that of patients. ~ Furthermore, new medical technologies in addition to their benefits near- ly always entail a certain element of risk. The beneficial or adverse effects of a medical technology are considered to be quintessentially different from those of many other technologies because, as Renee Fox observes, they affect "basic and transcendent axes of the human condition: life, conception and birth, body and mind, . . . and ultimately mortality and death" (9~. During development, the benefits and risks of a new technology are highly uncertain. To reduce this uncertainty, a new technology is subjected to continuous clinical evaluation. · Finally, health care professionals are usually reimbursed for their services not by patients but by third-party payers. Because patients and professionals traditionally have been insulated from the financial consequences of their deci- sions, there have been no strong incentives to consider cost in their decision making. In the present-day environment of cost containment this situation is in the process of changing.

COMPARING DEVELOPMENT OF TECHNOLOGIES 15

152 ANNETINE C. GELIJNS decision making. As mentioned in Chapter 1, concerns have emerged as to the quality of the clinical evidence that forms the basis for decision making. This appendix will therefore address the following questions: 1. What kinds of clinical evidence play a role in decision making during the development of a potential innovation? What endpoints are assessed during the different stages of development? 2. What are the methods by which these endpoints are assessed during the development of a potential innovation? 3. What are the implications of these evaluative strategies for the effective- ness and efficiency of the process by which research findings are translated into clinical practice? THE DEVELOPMENT OF DRUGS The number and kinds of new molecular entities entering development are, to a large extent, a direct result of the activities undertaken and the judgments made in the drug discovery phase. In view of the close relationship between research and development, let us consider some characteristics of drug research and discovery before going into the development process. Drug Discovery Although research in various biomedical disciplines relevant to drug discov- ery takes place in academic, governmental and industrial laboratories, the devel- opment process is largely industry sponsored and takes place in industrial divi- sions and in clinical research settings, often in academic institutions. Historically, close relationships between industry, academia, and government have been crucial to drug discovery and development (131. During the twenti- eth century the interdependence of industrial, academic, and governmental research has intensified (14,15~.8 On the one hand, industrial laboratories exploit basic biomedical and clinical knowledge accumulated in academic and governmental settings, including the discovery of biologically active com- pounds (16~. On the other hand, basic research findings are also made in indus- trial laboratories, and the availability of new drugs often permits advances in basic, non-industrial research to be made (17~. This reciprocal relationship refutes the popular perception that equates basic research with academia, and subsequently, in a linear fashion, equates applied research and drug develop- ment with industry. With the emergence of the biotechnology industry, the real- ity of this complex interdependence has received new prominence. Since the origin of the pharmaceutical industry in the nineteenth century, the nature of the drug discovery process has changed substantially. In the second half of this century drug discovery has, to a large extent, moved away from the random screening of thousands of compounds—the prevalent mode of operation

COMPARING DEVELOPMENT OF TECHNOLOGIES i, 153 in Paul Ehrlich's days to the more rational design of drugs. This transition was made possible by a burgeoning number of research tools (such as electron microscopes, x-ray crystallography, and molecular modeling), advances in bio- chemical theory, and an increasing knowledge of physiological processes in health and disease. However, both serendipity and empirical processes of trial and error remain important elements of drug discovery today (18~. According to Maxwell (19), four drug discovery approaches can be identified at present: 1. The basic approach This approach entails studies to elucidate new bio- chemical leads or biomedical hypotheses, which may result in the synthesis of new compounds. 2. Screening of compounds This screening is usually targeted, i.e., based on a distinct rationale, for instance, blocking of a particular receptor. Because compounds may show unexpected therapeutic activity in other areas, it can also be valuable to perform some general screening. 3. Molecular modification Because the first candidate in a therapeutic class s rarely optimal, the objective of molecular modification is to discover improved agents from a "lead compound" with, for instance, a longer duration of action and/or greater selectivity. Maxwell distinguishes between "enlightened opportunism" and "unenlightened opportunism." The former refers to the molecular modification of pharmacological compounds, identified at an early stage of their development, in order to develop an improved agent. The latter refers to making a close chemical variation of a specific drug, which often is already widely diffused on the market. This distinction, however, is not always easy to make (see below), since much of this research seeks to overcome shortcomings of the marketed drug. 4. Clinical observations The final source of new drugs can be the clinical observation that a compound, new or old, has unexpected therapeutic actions in patients.9 That these strategies are not mutually exclusive can be illustrated by the discovery and development of beta-blockers (see Box below). Over time, the drug discovery and research process has become increasingly complex and sophisticated. Interesting compounds are extensively screened both in vitro and in viva for pharmacological and toxicological effects.12 There has been a rapid increase in the number and kinds of toxicological tests (27,28,29~. Following short-term animal tests, long-term animal studies are ini- tiated to detect possible mutagenicity, carcinogenicity, and teratogenicity. These studies often continue for a number of years concurrent with initial human trials. For testing biotechnology-based drugs, however, toxicology studies in animals do not always make sense when the new biologicals are products of human genes and are functionally species specific. More in general, animal tests sometimes have variable relevance for predicting the effects of an agent in humans. The changes in preclinical testing are reflected in the time spent in this stage of the research process and the costs incurred. While the duration of preclinical

154 ANNETINE C. GELlJNS In the late 1 940s, clinical research on nerves revealed that the stimulation of one set of nerve pathways, producing epinephrine and norepinephrine, made the heart beat faster and increased the need for oxygen. This research also suggested the existence of To types of receptors in the human body, alpha and beta receptors, that mediate the effects of nore- pinephrine and epinephrine (21~. This work resulted in the hypothesis by Black, one of the 1988 Nobel laureates for physiology or medicine, that blocking one of these receptors would diminish the heart's demand for oxygen, possibly providing relief to angina sufferers. Black and his col- leagues at Imperial Chemical Industries (ICI) tried to develop analogues of an earlier discovered compound dichloroisoproterenol (22~. This com- pound had been found to induce beta-adrenergic blockade activity, but also had partial agonist (sympathomimetic) activity. They first developed pronethalol (23), which was found to induce considerable human side effects, such as nausea, vomiting, and light-headedness. They then developed propranolol (24), first marketed as Inderal, which was free of the agonist activity of dichloroisoproterenol and the side effects of pronethalol. The discovery and development of beta-blockers thus demonstrate the importance of the "basic approach" and the interaction with strategies 2 and 3. In the words of the Nobel committee's citation, 'while drug development had earlier mainly been built on chemical modifi- cation of natural products, they (the laureates) introduced a more rational approach based on the understanding of basic biochemical and physio- logical processes" (25~. Following the introduction of beta-blockers into clinical practice, it was observed that beta-blockers also played a role in lowering blood pressure and preventing heart attack and coronary death. Finally, the proliferation of various beta-blockers has resulted in a number of more selective drugs as well as some so-called "me-too" drugs (26~.10 (animal) tests was approximately one year in the mid 1960s, it increased to approximately three and a half years in the early 1980s, with a concomitant increase in costs (30,31~. Yet, uncertainty remains a crucial element in drug dis- covery and preclinical research: the attrition rate traditionally has been such that, of roughly each 10,000 compounds synthesized, 1,000 will go into animal research and only 10 will initiate human testing (32~. Drug Development In the United States, the decision to proceed with the development of a com- pound, including its clinical evaluation, initially involves a drug company and the FDA. Subsequently it engages clinical investigators, Institutional Review Boards (IRBs), and the research subjects themselves. The 1962 amendments to the Food, Drug, and Cosmetics Act require a sponsor to apply to the FDA for permission to initiate human testing with an Investigational New Drug (IND).

COMPARING DEVELOPMENT OF TECHNOLOGIES 155 The purpose of such an IND application is to protect human subjects, in part by making sure that the proposed clinical investigations are as efficient as possible to minimize the numbers of patients exposed to the risks of such trials. An IND application must contain essentially all of the information then known (the mean size of an IND is 1,250 pages) on the nature of the new compound, for- mulation and identification methodologies, stability information, manufacturing methods, the methods and results of preclinical animal studies, the proposed clinical development plan for trials, and the identity and qualifications of clini- cal investigators.13 The FDA classifies IND applications according to a compound's chemical We and its potential benefit, to determine priority for review. In principle, clin- ical trials can start 30 days after the FDA receives an IND application, unless the agency orders a "clinical hold." After an IND application has been approved, a multi-stage process of clinical investigation starts; the demarcation lines between the various phases are somewhat fluid. Human testing is initiated with Phase I studies, which ordinarily last between six months and one year. These studies usually involve 20 to 100 healthy human volunteers, except in the case of drugs with potentially high toxicity lev- els such as neoplastic or AIDS drugs where it is considered unethical to sub- ject healthy humans to the risk of these side effects, and thus patients are involved from the beginning. The objective of Phase I studies is to provide information on the dose of an experimental drug that might be used, how often, and especially on potential side effects. While drug absorption, metabolism, excretion, and some effects on tissues and organs are measured, a major concern is acute side effects in humans. Drug administration begins at very low single doses (for instance, one-eighth of the lowest dose that has caused a measurable effect in the most sensitive animal species), followed by multiple doses if no adverse effects are encountered as the dose is increased (351. Safety concerns in this phase may include acute cardiovascular reactions, gastrointestinal distur- bances, central nervous system disturbances, bronchopulmonary reactions, and anaphylactic reactions (291. These studies generally involve both laboratory testing and clinical observation. Development was discontinued during Phase I studies of 20 percent of the drugs that initiated human testing (361.14 The reasons for these discontinuations are safety (8 percent of the 20 percent), efficacy (6 percent of the 20 percent), and lack of commercial interest (6 percent of the 20 percent). Not uncommonly, chemical and pharmacological research on back-up compounds is pursued in case the compound undergoing development is discontinued due to side effects or lack of efficacy. For example, the anti-arthritic drug, piroxicam, was the third member of a new chemical series (the oxicams), but the first one to make it to the market. Simultaneous with Phase I clinical studies, technical development activities take place to improve a particular compound's formulation. In developing a suitable tablet or capsule formulation, a number of physical, chemical, and

156 AN'NETINE C. GELIJNS pharmacology issues need to be resolved, such as the use of stabilizing agents (e.g., anti-oxidants), micro-encapsulation, or the development of slow-release forms to achieve the optimum rate of absorption. Phase II clinical studies involve a few hundred patients and usually take sev- eral months to two years. The main emphasis in Phase II studies is to examine the efficacy of a compound in treating the clinical problem for which it is intended.15 At this point, the endpoints are selected that will be pursued both in Phase II and in Phase III studies. A major issue is the choice of endpoint; should one focus solely on intermediate endpoints, such as changes in biochem- ical, physiological, and anatomical parameters, or should one also include clini- cal endpoints, such as effect on mortality, morbidity, or quality of life. These decisions involve complex considerations regarding the disease, the time frame of treatment, and the scientific and regulatory acceptability of the relationship between intermediate endpoints and disease treatment. They can have a consid- erable impact on the scope of the development process. Traditionally, a number of intermediate endpoints, such as lowering blood sugar in diabetes or lowering blood pressure in severe hypertension, have been accepted as valid by the various parties involved in drug development. In other, more recent cases involving intermediate endpoints, such as clot lysis in myocardial re-infarction or the increase of hematocrit levels in anemic dialysis patients, there has been considerable disagreement about their value. For instance, in the development of recombinant erythropoietin, a stimulator of red blood cell development, a nine-center, 300-patient efficacy trial demonstrated significant increase of hematocrit levels, while none of the patients developed antibodies to erythropoietin. The FDA found hematocrit increase alone insuff~- cient proof of efficacy and required additional evidence of clinical benefit. The company was able to demonstrate a reduction in the number of transfusions and improvements in exercise tolerance and patient well-being. The license applica- tion is being reviewed (38~. A number of factors may influence the acceptabili- ty of the kind of endpoints to pursue. For example, in hyper-cholesterolemia clinical endpoints such as death from myocardial infarction may take a long time to develop, and thus practical reasons dictate the use of intermediate end- points such as reduction of low-density lipoprotein-cholesterol. In this case the acceptability of intermediate endpoints is heightened because the association between the intermediate endpoint and the clinical problem is perceived to be strong (39~. The crucial question, however, often is not whether to pursue intermediate or clinical endpoints, but which endpoint should be pursued at which stage in the development process (especially pre- or post-approval). This question is impor- tant because the traditional notion of what constitutes valid clinical endpoints is evolving. Since many therapeutic agents for today's chronic degenerative dis- eases only treat symptoms, the focus in clinical evaluations is shifting toward measuring long-term benefits and risks. Furthermore, it is increasingly apparent

COMPARING DEVELOPMENT OF TECHNOLOGIES 157 that risks and benefits should be measured not only in terms of reducing mortal- ity but also in terms of improving functional status and quality of life. Such quality of life studies are becoming more important in the pharmaceutical area. Recent examples are provided by quality of life evaluations of auranofin and captopril (40~. Phase II studies also attempt to detect short-term side effects. The safety concerns in Phase II and in Phase III studies include cumulative organ toxicity, hypersensitivity reactions, metabolic abnormalities, endocrine disturbances, and if women of childbearing age are involved, teratogenicity (29~. The Food, Drug, and Cosmetics Act requires "substantial evidence . . . of safety and effectiveness . . . consisting of adequate and well-controlled investi- gations." Most Phase II studies are double-blinded, randomized controlled clin- ical trials. While placebo control is the design of choice, the agency will accept no-treatment controls, standard treatment, and even historical controls (37~. The well-designed randomized controlled trial (RCT) is generally regarded as the statistically most powerful method to determine efficacy (42~.16 The essence of an RCT is that patients are randomly assigned to a treatment group which receives the experimental drug or to a control group which receives a placebo, standard treatment, or no treatment. According to Chalmers (43), a clinical trial is ideally quadruple-blinded: the therapy is disguised to physicians and patients (double-blinded), as are the randomization process and the ongoing results. Both randomization and blinding reduce bias17; the differences in health outcome can thus be attributed to the intervention, within the limits of statistical methodology. In a well-designed trial, the numbers of patients and the end- points are chosen to obtain clinically important and statistically significant results.18 The degree of complexity in determining efficacy and safety depends on the therapeutic class to which the experimental drug belongs. At one end of the spectrum are the anti-infectives. Efficacy testing of these compounds is a rela- tively straightforward assessment of whether the compound kills the microor- ganism at the site of infection. Due to the acute nature of most infections, there may be less need for chronic toxicity testing. At the other end of the spectrum are psychopharmacological drugs. Determination of efficacy in psychiatric dis- eases, with a complex interplay of neurobiological, environmental, and psycho- logical factors, is difficult. There are fewer objective tests for psychiatric disor- ders and one often deals with "soft" measures, making it necessary to subject these drugs to a wider range of tests. As these drugs may often be taken for long periods, chronic toxicity tests are needed. These varying degrees of com- plexity are reflected in the duration of the development process; for example, the development of psychopharmacological agents takes 3.1 years longer than for cardiovascular drugs, and 7.3 years longer than for anti-infective agents (30~. Within the total clinical development spectrum the highest dropout rate for

158 ANNETINE C. GELIJNS new molecular entities occurs during Phase II studies when 39 percent are dis- continued (36~. The FDA analysis lists as reasons for these discontinuations safety (13 percent), efficacy (12 percent) and economic considerations (15 per- cent). That efficacy and "lack of commercial interest" are prominent reasons for discontinuation is not unexpected if one considers that the main objective of Phase II studies is to determine efficacy, and that the line between "no efficacy" and "not enough efficacy to be competitive" may be quite fluid. With the rising costs of development, studies of the potential market for a drug increasingly occur during Phase II and Phase III studies. The relative prominence of safety as a reason is in part due to the fact that the results of long-term animal studies are usually obtained at this point in the development continuum. At the end of Phase II studies, a recent change in the U.S. regulatory scheme permits a sponsor to obtain a so-called treatment IND for compounds intended to treat immediately life-threatening diseases.19 This system makes experimen- tal drugs available at a reasonable cost before marketing approval for terminally ill patients not enrolled in clinical trials. A recent example of a Treatment IND drug is pentostatin, for patients with hairy cell leukemia. With drugs for very serious (but not immediately life-threatening) diseases, a sponsor may request a treatment IND in the course of Phase III studies. During Phase II and Phase III clinical studies much industrial effort is direct- ed, usually by chemists and engineers, toward process optimalization and 'scal- ing up' for production.20 The scaling-up for an efficient production process, involving pilot plant operations and various other process and quality control measures, is a crucial part of the development process. By the time an investigational drug is ready for Phase III studies, quite a good picture of its safety and efficacy has usually emerged, at least for a market approval decision. Only 5 percent of the compounds initiating Phase III trials are discontinued. These trials commonly involve up to several thousand patients (2,000-3,000), usually are multi-center trials, and are often multination- al in scope. On average they last between one and four years. The purpose of these controlled trials and open (uncontrolled) studies is two-fold: to further clarify a compound's therapeutic effects, for example by studying dose levels and schedules in larger patient groups, and to provide information on the side effects and possible toxicity of the drug candidate. These Phase III studies are important in determining what will be in a package insert for the drug, and thus what market claims can be made for a new entity in advertising. There are inherent limits to how much can be known about a drug prior to its general use in everyday practice. It is well accepted that the detection of delayed or rare (less than 1:10,000) adverse events may require long time peri- ods of exposure, a latent period to have expired, or the exposure of thousands of patients. Wardell et al. (45) point out that a sample size of 306,000 for each group would be needed to detect a difference between an incidence rate of 1/10,000 and 2/10,000 at the 90 percent power level (using a two-sided test, a 0.051. Some serious toxicity may occur much less frequently, for instance chlo-

COMPARING DEVELOPMENT OF TECHNOLOGIES 159 ramphenicol-induced aplastic anemia probably occurs only in 1:40,000 to 1:50,000 exposures (46~. However, for side effects of drugs that have less than fatal consequences but are medically important, the important difference to detect is between 1:500 and 1:1,000 or 1:10,000. Furthermore, as Wiener (47) argues, failure to detect adverse effects in Phase III studies may be more than a matter of time and numbers. Side effects may be influenced by environmental factors and variations in physician or patient characteristics (such as differing pharmacogenetic profiles or the use of other drugs, etc.~. The occurrence of these side effects may go unnoticed in carefully controlled and selected pre- marketing studies; detection will require actual patient care settings. While the full picture of the risks involved may become apparent only with the widespread diffusion of a drug, an equal argument can be made about benefits. The full range of information on effectiveness of a drug cannot be expected to emerge in Phase III clinical trials that are designed to test the null hypothesis of efficacy. The eligibility criteria for these trials almost invariably excludes a spectrum of at-risk patients, such as those with multi-morbidities, those using many drugs, and special patient groups, such as pregnant women, newborns, children or the very old.21 Thus, the findings of RCTs may not easily be appli- cable to the total patient population, especially if linearity cannot be assumed in extrapolation (48~. It follows that pre-marketing clinical studies are of necessity incomplete in developing information that can be used to optimize medical use of a drug. A marketing approval decision therefore can never be an all-benefits- known and no-risk situation.22 At the end of Phase III trials a New Drug Application (NDA) or, in the case of a biological, a Product License Application is usually submitted to the regu- latory agency, with a request for approval to market a specific compound for the indications specified in the application. The FDA ranks NDAs according to their review priority (49~.23 A drug, for instance, that is a "new molecular enti- ty" not previously marketed in the United States and that promises to provide "important therapeutic gain" (i.e., may diagnose or treat a disease not adequate- ly treated or diagnosed by any marketed drug) receives the highest priority rat- ing. An NDA contains detailed information on the laboratory formulation and chemistry of the drug, the results of all investigations, the manufacturing pro- cess, quality control procedures, the labeling of the drug, and samples of the drug in its proposed dose and form. Commonly an NDA encompasses over 100 volumes of information containing 60,000 pages each. Electronic NDAs, which contain the data in machine-readable form, are becoming more common and may prove important in facilitating the FDA review process. This review pro- cess involves a team consisting of at least a medical officer, a pharmacologist, and a chemist.24 If applications concern significant new drugs or involve com- plex issues, they may be referred to an advisory committee for review and rec- ommendations. With regard to biologics, licensing committees are used to pro- vide the expertise as appropriate to the product. The FDA review time takes 2.5 to 3 years on average (30~.

160 ANNETINE C. GELIJNS Drug Diffusion and Post-Marketing Surveillance After a new drug is approved for marketing, coverage and reimbursement decisions by third-party payers can affect the diffusion of a drug and hence the development continuum.25 These decisions should be placed within the context of a country's health care reimbursement policies. At present, these policies are changing in an attempt to contain health care costs; see, for instance, Medicare's prospective payment system, pro-generic substitution laws, and restrictive hos- pital formularies. With these changes, coverage is becoming a more important decision point in the process, as illustrated by the heated debate surrounding Medicare's decision not to authorize extra reimbursement for tissue plasmino- gen activator. One consequence is that cost analyses and cost effectiveness studies are becoming a much more prominent part of a drug's evaluation. However, these analyses and their influence on decision making are outside the scope of this paper.26 Following the marketing approval decision, a new drug generally diffuses into clinical practice (with the active help of marketing professionals). With the present-day chronic diseases some of the most important therapeutic informa- tion, both on rare and delayed side effects and on long-term effectiveness, can be provided only after a new drug has been used in everyday practice. The objective of so-called Phase IV (or post-marketing) studies is to provide this information. This can be done by performing additional controlled clinical tri- als or by using observational (non-experimental) surveillance systems. The importance of these studies is underlined by the fact that new indications often are discovered only in clinical practice and subsequently drugs may be pre- scribed for these unapproved indications. One should realize that the FDA only regulates the introduction of new drugs and not their use in medical practice. Only experimental Phase IV studies may be used to request approval for a new indication and to change the drug labeling. In addition, these studies have sometimes been encouraged by manufacturers from a marketing standpoint, to create a pool of physicians familiar with the drug (52~. Most industrialized countries have some kind of post-marketing surveillance system to detect potential adverse effects. Such a system generally depends on a variety of methodological approaches, as no single method is fully effective. One approach depends on adverse effect reporting (53~. In the United States, physicians traditionally report suspected adverse effects voluntarily to the com- pany (the Food, Drug, and Cosmetics Act requires manufacturers in turn to immediately report these effects to the FDA). In addition, physicians may vol- untarily report suspected adverse effects directly to the regulatory agency, to the medical literature, or to disease or specialty registries.27 While advantages of adverse effect reporting are its potential coverage of the entire population and low operation costs, important weaknesses are found in incompleteness and inaccuracy. For example, due to a variety of factors, there is considerable under-reporting; the overall return on the U.K. Yellow Card System is estimated

COMPARING DEVELOPMENT OF TECHNOLOGIES 161 to be only 10 percent. With adverse effect reporting one also cannot measure the incidence of the risk. Furthermore, this reporting is by nature a hypothesis- generating activity; the subsequent testing of the hypothesis will depend on other methods. Methodological approaches for further analysis of the adverse events report- ed by physicians or manufacturers (and for monitoring signals of suspected adverse effects) include experimental and observational methods. While experi- mental methods have especially been applied to further examine efficacy post- approval, risk measurements in specific patient populations are sometimes also undertaken. At present, however, there is increasing interest in epidemiological studies, such as case-control and cohort studies, to measure adverse drug effects.28 The advantage of cohort studies is that they can establish the likely incidence of the risk. Disadvantages are that they are potentially expensive and may yield the results more slowly than case-control studies. Case-control stud- ies are useful if the frequency of events is very rare (up to 1:10,000~. Disadvantages are that controls are often difficult to establish and the studies cannot establish absolute risk. The proliferation of large-scale automated data bases, such as those main- tained by health maintenance organizations or Medicaid, may open up exciting opportunities to study a drug under general conditions of use. These data bases may contain demographic data, drug prescription data, or patient hospital admission and discharge data. With advances in computer capabilities it is increasingly possible to link different data bases, for instance, pharmacy records with medical record data bases (55~. In essence, the Drug Surveillance Research Unit, initiated by Inman (56) in the United Kingdom in 1980, is based on this principle.29 In the same vein, the FDA has carried out a number of hypothesis-testing studies using Medicaid and other medical record linkage data bases. Industry is also increasing its efforts in pharmacoepidemiological research. As these large- scale data bases exist for other reasons, their operating costs are much lower than those associated with registries. In addition, they may lack the reporting bias and the inadequate follow-up that renders case studies problematic (57~. However, limitations exist in the adequacy of the data collected in these data bases (see below). As argued above, Phase IV studies also need to examine the long-term effec- tiveness of a drug. Since the early 1970s, the FDA has sometimes requested post-approval research as a condition of approval, often with good reason (see, for instance, the approval of levodopa). Studies done post approval to examine the benefits of a drug in different patient populations or with different dosages are usually an extension of the type of studies done before marketing approval (58~. In addition a number of large-scale randomized trials have also been undertaken post approval that were funded not by the sponsor but, for instance, by the National Heart, Lung, and Blood Institute. In view of the very high costs associated with these large-scale trials (between $10 million and $100 million),

162 ANNETINE C. GELIJNS the number of such RCTs is limited (59~. Furthe~ore, as mentioned above, the RCT may not always be most helpful as a foundation for therapeutic decisions. It has therefore been proposed that modern observational methods could play an important complementary role to the RCT for assessing the effectiveness of a drug. Major weaknesses traditionally associated with these methods have made the determination of the cause-and-effect relationships between drug use and outcomes more diff~cult.30 However, in recent years there have been advances in the design and the execution of observational studies, which may address some of these weaknesses. THE DEVELOPMENT OF MEDICAL DEVICES Over the past quarter century there has been an acceleration in the develop- ment of new medical devices, in part because of rapidly expanding scientific and engineering knowledge (61~. In view of the reciprocal relationship between research and development, we will briefly consider the interactions between the research and invention phase of medical devices and the development phase. The Invention of Medical Devices In addition to basic biomedical and clinical research, bioengineering research, which builds on advances in the physical sciences, mathematics, and engineering in other sectors, provides an important contribution to the knowl- edge base underlying medical device development.31 Basic bioengineering research predominantly takes place in university and government laboratories. In contrast to some European countries and Japan, funding for fundamental research in biomedical engineering is relatively small in the United States (e.g., 1 percent of the National Institutes of Health budget) and is dispersed through a number of agencies (62~. Compared with the United States, the Federal Republic of Germany has nearly double the amount of space and equipment for bioengineering research (63~. Some recent efforts, however, may ameliorate this situation. The National Science Foundation, for instance, has established a pro- gram to fund high-risk fundamental bioengineering research. On the applied research side, the Small Business Innovation Research program was established in the early 1980s by the National Institutes of Health to provide R&D grants or contracts to small businesses. According to an OTA analysis, 40 percent of grant applications in 1983 concerned medical devices and 23 percent of these applications were funded (64~. Federal support is complemented by private investment, especially in applied research. In 1986, the medical device industry invested on average 7.5 percent of sales in R&D (65~. In 1979, medical device firms in the five medical device Standard Industrial Classification codes (x-ray and electromedical equipment, surgical and medical instruments, surgical appliances and supplies, dental equipment and supplies, ophthalmic goods) reported that 3.7 percent of their

COMPARING DEVELOPMENT OF TECHNOLOGIES 163 company-sponsored R&D budget was basic research and 23 percent was applied research (64,66~. The investment in research depends on the type of device as well as the kind of firm involved. As Spilker observes, medical devices are a much more het- erogenous group of products than drugs in terms of design, use, and purpose; many devices never come in contact with patients, some do briefly, and others do permanently (67~. There are roughly 1,700 different types of medical devices and 50,000 separate products (64~. Clearly, the research required for the invention of disposable needles is very different from that for the invention of a CT scanner.32 Equally, there is much more variety in the kinds of firms that invent and develop medical devices than is the case with drugs. The industry is characterized by a large number of small firms; approximately 50 percent of U.S. medical device manufacturers have fewer than 20 employees. Large com- panies, however, dominate the industry in terms of sales (68~. According to Roberts, small funs and even individuals produce most of the innovations in the early stages of developing a new class of medical devices, whereas larger firms play an especially important role later on in the development process (sometimes through the acquisition of small firms). As Roberts put it, the invention of "medical devices is usually based on engineering problem solving by individuals or small firms, is often incremental rather than radical, seldom depends on the results of long-term research in the basic sciences, and generally does not reflect the recent generation of fundamental new knowledge. It is a very different endeavor from drug innovation, indeed" (68~. This observation, however, is not as easily applicable to radical innovations, such as those in mod- ern imaging devices, which require large-scale investments in research and development. Such resource-intensive innovations usually take place in large firms. After a product is invented, a patent application may be filed. While patent protection is extremely important to the pharmaceutical research and develop- ment process- partially because of the long duration of the R&D process and the relative ease with which drugs can be copied- the value of patent protection in medical device development is much less evident. In the device area, it prob- ably is easier to invent around a patent, and the research and development time is generally much shorter. Furthermore, with devices that require large capital costs, the need for large-scale investments may prevent competitors from enter- ing the market, and small firms may depend more on trade secrets. Whereas the potential users of new medical devices, i.e., the physician- researchers, may play an important role during the development process, they also may be crucial to the invention of medical device prototypes. Not only do they identify the clinical need for a new device or for improvements in existing devices, but they may also be the innovators and builders of the original proto- type. Von Hippel first described the importance of users in the invention of such scientific instruments as gas chromatography, nuclear magnetic resonance, ultraviolet spectrophotometry, and transmission electron microscopy (69~. He

16;4 ANNETINE C. GELIJNS concluded that 80 percent to 100 percent of the key innovations in these four fields were originated by users and not the ultimate manufacturers. Von Hippel and Finkelstein underlined the importance of users with regard to the automated clinical chemical analyzer (70~. For example, the initial prototype of an auto analyzer was developed by Skeggs in the pathology department of Case Western Reserve University; Technicon then made a licensing agreement with Skeggs to patent the auto analyzer and further developed and marketed the machine (71~. Shaw, who analyzed 34 medical equipment innovations in Great Britain presented similar results33 (72~. It follows that close interactions between clinicians and industry are important to the development of medical devices.34 Roberts and Peters, however, found that academicians in the Massachusetts Institute of Technology physics, mechanical engineering, and chemical engineering departments and in two large research laboratories did not readily transfer their ideas for commercial development (74~. This finding was repeated in an analysis by Roberts (68) of two major medical centers in the Boston area, although this may change somewhat in the present-day climate, where universities and their medical centers are becoming more market-oriented (75~. These considerations affect industrial decision making. In the pharmaceuti- cal industry, the decision to invest in particular research areas involves "poten- tial demand" for a pharmaceutical as an important criterion. If the user-domi- nance paradigm of Van Hippel plays an important role in some parts of the med- ical device industry, manufacturers' decisions will be made later in the R&D continuum. The decision whether to pursue development of a prototype involves both technical and market factors.35 Medical Device Development In most industrialized countries, the development of new medical devices is governed by regulatory schemes, either in the form of standards or extended pharmaceutical laws, which focus mainly on safety. In contrast, the United States has passed a specific law governing the development of medical devices. Prior to the passage of these amendments to the Food, Drug, and Cosmetics Act in 1976, the FDA asked Arthur D. Little consultants to provide insight into the safety and efficacy testing practices of medical device funs (77~. This analysis revealed that most devices were tested during their development, but that the extent and nature of clinical evaluation varied considerably among products. In addition, there was considerable variety within product categories. For exam- ple, one developer of an artificial knee undertook clinical trials in 200 patients, another performed informal trials with 75, and the third used only 50 patients with no set protocol. Furthermore, in comparison to drug evaluations, the crite- ria of clinical evaluation may differ with new clinical devices. Criteria more often include user acceptability, either of the design or of the reliability and ease of use in the clinical setting, and the competitive advantages of a new device

COMPARING DEVELOPMENT OF TECHNOLOGY 165 versus alternative devices. Finally, in most cases evaluations did not include the classical randomized clinical trial, review by Institutional Review Boards (IRBs), or were conducted with patients who signed consent forms. Because medical devices are a much more heterogeneous group of products than are drugs, it is understandable that some variation in clinical evaluation exists. The existence of considerable variation within device categories and the fact that half of the clinical investigations had no formal protocol indicate room for improvement. In addition, the risks associated with some devices, such as certain cardiovascular implants or IUDs, became evident in the 1970s. The Cooper Committee was established to recommend device legislation. It pro- posed different levels of regulatory control based on the likelihood of risk inher- ent in specific classes of devices, with more rigorous regulation for devices with higher risk potential. In 1976 the Medical Device Amendments (Public Law 94-295) were passed to ensure that new devices were "safe and effective" before they were marketed (78~. These amendments divide medical devices36 into three classes (80,81~. Approximately 30 percent of all types of medical devices are in Class I. Class I devices include such instruments as tongue depressors, which do not support or sustain human life and do not present a potentially unreasonable risk of illness or injury. They are subject to the general controls used before passage of the Medical Device Amendments, such as regulations regarding registration, pre-marketing notification, record keeping, labeling, and Good Manufacturing Practice (GMP) regulations. About 60 percent of devices are in Class II. They may involve some degree of risk and are subject to federally defined performance standards (such as x-ray devices). To date, however, no performance standards have been issued by the FDA, and existing national or international product standards apply. Finally, all devices that are life supporting or sustaining, that are of substan- tial importance in preventing impairment of health, or that have a potential for causing risk of injury or illness are in Class III. For these, the sponsor needs to demonstrate safety and efficacy before the FDA grants marketing approval. Approximately 10 percent of medical devices are in Class III, such as the artifi- cial heart, DNA probes, or laser angioplasty devices. According to the law, devices introduced since 1976 are automatically placed in Class III, unless the sponsor successfully petitions the FDA to reclassify it as "substantially equivalent" to a device that was on the market before the amend- ments took effect. The substantial equivalence provision has provoked uncer- tainty, as the law did not specify if this equivalence referred to safety and eff~ca- cy, or to equivalence of the physical characteristics of a device. FDA regula- tions issued in 1986 state that devices with new intended uses require pre-mar- keting approval. Post-amendment devices with intended uses similar to those of pre-amendment devices may be found to be substantially equivalent only if the new technological features of a device can be shown not to decrease its safe- ty and efficacy (81~. This may be demonstrated through descriptive, perfor-

166 ANNETINE C. GELIJNS mance, and even clinical data. This is called a 5 lO(k) submission. If a device is found to be substantially equivalent, the manufacturer may rely on pre-market- ing notification. This route to the market is much more expeditious than the pre-market approval route, and the impression is that sponsors will attempt to change the design of devices accordingly. Indeed, a recent GAO report found that roughly 90 percent of medical devices reviewed by FDA were marketed through 510(k) review, while 10 percent underwent the full pre-marketing approval process (82~. To support a marketing approval decision, or in some instances a 510(k) sub- mission, a sponsor is required to conduct clinical studies. Clinical investiga- tions of devices are subject to the two basic elements governing clinical research in general: informed consent and institutional review. In comparison with drugs, however, IRBs play a more important role in device evaluations. They review all clinical device studies, decide if the device poses a "significant risk," and approve clinical studies for their institution.37 The IRB determines if a device poses a significant risk on the basis of an investigational plan. The plan includes a description of the device, the objectives and duration of the investigation, the investigational protocol, a risk analysis, monitoring proce- dures, and informed consent materials, and it also identifies all involved IRBs. If a device poses a significant risk, a request for an Investigational Device Exemption (IDE) is submitted to the FDA (83~. Such an IDE application con- tains the investigational plan, information on prior investigations, the manufac- turing process, and the amount to be charged for the investigational device. In comparison to the 1,250 pages of an average IND, the average size of an IDE is 150 pages. After an IDE has been approved clinical investigations can be initiated.38 Data from a random 10 percent sample of IDEs submitted between 1980 and 1986 indicate that most clinical evaluative studies are concentrated in a few product categories; ophthalmic, cardiovascular, and obstetrics/gynecology prod- ucts account for nearly 60 percent of all IDE investigations. The range of prod- ucts requiring an IDE, however, is increasing (84~. In contrast to pharmaceuticals, the final version of a medical device is often not created de nova; instead a device prototype is usually modified technically as a result of initial clinical testing (67~. The period of learning necessary before a device can be used properly and efficiently may be longer than with drugs. Therefore, according to SpiLker, clinical testing usually first involves an initial pilot stage during which the protoWpe's design and materials are further developed and tested. The main questions are whether the device produces the postulated effect in humans and whether it seems to be clinically useful. These evaluations usually are based upon non-formal experiments (see below for dis- cussion of the argument to randomize the first patient). In addition to clinical evaluations, this stage also involves technical testing; for example, the electrical and mechanical components of infusion pumps are subject to technical evalua-

COMPARING DEVELOPMENT OF TECHNOLOGY 167 lions, and a number of bench tests are performed to determine a pump's accura- cy and reliability. After the technical development has become more or less stabilized, a series of safety and efficacy evaluations of the final "initial" product can be initiated. This decision is sometimes a difficult one, as clinical evaluations usually reflect risks and benefits at a fixed point in time. Too early assessments may not reflect the true risks and benefits of an evolving device and the results of the study may be obsolete before the evaluation is completed, whereas the results of evaluations done too late in the life-cycle may be irrelevant for health care deci- sion makers. In medical device evaluations, a distinction needs to be made between diag- nostic and treatment devices. With the former, it is usually not direct patient benefit, but benefits in terms of clinical utility (i.e., its contribution to further diagnosis or therapy) that are to be evaluated. Fineberg has formulated a hierar- chy of criteria for diagnostic technology evaluations: technical capacity, diag- nostic accuracy, diagnostic and therapeutic impact, and patient outcomes (85~. Generally, evaluations provide information on the technical and diagnostic per- formance (not the more comprehensive clinical utility or patient outcomes) of a diagnostic device, and possibly on its risks and complications. The main mea- sures of diagnostic performance are sensitivity (ability of a test to detect disease when it is present) and specificity (ability of a test to correctly exclude disease when it is absent).39 In clinical practice, however, the question of interest is, if the patient has a positive test how likely is he or she to have a specific disease? (86) Therefore two additional measures, the predictive value of a positive test result (i.e., number of true positives/true positives plus false positives) and the predictive value of a negative test result (i.e., number of true negatives/true neg- atives plus false negatives) play an important role. These measures indicate the likelihood of the presence or absence of a disease in a tested individual from a given population with a particular prevalence of the disease. In order to com- pare the sensitivity and specificity of two or more diagnostic devices, the receiver operating characteristic (ROC) curve analysis can sometimes be used.40 However, ROC analyses require simple models and large numbers of cases, and Friedman observes that they are often very difficult to undertake (88~. Thus, most diagnostic tests are evaluated only in terms of their technical and diagnostic performance before marketing. Furthermore, according to Schwartz, the range of patients tested may be inadequate, as they usually involve those with advanced disease, and a few young healthy controls (89~. A diagnostic test, however, may not perform as well with patients with earlier dis- ease, which indicates the need for more comprehensive evaluations (60~. As with drugs, the question here concerns what endpoints should be evaluated in pre-approval and/or post-approval trials. Traditionally, most device evaluations lack randomized control groups (671. While this may in part be due to less sophistication in clinical research on the

168 ANNETINE C. GELIJNS part of many device manufacturers, it may also result from inherent characteris- tics of device development that make the classical RCT more difficult to per- form. The statutory standard recognizes this and is less rigorous than with regard to new drugs; i.e., safety and effectiveness information for devices may be provided through "well-controlled scientific studies" or through "valid scien- tific evidence." The randomized placebo-controlled, double-blind clinical trial, optimally suited to provide pre-marketing efficacy information on drugs and biologicals, indeed has more limitations with new devices. This holds especially for diagnostic but to a certain extent also for treatment devices; for example, a placebo may be unethical (as with heart valve replacements) or certain situa- tions may not be amenable to observing a placebo effect (as when the patient is unconscious or the interaction between patient and device is minimal). Another essential characteristic of RCTs as used in drug evaluations, patient and physi- cian blinding, may also cause more difficulties with devices. However, creative techniques to eliminate bias are emerging. For example, one physician may insert an implant device while another physician evaluates its benefits and risks. Thus if RCTs are possible at this stage their use or that of otherwise well-con- trolled study designs, such as parallel study designs or crossover designs, should be stimulated. Generally the sample sizes used in these clinical studies are con- siderably smaller than is the case with drugs. Ophthalmic IDEs, for example, called for an average of 280 patients, while all other IDEs involved about 150 patients (84~. During clinical studies, much industrial effort may be directed towards scal- ing-up for production. The necessary production capacity may vary widely, ranging from 10 to 100,000 devices a year. Depending on the kind of devices, specific manufacturing requirements may exist, such as the need for sterility or for a certain shelf life. Good Manufacturing Practice regulations41 govern the manufacturing process in general. International and national standards may also exert an important influence on the manufacturing process, for instance those set by the Association for the Advancement of Medical Instrumentation or the International Electrotechnical Commission (64~. On the basis of the results of clinical investigations, a device may be approved for marketing.42 In contrast to drug regulation, the device amend- ments require that advisory committees participate in the pre-marketing approval (PMA) decision for Class III devices (90~. In general, the PMA is an individual license to the developer for a particular device. Other developers of similar types of devices need to submit a separate PMA, with adequate clinical data. Data of previously approved PMAs cannot be used, unless they are pub- lished and generally accepted by the medical community. This policy protects each manufacturer's investment in the development process, but it also may stimulate the duplication of investigational efforts, including the performance of unnecessary trials. However, the next model of a medical device often differs in materials and/or design, and these differences may affect clinical risks and ben- efits. Recently proposed amendments to the 1976 medical device legislation

COMPARING DEVELOPMENT OF TECHNOLOGIES 169 would allow the FDA to waive data requirements for PMAs following that of the innovator. Adoption of these amendments could lessen the incentive for innovative R&D. Device Diffusion and Post-Marketing Surveillance An important decision point in the course of development concerns the adop- tion of a new device by physicians and hospitals, which is influenced by a com- plex set of medical, economic, regulatory, and social factors (11,91~. In the 1970s a number of health planning laws, such as Certificate of Need laws and rate regulation, were enacted to control the adoption of "big-ticket" devices. But after a decade some of the drawbacks of such planning laws surfaced, in part because the numbers of new devices expanded beyond the scope of regula- tion. At the same time, policy attention increasingly turned towards the pay- ment method as an important tool for influencing the adoption and use of devices (8~. Most industrialized countries are moving away from a cost-based, essentially open-ended reimbursement system to a prospective payment system (PPS). This transition has probably been most prominent in the United States with the establishment of Medicare's PPS for hospitals, based on diagnosis-related groups (DRGs). Under PPS hospitals have a strong financial incentive to pro- vide the least resource-intensive treatment. The system promotes a significantly lower level of growth in service intensity than traditionally has been the case, and the recalibration of DRGs is lagging behind changes in medical practice (92~. Although the price system is intended to be neutral under PPS, this is not always the case. For example, the lithotriptor was covered as a medical treat- ment for kidney stones under DRG 323. But this DRG pays only half as much as DRG 308 for the surgical treatment of kidney stones. Thus, although lithotriptors may improve the quality of care and may be cost-effective for some indications, hospitals have less financial incentive to invest in these machines (93,94~. Also, because PPS deals only with payments for inpatient hospital care, there is an incentive for hospitals to utilize technologies that are cost- effective over the short term of hospitalization. There is little incentive for hos- pitals to use technologies which have long-term benefits, even though they may ultimately have a greater impact on the efficiency of the system as a whole. As the existing reimbursement system affects the market for new medical products, changes in this system may exert strong feedback signals to the development process, e.g., it has been observed that medical device manufacturers react to the demand for products that are cost-effective over the short term and neglect R&D projects dealing with products that are cost-effective over the longer run.43 With these changes in reimbursement, the coverage decision by the Health Care Financing Administration has become a more important factor in the development process.44 Traditionally, the coverage decision making process

170 ANNETINE C. GELIJNS was based on generally subjective evidence provided by medical expert panels; increasingly, however, formal evidence becomes the basis for these decisions (95~. This evidence includes safety and efficacy considerations narrowly defined, but there is a tendency to consider the effect of devices on the quality of life of patients (including their preferences for certain outcomes) and their cost-effectiveness. As such, the change in decision making provides an impor- tant incentive to undertake evaluative studies after pre-marketing notification or FDA approval. Phase IV studies include a number of post-marketing surveillance mecha- nisms to detect adverse device reactions.45 The FDA maintains a Device Experience Network that receives reports on device hazards from health profes- sionals and manufacturers. Device manufacturers are required to keep records of complaints as part of GMP regulations. On the basis of adverse reaction reports, the FDA may require removal of designated devices from the market or restrict their sale or use. In comparison to drugs, acute injuries are probably more easily associated with a particular device. A major issue, which needs to be examined, is whether the adverse reaction or event is a consequence of the skill of the professional or inadequate maintenance of the device, or can be attributed to a defect in the device itself (96~. In addition to these surveillance mechanisms, a number of epidemiological methods may be used to detect possible risks of device use. As discussed above, the potential of using observational methods for risk detection is increas- ing. In addition, information on effectiveness is needed; such information can be provided by experimental or observational studies. Because the life cycle of a device is short and next-generation versions of a particular device may emerge relatively quickly (as with diagnostic pregnancy kits, for instance) the applica- bility of RCTs may be more limited. An advantage of using modern observa- tional data bases is that they represent continuous monitoring of the use of devices in practice and their outcomes. Uncertainty, however, remains as to the strengths and weaknesses of these methods in providing reliable evidence (97~. THE DEVELOPMENT OF CLINICAL PROCEDURES The last 25 to 30 years have seen rapid advances in basic biomedical research,46 strengthening the scientific underpinnings for the development of new clinical procedures in the years to come. A clinical procedure can be defined as any practice of a health practitioner that involves a combination of special skills or abilities and may require drugs, devices, or both. As clinical procedures involving new drugs or devices, such as laser angioplasty, have been considered above, this section will especially focus on those clinical procedures which are not to a large extent dependent on new health care products but on the technique of the provider performing the procedure. For example, the develop- ment of certain surgical procedures (although they may involve the use of scalpels, clamps, and drugs) or psychotherapy.

COMPARING DEVELOPMENT OF TECHNOLOGIES Clinical Procedure Development and Adoption 171 The development process of clinical procedures is very different from that of drugs and medical devices. Analytically, the distinction between the develop- ment of radical or breakthrough innovations and incremental innovations is use- ful. Radical innovations frequently arise in academic or academic-associated centers, where physical and professional resources are available and clinical development is stimulated. The development of incremental innovations usual- ly occurs in a much more decentralized fashion, involving numerous physicians refining and modifying an existing procedure in everyday clinical practice. In contrast to medical device innovation, which requires as C. P. Snow would say—the bridging of "two cultures" (that of engineers and that of clinical researchers), the distinction between "developers" and "evaluators/users" may be very fine or even non-existent in the development of clinical procedures. Within the hospital those involved in experimental medicine may be physically down the hall from their clinical colleagues, but often they are embodied in the same person. Physicians who treat patients may at the same time be engaged in the development of clinical procedures. This sometimes may lead to difficult conflicts of interest between the therapeutic and investigational role of a physi- cian. As Swazey and Fox (99) observe " . . . their double-edged role causes stress for most physician-investigators. The strains that they experience are intensified by their typically close and continuous relations with the patients who are also their subjects; by colleagues' scientific and ethical judgments of their work; and by a certain vested interest not only in protecting their profes- sional reputations, but also, in advancing them through recognition for being eminently successful with breakthroughs in knowledge or technique." In spite of the enthusiasm and fascination generated by potentially radical procedures, the initiation of first human application often remains inherently premature (particularly in the absence of a satisfactory animal model) (99~. Therefore this transition often is controversial, as recently illustrated by trans- plants of dopamine-producing cells into the brain region (in need of that specif- ic transmitter) of very severe Parkinson's disease patients. Sladek and Shoulson, in a review of the initial clinical application of this procedure in Science, argue strongly that although " . . . the scientific rationale continues to build for neural grafting as a therapy for neurological disease . . . we could ben- efit from more patience than patients (100~. Fox and Swazey, in their book The Courage to Fail, have described the scientific and emotional controversies that may arise during the development of clinical procedures such as kidney dialysis and transplantation. Their work indicates that radical innovations usually are first applied to life-threatening or very serious diseases, which often have no alternative treatment (101~. In these cases the considerable uncertainty, and potential risks, associated with the clinical application of the innovative proce- dure may be considered more acceptable. Their analysis also indicates that during their development procedures may

172 ANNETlNE C. GELIJNS often be subject to a partial or complete "clinical moratorium," i.e., human use of a still experimental procedure on patients is suspended (991. For example, mitral valve operations have been performed on animals since the turn of this century. The first application to humans occurred in 1923, but a clinical mora- torium was invoked in 1928, in part due to the high mortality associated with the procedure. Following a series of drug, device, and surgical advances such as those in cardiac catherization, anesthetic techniques for intrathoracic surgery, ligation of the patent ductus, and antibiotic drugs, the clinical development of mitral valve surgery was resumed in 1945 (despite initially high mortality rates). Over time, as surgical experience increased and different patient groups were accepted, mortality declined and the technique became established. Comroe and Dripps have equally underlined how the development process of procedures for cardiovascular-pulmonary medicine depended on numerous advances in differ- ent areas of science and technology (1021. In contrast to drugs or devices, no formal governmental regulatory system exists for the development and evaluation of clinical procedures. Their devel- opment has traditionally been placed in the context of the physician's clinical autonomy and the trust relationship between patients and physicians. Evaluation of these procedures during development therefore depends heavily on professional self-regulation (for instance, through peer review and IRBs).47 In this respect, the difference between radical and incremental innovations may also be of importance. In the case of incremental innovations, the line between experiment and individualized therapy often is difficult to draw clearly (103), and IRBs are usually not approached to give their approval for the evaluation of slight modifications of existing procedures. This is different regarding radical innovations, and their development and evaluation (at least for those that are federally funded) is generally subject to the approval of IRBs. IRBs, however, do not usually conduct in-depth examinations of the research design (1041. To date, the potential safety, efficacy, and effectiveness of many procedures have not been evaluated systematically during their development. Surgical techniques in the first half of this century were developed by pioneering sur- geons on the basis of their intuition and insight, and were tested by trial and error. Many of these procedures attained acceptance in the medical community and resulted over time in useful treatments. A number were discarded, howev- er, often after years of clinical application, such as surgery for constipation. According to Barnes, this pattern of development is due to a number of factors (1051. Historically, there was often a poor understanding of disease processes and an uncritical acceptance of established dogma as dictated by leaders in the field. In addition, the analytical underpinnings of clinical investigations, in terms of sample bias, observer objectivity, or standards for adequate follow-up, were often still rather weak. As Bunker et al. conclude in their important work on the costs, risks, and benefits of surgery: "In this respect, surgery shared with other branches of medicine at the time a process for groping for effective thera- pies, a process that did not have the help of extensive knowledge in the basic

COMPARING DEVELOPMENT OF TECHNOLOGIES 173 biological sciences or the understanding of sophisticated experimental designs to permit logical inductions from multivariate clinical circumstances" (106~. In the second half of this center, rapid advances were made in the method- ological underpinnings of clinical investigations. At the end of the 1970s, how- ever, Bunker, Hinkley, and McDermott conclude that surgical development was still often based on inadequate evaluation (107~. Examples of procedures that diffused into health care and only later were to be found ineffective for treating certain conditions include prefrontal lobotomies for schizophrenia, colectomies for epilepsy, and more recently, EC/IC bypass surgery to prevent stroke. In a recent article Eddy and Billings provide an extensive argument for the often weak evidence underlying a number of important present-day clinical proce- dures (108~. The Use of Controlled Clinical Studies According to Wennberg, many procedures have not received careful feasibil- ity studies during their initial application in humans (109), but have been intro- duced on the basis of investigations involving historical controls or more anec- dotal evidence. Generally, the results of such investigations tend to be more optimistic regarding the benefits of a new procedure (110~. On the basis of such optimism and a complex set of sociological, economic, and scientific factors a procedure then may diffuse into more widespread use. Over time, uncertainty regarding the risks and benefits of a procedure, as used in specific patient groups and for various indications, may increase and clinical trials may then be undertaken. At that point in time, however, the acceptance of the trial results has become inherently difficult as an advocate group for a procedure generally has been created.48 Chalmers, therefore, has proposed to "randomize the first patient" receiving a new procedure (114~.49 This proposal has not received wide acceptance, because during the initial stage the practitioner's skills and expertise with a pro- cedure still evolve and the risks and benefits associated with the procedure may change considerably. In view of this "learning curve" phenomenon, the initial application of a new procedure will probably need to involve methodologically sound non-formal experimental studies.50 Such early careful and comprehen- sive reporting of clinical experience may form the basis for the design of subse- quent RCTs, if necessary, or of otherwise well-controlled trials to determine a procedure's efficacy and safety. The above does raise the question of the timing of these studies; when exact- ly in the development process should RCTs or otherwise well-controlled studies be undertaken? If an RCT is undertaken too early, the results may be obsolete before the trial is finished. For example, 15 years ago a randomized trial was initiated to compare the Ginsberg procedure with medical treatment for coro- nary artery disease. Two years later the trial was abandoned because the tunnel implant had been replaced by coronary artery bypass grafting (116~. If an RCT

174 ANNETINE C. GELIJNS is delayed, however, a constituency for the procedure may have formed. Bunker et al. therefore suggested the establishment of a reviewing authority to initiate and coordinate such trials as appropriate (106~. With regard to RCTs, one should bear in mind that some real conceptual, practical, and ethical difficulties may exist regarding their use in the develop- ment of new clinical procedures (117,118~. Double blinding, for instance, is more difficult to achieve. One possible solution may be to have one physician perform the procedure while another evaluates its effects. Controls may include standard accepted surgery or alternative treatments involving drugs or devices; it is generally accepted today that use of sham-operations is unethical.51 Surgical procedures will also depend much more strongly on the technical skills of the surgeon, who might be better at one type of surgery than another. Van der Linden (27) suggested that patients should be randomized to different surgeons who would perform the surgery they do best. Furthermore, if alternative treat- ment modalities are being developed with the aim of improving quality of life, while the different interventions are associated with variable risks and benefits, randomization may be considered unethical. As Relman notes, from the patient's point of view, surgical and medical therapy are not simply comparable arms of a clinical trial. They are vastly different treatments with very different personal consequences (113~. In these cases, Wennberg has argued that assign- ment according to patient preferences may be the ethically necessary choice. This would require systematic analysis of how patients value different Apes of health outcomes (an understanding that today is not yet available) and an in- depth examination of how one will be able to understand the "biases" associated with actual patient choice. Post-Marketing Surveillance Finally, as argued above, the full range of information on the effectiveness and safety of a procedure may not emerge in randomized clinical trials, as these trials may exclude a spectrum of at-risk patients. For example, Hlatky et al. (120) compared the patient population in their cardiovascular disease data bank with the patients enrolled in some large RCTs of coronary artery surgery. They found that only 8 percent of their patients met the eligibility criteria for the European Cooperative Surgery Study, 13 percent met the criteria for the large Veterans Administration study, and 4 percent met those for the Coronary Artery Surgery Study. This indicates that the trial results may not always form a suff~- cient basis for clinical practice decision making. Therefore, following randomized or otherwise well-controlled efficacy and safety trials, long-term surveillance should be undertaken of the safety and effectiveness of new procedures as they are used in everyday clinical practice. These studies may involve experimental or observational methods. In view of some of the logistical problems involved, it may be especially useful to depend on modern observational methods that enable one to monitor clinical practice

COMPARING DEVELOPMENT OF TECHNOLOGY 175 and changes in health outcomes. In recent years the use of such observational studies for assessing outcomes of clinical procedures has increased. For exam- ple, Wennberg et al. (57) and Roos et al. (121) have used claims data to evaluate health outcomes following prostatectomy, hysterectomy, and cholecystectomy. Given the increased availability of computerized data banks, the possibilities of inexpensive monitoring are appealing. A more extensive examination of the advantages (such as lowered costs, ease of patient follow-up over long periods of time, and the absence of reporting bias) and the disadvantages (such as ade- quacy of the data for case-severity adjustment and lack of outcome information on quality of life and functional status) is needed. IMPROVING THE TRANSLATION OF RESEARCH FINDINGS INTO CLINICAL PRACTICE: SOME OPPORTUNITIES FOR CHANGE The increase in knowledge concerning human health and the mechanisms of disease has been so rapid during the second half of this century that the present era has been described as that of the biological revolution. This biological revo- lution may prove as decisive for the future of medicine as the industrial revolu- tion was for economic development in the past (122~. The extent to which this occurs, however, depends in part on the effectiveness and efficiency of the pro- cess by which advances in biomedical research are translated into clinical prac- tice. As indicated earlier, in medicine this translation or the development pro- cess includes three components: the development of new drugs, of medical devices, and of clinical procedures. This paper describes the similarities and differences that exist among the development processes for drugs, medical devices, and clinical procedures. A primary difference concerns the asymmetry of the evaluative strategies employed: over the last quarter century drugs have been subjected to rigorous clinical testing before their introduction into general use, while clinical proce- dures are still being assessed only in a more ad hoc fashion, and new medical device evaluations are to be found somewhere in between. It might be expected that this asymmetry reflects important differences in the effectiveness and eff~- ciency of the three different processes by which research findings are translated into clinical practice. Following are some major observations with regard to these differences, and some inferences as to opportunities for improvement. The Development of Drugs In comparison to the medical device industry, the multinational pharmaceuti- cal industry is older, highly regulated, and very research-intensive. The phar- maceutical industry annually invests approximately $6.5 billion in R&D in the United States (about 17 percent of sales52), roughly $1.5 billion of which goes to pre-marketing clinical testing. The investments in research, but especially those in development, are consistently increasing (since 1980, an increase of

176 15 14 13 12 11 i,, 10 9 is O 8 111 7 m is _ ~ Europe _ ~ _ ~ 5 ~ U.S. 4 ~ e ~ 3 Japan 2 it/ e. \ \ \ . . ·. 1 1964 1966 1968 1970 ANNETINE C. GELIJNS A\ . a _' 'I _/ . . . 1972 1974 1976 YEAR OF IND FILING 1978 1980 1982 FIGURE A.3 Origin of NCEs on which INDs have been filed by U.S.-owned Finns. SOURCE: Mattison N. Trouble AG, Lasagna L. New drug development in the United States, 1963 through 1984.ClinicalPharmacology and Therapeutics1988;43:290-301. roughly 30 percent). According to the Wiggins analysis, the R&D process is -Jo- ~ -- or- ~~ = ~ ~~ - , ~ ~ estimated to now cost ~l;~3 mllllon per marketed new chemical entity (NUb) (123). On the input side, the resource commitments to drug R&D, the relevant sci- entific knowledge, and the technical capabilities have all grown impressively since the 1950s, but this growth has not been reflected on the output side, at least not quantitatively (124~. The number of NCEs entering human testing fell from a mean of 89 a year in the period 1950-1962, to 35 a year in 1963-1972, to 17 a year in the period 1975-1979 (an overall reduction of 81 percent) (30, 31,125~. In recent years, IND filings in the United States are increasing again, especially those for biological drugs (126~.53 But these INDs are increasingly acquired from non-U.S. sources. Especially noteworthy is the fact that Japan has been increasing as a source since the early 1970s, by the end of the 1970s surpassing Europe, traditionally a stronghold for producing new chemical com- pounds (see Figure A.3~. A similar trend occurs with regard to the number of new drug approvals.54 Over the years these output measures of the development process have been extensively reviewed in the literature and the halls of Congress (see Addendum). The dates given above indicate that the beginning of the decline in

COMPARING DEVELOPMENT OF TECHNOLOGIES 177 the number of new U.S. drug introductions occurred at roughly the same time as the introduction of the 1962 amendments to the Food, Drug, and Cosmetics Act. A substantial body of policy analysis was undertaken to consider the causal effect of these regulatory changes on the declining number of new drug approvals. Originally it was concluded that the decline in drug introductions could be fully attributed to changes in regulatory requirements for evaluative practice during development. Currently, however, in view of the increasing recognition of social, economic, managerial, and political factors as determi- nants of the decline, it is apparent that no such straightforward link can be established (127~. Nonetheless, regulatory requirements and their interpreta- tion by the regulatory agencies concerned remain an important factor in the potential rise and fall of new drugs, not to mention the scientific, commercial, and public perceptions of such regulations as determinants of whether and how a drug is developed. Generally speaking, these regulatory configurations and the resulting clinical evaluations have led to important benefits.55 Under social and political pres- sures, these requirements have become increasingly detailed over time. As a result the time-span of pre-marketing development has increased from about 4.5 years in 1964 to 9 years in 1984 (311. This interval has reached a point where access to useful new drugs may be delayed. The tension between increasingly thorough pre-marketing evaluations and early availability becomes urgent in the case of life-threatening disease. For example, the prominence of AIDS has raised two fundamental issues as to the clinical basis on which decisions are being made. The first concerns the endpoint issue mentioned earlier, i.e., considerable uncertainty exists as to what endpoints should be evaluated during which stage of the development process. For instance, to expand on the AIDS example, the question concerns whether and in which cases intermediate endpoints (instead of survival) should be evalu- ated in pre-marketing trials. Equally, the question concerns whether and when quality of life endpoints should be built into the developmental evaluation pro- cess. The second issue concerns the balance between pre- and post-marketing evaluation, regarding which some new initiatives have recently materialized. The FDA, for example, has proposed to streamline the drug approval process for life-threatening diseases by shortening the pre-marketing evaluation stage (Phase II and Phase III clinical trials will be merged into more definitive Phase II trials), and by emphasizing more strongly the post-marketing evaluation stage (Phase IV) for providing safety and effectiveness information on a new drug. Even apart from life-threatening diseases, there is a general need for such Phase IV information, because the full range of a drug's risks and benefits will emerge only when it is used in actual circumstances of clinical practice. Drugs, once marketed, are subject to empirical innovation—just like devices or clinical pro- cedures. That is, in the hands of physicians trying to solve problems, new theo- ries are spun out and drugs are used as if those theories were true (e.g., cimeti-

178 ANNETIlIE C. GELIJNS dine). It is only through Phase IV monitoring and surveillance broadly con- strued (i.e., regarding general use) that the identification of these theories can be accelerated and steps taken to assure their timely testing. As argued above, the Phase IV studies that would provide this information will depend heavily on observational methods. In recent years, methodological advances (see below) have opened up new opportunities for inexpensively mon- itoring the use and long-term risks of drugs. These methods may well be useful, not only in providing risk information, but also in providing effectiveness infor- mation. So far, however, uncertainty prevails as to the scientific value of the practical application of these methods to medicine in general and drugs in par- ticular. In view of the potential effects of these methods on shifts between "development" phases and the subsequent implications for medical innovation in general, any serious investment strategy for medical technology development must address the possible promise of such an application. A broader argument exists to carefully consider the potential and problems of post-marketing evaluation.56 The increasing time-span of development has not only made the process more costly but also has decreased the return on invest- ments by lowering the effective patent life of new pharmaceuticals. The aver- age patent life of NCEs, from date of approval to expiration of the patent, was 16.3 years in 1960 and roughly 9 years in the mid 1980s.57 The need to consid- er patent life was recognized in the United States, and the Drug Price Competition and Patent Term Restoration Act was enacted in 1984. The law, however, restores only some of the patent life lost during the regulatory process, as it also allows generic drugs to receive more speedy marketing approval through a system of abbreviated NDAs.58 The overall result of this law is that the effective exclusive marketing time of innovative products has not increased because generic drugs can be marketed more rapidly. The impact of this law may be considerable since 81 of the most important 100 drugs used currently in the United States will go off patent by 1991, and will thus become generic drugs.59 Furthermore, the economic climate is becoming much more price competitive, e.g., most states have passed pro-generic substitution laws allow- ing pharmacists to dispense generic drugs for the brands specified on the pre- scription forms.60 Whereas industries other than the pharmaceutical, such as electronics or optics but also the medical device industry, can react to more competitive environments by decreasing the turn-around time of their innova- tive cycles, such a strategy will be much more difficult in a pharmaceutical industry subject to long and relatively fixed R&D cycles. Although the pharmaceutical industry has generally been very profitable and recent advances in biomedical research seem to present exciting opportunities for the development of new drugs, the trends visualized in Figure A.4 may con- stitute an impediment to drug development in the long run. Whereas the effec- tive translation of research findings into clinical practice will require informa- tion on the health outcomes of a drug in general use, the above underlines the necessity to provide this information as efficiently as possible.

179 En a) In a) ._ o ._ _ ._ a) Q O . it. _ , o 1 o o o Con ~ To ~ =8 o ~ COM Oi<D ~ ~ 1- . ~ i= CM o . ~~ ~ O -0 E a) ~ <: ~ a) c ~ _~ is/ a) ._ In O CD c of o tin ~ ,~ ;, 0 . CO . . Do D Cal ._ o . . o CQ . o on A> o 4— Cal .= .s ._ U) o Pa lo:

180 ANNETINE C. GELIJNS The Development of Medical Devices Moving from the pharmaceutical to the medical device industry, the latter is younger, more nationally oriented, and characterized by smaller firms; the image of the innovative engineer developing a device prototype in a basement, garage, or study still has some relevance, although it may become a metaphor in the 1990s. In recent years the market for medical devices has been growing rapidly. The overall U.S. medical device market is estimated at over $20 billion dollars in 1987, and parts of that market are expanding at annual rates ranging from 10 to 25 percent (63~. Investment in medical devices R&D is smaller than for pharmaceuticals. On average, medical device funs invested 7.5 percent of sales in R&D in 1988 (65), and it appears that this percentage remains relatively stable. Differences in R&D investment can be observed to depend on the size of the few, and the particular type of device under development. For example, small firms invest almost double the industry average (130~. In comparison with drugs and biologicals, there is a much greater hetero- geneity in medical devices in terms of design, purpose, and use, and conse- quently much more variation in the kind of clinical evaluations undertaken (67~. In view of this heterogeneity, the medical device amendments to the Food, Drug, and Cosmetics Act divided devices into three classes and differentiate the level of regulatory control according to the likelihood of risks inherent in a par- ticular device class. Whereas 90 percent of all new medical devices are subject to "general controls" (including good manufacturing practices, pre-marketing notification, and, potentially, technical performance standards), only 10 percent of new medical devices are subject to full pre-marketing review for safety and efficacy (82~. It is interesting to observe that, unlike the Drug Act, the device amendments explicitly incorporate in their mandate the need to encourage med- ical device development.61 In contrast to pharmaceutical innovation, there is a steady growth in the num- ber of medical devices entering clinical investigation (84~. In the past few years, IDEs62 include an increasingly broad range of investigational devices. But, since 1980, the proportion of IDE devices that in the first instance successfully com- pleted their developmental evaluation has decreased to 30 percent (although eventually 80 percent are approved). This decrease reflects the more complicat- ed safety and efficacy issues surrounding new investigational devices, a more stringent regulatory climate, and the relative inexperience of many device manu- facturers (80 percent of device developers have only submitted one IDE since 1980~. In terms of efficiency and effectiveness of development, those who had submitted seven or more IDEs since 1980 had an approval rate twice as high as those who had submitted only one IDE (84~. The impression exists that the R&D cycle of incremental device innovations is only two to three years, whereas with radical device innovations (such as ultrasound or magnetic resonance imaging) it is more like 10 years. Development times can then be estimated to range rough- ly from one year for incremental devices to five years for radical devices.

COMPARING DEVELOPMENT OF TECHNOLOGIES 181 Unlike drugs, a medical device is generally not created "de nova" but arises in a development process typically representing continuous technical modifica- tion and incremental improvement. Clinicians provide significant input by eval- uating the clinical efficacy and safety of a device as well as by suggesting tech- nical improvements to enhance its clinical utility. For certain devices the users are also the innovators, designing and building the original prototype (69~. Although already fairly common, close interaction between device manufactur- ers and the clinical community is an even more crucial prerequisite for effective and efficient device innovation than is the case with pharmaceutical innovation. The initial evaluation stage of a new device usually depends on careful clini- cal observations based on informal experimental methods. The main question at this stage concerns whether the device prototype has the postulated effect in humans and may be clinically useful. As indicated above, this pilot stage then normally leads to improvements in a device's design and materials. Chalmers (114) has argued for randomizing the first patient who receives a procedure involving a new medical device (or a surgical technique). In view of the techni- cally evolving nature of the device and the fact that investigators usually must be educated and trained in how to operate certain equipment, randomization of the first patient is generally not considered feasible (this equally holds for surgi- cal techniques). This then does raise the question of the timing of (a multicenter stage of) safety and efficacy evaluations, and when exactly in the development process these evaluations should be initiated. If these studies are undertaken too early, the changing characteristics of the device may render the results obsolete. If they are undertaken too late, the results may be unimportant for decision making. Ideally, these studies would be based on randomized controlled trials. However, in many cases, the RCT as generally used in drug studies (involving double blinding and placebo controls) will be much more difficult to undertake, especially with diagnostic devices. In such cases other well-controlled study designs will have to be used to evaluate efficacy. Until recently, however, new device evaluations often are uncontrolled (67), and the use of adequate controls needs to be stimulated. Once the device has been approved and diffuses into more general practice, its long-term safety and effectiveness should remain parameters of the devices's long-term development evaluation. Because device development often involves incremental innovation for a considerable part of its lifespan, and because RCTs are not ideally suited to provide information on a slightly different version of a device, these studies will usually depend on obser- vational methods. When comparing the rationality and efficiency of device development to drug development and considering the well-known methodological weaknesses of traditional observational methods, it is timely to assess the strengths and weaknesses of new non-experimental methods for providing reliable informa- tion about the health effects of new medical devices. In the quest for improved and more reliable methods of clinical device evaluation, however, it is necessary to consider the importance of small device firms in medical device innovation.

182 ANNETINE C. GELIJNS One will need to keep in mind the potentially differential impact such require- ments could have on small versus large firms in terms of viability, innovation potential, competitiveness, etc. The Development of Clinical Procedures As discussed above, the distinction between "developers" and "evaluators/users" is a thin line in the development of clinical procedures. In comparison to drugs and devices, no governmental regulatory system governs their development. The evalua- tion of developing clinical procedures is based in principle on the trust relationship between physicians and patients. Ihitiation of development and its evaluation thus depends heavily on professional self-regulation. In this respect the difference between radical and incremental innovations is important. Radical innovations may frequently originate in academia or academic-associated centers, and are generally subject to approval by IRBs. A large part of developmental efforts, however, concern incremental improvements in existing procedures. In cases of incremental improvement the line between experiment and individualized thera- py generally is difficult to draw clearly, with the result that IRBs are not usually approached for approval. There is very little information available on investments in R&D for clinical procedures. It appears, however, that considerable changes may be taking place as to the source of funding in this area. For example, whereas traditionally the development of many procedures was cross-subsidized through patient care rev- enues, with the changes in hospital reimbursement these funds are decreasing. Very little information is also available on the aggregate number of new proce- dures being developed as well as on the average time needed for development. During the development process new clinical procedures generally have not been systematically evaluated in terms of safety, efficacy, and effectiveness. Traditionally, their evaluation during the development process often depended on non-formal evidence or the use of historical controls; this usually leads to more optimistic results as to the potential benefits of a new procedure than would have been the case from well-controlled studies (1101. As a result, the scientific evidence normally assumed to support day-to-day clinical practice is not always provided in a systematic and timely fashion. For example, a number of procedures were discarded only following their widespread use, when they were found to be ineffective on the basis of well-controlled studies. For some of these procedures, the weak quality of their clinical evidence is illustrated by considerable geographic variations in their use, such as those for coronary artery surgery, hip replacements, or lower back surgery. To achieve an effective development process for procedures, more systematic and improved evaluative strategies are needed. Such an improvement will need to take into account that the development of clinical procedures is a very different endeavor from that of drugs and devices. The development of especially incremental innovations often occurs in a decen-

COMPARING DEVELOPMENT OF TECHNOLOGY 183 tralized manner, involving change and refinement of a particular procedure by numerous physicians. During the initial development of a new procedure, the skills and experience with a technique continue evolving and the risk/benefit ratio may change considerably. Pilot or feasibility studies at this stage will have to include "systematic and comprehensive collection of clinical experience" to determine whether a procedure works and to differentiate patients according to prognostic factors (107~. However, many clinical procedures do not now receive the careful Phase I studies required for drugs (109~. If and when the feasibility of a new procedure has been established, randomized clinical trials or otherwise well-controlled studies should be undertaken at selected institutions. The transition from the feasibility study by a few developers to multi-center investigation is more difficult to determine with clinical procedures than is the case with drugs and even devices. Systematic surveillance of early clinical evi- dence could facilitate the timely implementation of such well-controlled studies. After efficacy has been determined under such trials, evaluation of a proce- dure's effectiveness and safety should again remain parameters of development evaluation. Comparing the Outcomes of Drugs, Devices, and Clinical Procedures Ultimately clinicians and patients, of course, are concerned with choices among a spectrum of alternative diagnostic and treatment technologies and want to know, for a clinical condition, which treatment is best for which patient. The rational assessment of technology thus requires a balanced strategy for assess- ment that provides comparable information about relevant outcomes for all rele- vant technological options. This chapter has already noted the present imbal- ance in regulatory assessment strategies which provide extensive documentation of (at least some) outcomes for drugs compared to other drugs or to placebos, while little attention is given to understanding the relative merits of drugs com- pared to devices or to clinical procedures. The treatment of angina, gallstones, or prostatism are examples where all three types of technology have been devel- oped, but have not as yet received comprehensive and ongoing evaluation. This chapter does not intend to imply the need for a federal regulatory sys- tem governing the development of procedures. Alternatives to such a system have been proposed. Bunker et al. have suggested the establishment of a central reviewing authority (under which the various IRBs could resort) to initiate and coordinate clinical procedure trials as appropriate (1071. The initiation and coordination of studies determining effectiveness and (long-term) safety of a procedure would also be part of such an authority's mandate. The Bunker model does not, however, call for the systematic comparison of all technological options (including drugs and devices). A more recent model may be found in the assessment teams which have been established by the National Center for Health Services Research (now the Agency for Health Care Research and Policy) to evaluate alternative technological options available in

184 ANNETINE C. GELIJNS TABLE A.1 Comparison of rationality/eff~ciency of technology development Clinical Drugs Devices yes R&D investment +++ + ? Development time ++ + ? Number of new innovations entering health care + ++ ? Clinical basis for decision making: pre-diffusion +++ ++ + post-diffusion + + the management of clinical conditions. These teams are to undertake the equiv- alent of Phase I and Early Phase II studies now undertaken for drugs, make rec- ommendations for clinical trials (Phase III), and conduct Phase IV studies for new as well as established clinical procedures. These teams will focus on spe- cific clinical conditions, such as benign hyperplasia of the prostate and stable angina. They will assess all relevant treatments and thus provide information on the relative safety and effectiveness of drugs, devices, and procedures. Drug and device manufacturers could be expected to have interest in helping to fund these teams. Whereas—on the positive side this scenario would imply that the stronger financial sectors of our health care system would share the finan- cial burden of performing evaluations of clinical procedures, their involvement could result in possible conflicts of interest. This policy question will need to be addressed if the assessment team approach is to prove a realistic mechanism for the systematic evaluation of alternative medical technologies. In conclusion, serious inconsistencies exist in the evaluation of drugs, devices, and procedures during their development process. The above indicates that these inconsistencies may have contributed to shortcomings in the effec- tiveness and efficiency by which biomedical research findings and clinical theo- ries are translated into clinical science and useful clinical practice (see Table Add. Furthermore, these inconsistencies may also have contributed to unneces- sary health care costs, if one takes into account that the least systematically evaluated technologies, clinical procedures, are also the most costly.63 Although these inconsistencies are to a certain extent the result of inherent dif- ferences among the development processes of drugs, devices, and procedures, these differences do not seem to preclude a more balanced approach to assess- ing all medical technologies. Such an approach would strengthen the clinical evidence on which development decisions are made, and probably would improve the cost-effective use of health care resources.

COMPARING DEVELOPMENT OF TECHNOLOGY 185 This chapter concludes that to achieve a proper balance three issues can be identified that need to be addressed. The prst issue concerns the criteria or end- points of development evaluation. With regard to determining a technology's safety and efficacy, the role of intermediate endpoints in comparison to mortali- ty, functional status, or quality-of-life endpoints should be clarified. In addition to clinical and scientific considerations, the endpoint issue raises economic con- cerns; i.e., these decisions may have large consequences for the length and costs of pre-marketing development. Both these considerations would need to be taken into account. Furthermore, following the approval decision for new drugs and devices and the more widespread diffusion of new procedures, it will be increasingly important to include health outcomes in "real world" clinical prac- tice as important evaluative endpoints (n.b. for diagnostic technologies this may be inappropriate). In view of the increasing numbers of alternative or compet- ing technologies being developed, it seems especially important to provide com- parative evaluations of the relative safety and effectiveness of technological options available in the management of clinical conditions. Inherent in these evaluations would be the need to incorporate patient preferences for the health benefits and risks associated with alternative technological interventions. The second issue concerns the methods for providing such information. Evaluation of the risks and benefits of new technologies during their develop- ment will have to rely not only on experimental methods (including randomized controlled clinical trials), but also on improved observational methods of clini- cal evaluation. This applies for devices and especially clinical procedures, but also to drugs; for example, these kinds of studies can provide needed informa- tion on the long-term health outcomes of drugs in everyday use. In comparison to RCTs, these observational methods are usually considered to be the weaker methods of clinical evaluation. However, recent methodologi- cal advances may have addressed some of these weaknesses. It has been observed that (109~: 1. Advances in statistical methods, for instance those in Bayesian statistics, make it possible to assess outcomes for alternative treatment strategies. These methods are useful for assessing outcomes in non-experimental study designs. 2. The increased availability of large-scale automated data systems and improved methods of data base linkage make it possible to inexpensively moni- tor use and outcomes. 3. Advances have occurred in measuring the effects of a new technology on functional status and the quality of life of patients. 4. Advances in decision analysis provide means to assess the importance of patient preferences and of the uncertainties about the probability for specific health outcomes. In view of these advances, it seems especially timely to explore the strengths and weaknesses of modern evaluative methods within the wider context of existing methodologies.

186 ANNETINE C. GELIJNS Third, depending on their strengths and weaknesses, a policy and an institu- tional framework will have to be established for assuring the application of non- experimental methods as appropriate. It is only by addressing these complicat- ed issues that we will be able to improve the effective and efficient transfer of research findings into clinical practice, and thereby strengthen a crucial link in the medical innovation chain. ADDENDUM The major decline in the number of new U.S. drug introductions occurred at roughly the same time as the introduction of the 1962 amendments to the Food, Drug, and Cosmetic Act. A substantial body of policy analysis was undertaken to consider the effect of these regulatory changes on the number of new drug approvals. The early literature, however, has some major weaknesses. One of the initial studies by Peltzman (131), for example, indicated that all of the dif- ferences in introduction rates between the 1960s and the 1950s could be attributed to the effects of regulation. One major weakness of his model is that it assumes that new drugs are supplied at a constant rate, and therefore changes in supply side factors that would cause the introduction rate to fall are not incor- porated. Using a supply side model (a production function approach), Baily subsequently argues that introductions are a function not only of regulation, but also of industry research expenditures and research opportunities (132~. He concludes that regulatory requirements have significantly decreased new drug introductions. The measures used to examine the effect of regulation and research opportunities, however, are not very refined.64 Wiggins (134), in a careful analysis of the subject, subsequently argues that to determine the specif- ic influence of regulatory factors versus non-regulatory factors on the develop- ment process one should desegregate the new drug approval data according to therapeutic class. These data indicate that there were changes over time not only in the numbers of INDs filed, but also in the pharmacological types of NCEs entering human testing. If one compares the mid 1960s with the early 1980s, for example, the number of anti-infective and psychopharmacological drugs decreased markedly, while cardiovascular drugs initially decreased some- what and then increased again, and antineoplastic and gastrointestinal drugs increased steadily.65 The primary source of the overall decline can be found in psychopharmacological drugs, especially tranquilizers, and in anti-infectives. The question then arises whether these categories were more stringently regulat- ed than other categories. According to Wiggins,66 it appears that these cate- gories were not regulated more stringently, and thus non-regulatory factors must have also played a major role. Peter Temin specifies the argument as follows (127~. He underlines the fact that by far the largest decline can be found in the area of tranquilizer drugs. In addition to non-regulatory factors (such as the strong patents held in this area), he asserts that "the thalidomide tragedy was the proximate cause in the decline, acting quickly through its effects on the direc-

COMPARING DEVELOPMENT OF TECHNOLOGIES 187 lion of drug industry research and more slowly through the governmental regu- latory process." In short, this literature does not specify the exact effect of changing regulatory requirements on the decline, but does demonstrate that reg- ulatory requirements are an important factor in detennining how, and whether, a drug is developed. At the same time, a number of studies (41,135,136,137) have approached this issue from a different angle. They have compared the number of drug introductions in the United States with other European countries, most notably the United Kingdom. Their results indicated that on average more NCEs were introduced in, for example, the U.K. market than in the U.S. market. Furthermore, of the drugs introduced to both U.S. and European markets, most drugs were first introduced in Europe. This phenomenon has been described as the "drug lag." Most of these analyses, however, focus on the 1960s and early 1970s, while these differences have diminished since the early 1970s. While the differences in market withdrawals were never anywhere near so marked as those of new drug introductions, the withdrawal rates also converged over time. For example, between 1964 and 1983, eight drugs in the United Kingdom and five in the United States were discontinued due to safety reasons, while after 1974 the discontinuations in both countries are similar (138,1391. Although international regulatory differences thus seem to diminish, a recent analysis by Berlin and Jonsson demonstrates that the Scandinavian countries and the United States are among the more stringent regulatory approval sys- tems, both in terms of the length of the review times as well as in dates of mar- keting approval (1401. NOTES The characterization of a human being as a tool-making animal should be quali- fied. Animal species have been found to use a wide variety of tools, although as far as we know no animal species exists capable of handling fire in essence one of the first human technologies to its own benefit. A fine distinction between human beings and animals in this respect may be the human ability to use tools to make tools, and to communicate from one human to another the knowledge of how to develop them. 2. Kay's "flying shuttle" in the textile industries was one of the early instances of a machine replacing human labor with technological labor and introduction of the economic concept of work-without-workers. 3. Rosenberg: "In a fundamental sense, the history of technical progress is insepara- ble from the history of civilization itself, dealing as it does with human efforts to raise productivity under an extremely diverse range of environmental conditions" (31. 4. Within the development process, clinical investigation is essentially initiated with the first testing of a potential innovation in humans. In the development process of drugs and biologics these initial studies in humans have been designated Phase I studies, which are generally followed by Phase II and Phase m clinical studies before a drug or biological can be marketed. Phase IV studies, conducted after an innovation diffuses into more widespread use, may reveal important information

188 ANNETINE C. GELIJNS on the (cost-) effectiveness and (long-term) safety of an innovation, which subse- quently may be an important impetus for further developmental activities. In this paper we will also apply the terms Phase I to Phase IV clinical studies to the development of devices and clinical procedures. 5. The federal Food, Drug, and Cosmetics (FD&C) Act of 1938, which provided for the premarket clearance of new drugs to ensure their safety, in its amended form still governs drug development today. In comparison, biologicals are governed by a separate law, the Public Health Service Act of 1944. Major changes to the FD&C act were provided by the 1962 Kefauver-Harris amendments, which increased the role of the Center for Drugs and Biologics of the FDA in the devel- opment process. The medical device amendments were enacted in 1976 and are implemented by the Center for Devices and Radiological Health. 6. Nelson and Winter have developed a theoretical structure, incorporating both "uncertainty" and "institutional structure" as essential elements of technology development (10~. 7. The decision making processes for individual or organizational adopters of tech- nological innovations vary greatly (11~. 8. A powerful incentive for industrial collaboration with federal laboratories, such as the National Institutes of Health, in R&D projects was provided by the federal Technology Transfer Act of 1986. 9. The following examples would come under this category: "all the main classes of psychotherapeutic drugs (tranquilizers and anti-depressants); thiazide drugs for diabetes insipidus; anti-Parkinson action of amantadine; anti-inflammatory action of steroids and phenylbutazone; anti-gout action of allopurinol; anti-arrhythmic action of phenytoin and lidocaine; uricosuric action of probenecid; acetazolamide for glaucoma and epilepsy; diazepam for status epilepticus; protective effects of beta-blockers (and the probable protective effects of platelet modulators, includ- ing aspirin) against myocardial infarction and coronary death; use of aspirin and sulfinpyrazone in preventing stroke; non surgical closure of patent ductus arterio- sus in premature babies by indomethacin" (201. 10. One caution needs to be made in this respect. As the research and development process is so lengthy, a number of companies may start working on a clinical problem at roughly the same time, but reach the market at somewhat different times. Regarding beta-blockers, for instance, the British ICI and the Swedish Astra started at roughly the same time, but ICI was first to market. Astra's subse- quent beta-blocker can not be simply defined as a me-too drug. 11. Drug discovery and preclinical research is governed directly by federal Good Laboratory Practices regulations; however, the investigational new drug regula- tions exert strong feedback pressures on how research is undertaken, especially toxicological research. 12. Patent protection is extremely important to drug research. Usually patents are filed early in the research process, preferably when there is a clear distinction between the active and inactive compounds. There are three types of patents: of a compound; of the use of a compound for a specific purpose; and of procedural methods of manufacture. 13. Part 312, Title 21, the Code of Federal Regulations specifies the procedures sur- rounding a "Notice of Claimed Investigational Exemption for a New Drug." Over the years the IND regulations have continuously been revised, resulting in a

COMPARING DEVELOPMENT OF TECHNOLOGIES 189 very complex system of requirements. Concerns were put forward that the inter- pretation of these regulations was unduly delaying the drug development process. An attempt was therefore made to rewrite these regulations in 1987, but accord- ing to the former legal counsel of the FDA this rewrite did not result in any sig- nificant changes (33,341. 14. This FDA study (36) analyzed a cohort of 172 new chemical entities that under- went human testing from 1976 through 1978. Not unexpectedly, new molecular entities developed outside the United States are less likely to be discontinued than U.S.-developed ones (14 percent versus 24 percent), as the foreign-developed entities usually have already been clinically tested outside the United States. 15. One of the major changes embodied in the 1962 amendments was to include the provision that a sponsor needs to provide "substantial evidence" of "effective- ness" as well as of"safety" (federal Food, Drug, and Cosmetics Act, as amended, Sec. 505 Idly. While effectiveness refers to the probability of benefits under aver- age conditions of use, efficacy refers to this under ideal conditions of use. The law uses the term effectiveness to make explicit that drugs are approved and labeled for use under the general conditions of medical practice, not the more ide- alized conditions often found in an investigational sewing (371. Extending this argument, it is for this very reason that we will use the term efficacy in the con- text of pre-marketing clinical investigations. 16. There are a number of design variations, such as crossover, stratified, matched, and factorial designs (41~. 17. Randomization reduces selection and blinding reduces observer bias. 18. For example, the size should be such as to avoid both Type I errors (the likelihood that an observed difference is due to chance) and lope II errors (the chance that a difference of interest is missed due to too few patients). 19. The agency already had some experience with such an approach. For instance, since the mid 1970s promising anti-cancer drugs (so-called group C cancer drugs) were distributed on a limited basis prior to approval through the National Cancer Institute (441. 20. The process by which a compound is initially synthesized, and milligrams to grams of materials are made at the laboratory bench, is not only quantitatively but also qualitatively different from the large-scale production process. For instance, laboratory chemists may use reagents in preparing small quantities of a com- pound that cannot be used in a large-scale production setting, which may need to produce a ton of a particular compound per year. 21. Increasingly, if a drug is intended for extensive use in a particular population such as the elderly, it is studied in that specific population. 22. Unless, of course, one is willing to delay the marketing of new drugs for extreme- ly long periods of time. This, however, would increase another kind of risk, i.e., the risk of not having a new or improved drug available on the market. 23. The following classification of INDs, and also of New Drug Applications, exists according to chemical type: (1) a new molecular entity not marketed before in the United States; (two) a new derivative from an active ingredient already mar- keted; (3) a new formulation of a drug already on the market; (4) a new combina- tion of two or more compounds; (5) a duplicate of an already marketed drug; and (6) a new indication of use for an existing drug. With regard to the potential ben- efit, the following distinction is made: (A) "important gain," i.e., may effectively

190 ANNETINE C. GELIJNS treat or diagnose a disease not adequately diagnosed or treated by any marketed drug; (B) "modest gain," i.e., offers modest but real advantage over existing prod- ucts; (C) "little or no gain," i.e., essentially offers therapeutic benefit similar to that of an already marketed drug. Orphan drugs, i.e., drugs developed for rare diseases (in principle with less than 200,000 American patients) are handled under a different system, which explicitly incorporates marketing and tax advan- tages for the sponsor. Such systems also exist in other regulatory schemes, e.g., the "fast track" system within the U.K. Committee on the Safety of Medicines. 24. The chemist in the team, among other things, requests that an inspection report be made to ensure that the sponsor adheres to good manufacturing practices (50~. 25. At this point a firm also needs to determine its price. The pricing mechanisms and the subsequent drug prices, as well as the health insurance or social security schemes, differ considerably by country. In the United States, there are few gov- ernment restrictions on setting drug prices. In Britain, however, the prices of drugs are controlled under the Price and Profit Regulation Scheme. Under this scheme, the government and the specific pharmaceutical industry agree upon a reasonable rate of return. This scheme thus institutes a target rate of return (in essence controlling profits), and only allows price increases to work through new products, thus providing an incentive for innovation (51~. 26. One development deserves mentioning as it directly influences drug development. In view of rising health care costs, third-party payers are sometimes refusing to reimburse even the routine costs of medical care associated with clinical trials of experimental drugs. 27. In the United Kingdom, for example, the well-established system of physician reporting to the Committee on the Safety of Medicines operates through the so- called Yellow Card System. 28. Cohort studies compare people exposed to a drug with those unexposed, and ana- lyze differences in adverse events between both groups. Case-control studies compare groups exhibiting a particular event with those not exhibiting this event, and then they examine differences in exposure to a particular drug. See the Report of the Joint Commission on Prescription Drug Use for an extensive dis- cussion of these methods (54~. 29. The DSRU system system has become a second national scheme to detect adverse Snug reactions greater than 1 in 10,000, and to evaluate the balance of risks and benefits of a drug. Using prescription-based cohorts as a starting point, this sys- tem actively solicits responses from physicians. The response rate is 70 percent, approximately 22,000 general practitioners report regularly, and the system catch- es nearly 50 million people. Monitored events are followed up by analysis of the medical records of the patients. 30. For example, the series of cases has been found to be subject to different kinds of physician and patient bias. Cohort studies, for example, may include limitations such as the exact specification of the cohorts, the quality of the data in terms of reproducibility and validity, the difficulty of analyzing the attributable agents, and the occurrence of detection bias. The U.S. Surgeon General's first report on smoking listed five supporting criteria to establish a cause-effect relationship: consistency of the association; temporal relationship between cause and effect; coherence with existing insights; specificity of the relationship; and strength of the association. See also Feinstein (60~.

COMPARING DEVELOPMENT OF TECHNOLOGIES 191 31. Bioengineering research will be defined as the application of engineering knowl- edge and concepts to the understanding of the human body and its interactions with machines, and to the development of new and improved medical devices. This definition is very similar to a definition provided in a recent National Research Council report (62), except that the scaling-up and production of new products derived from advances in biology (i.e., the engineering aspects of biotechnology) are excluded. Those aspects of engineering are discussed in the previous section. 32. In view of the heterogeneity of medical devices, the type of device determines if animal research will be undertaken before a device prototype is evaluated in humans. 33. Shaw found that half of the initial prototypes were produced by users. 34. Allen (73) established the importance of intra-organizational (e.g., between R&D and manufacturing divisions) and inter-organizational communication for R&D performance. With regard to the latter, a recent analysis of the development of devices demon- strated that half of the device firms considered used a formal financial analysis of the expected returns on investment or at least some form of market survey. Many firms, however, relied on informal decision making processes, usually based on a firm's experience in the market for the product (76~. 36. According to Kennedy (79), the term medical devices includes all of the items readily identified as devices as well as in vitro diagnostic devices used in clinical laboratories and some products previously regulated by the FDA Bureau of Drugs, such as IUDs, or by the Bureau of Biologics, such as arterial grafts. 37. A "significant risk" device is legally defined as an implant and presents a poten- tial for serious risk to the health and safety or welfare of a subject; is purported or represented to be for use in supporting or sustaining human life and presents a potential for serious risk to the health and safety or welfare of a subject; is for use of substantial importance in diagnosing, curing, mitigating, or treating disease and presents a potential for serious risk to the health and safety or welfare of a subject; or otherwise presents a potential for serious risk. 38. In some cases an IDE application is not necessary but clinical trials are con- ducted. 39. Determining technical performance involves replicability and reliability as impor- tant criteria. 40. The ROC analysis allows one to compare the technical performance of diagnostic tests over a range of different cutoff points or reference values that denote a posi- tive test result. This test displays the true positive ratios and the false positive ratios for these different cutoff points. See McNeil et al. (87~. 41. One needs to distinguish between critical and non-critical devices. Most rigorous GMP regulations apply only to critical devices. 42. On average the FDA takes a year to approve a PMA (81~. 43. As mentioned before, the economic environment in general and cost analyses of devices in particular are outside the scope of this paper. 44. The statutory provision indicates that this decision should be based on whether a device is considered "reasonable and necessary," which has been translated to mean "accepted by the medical community as a safe and efficacious treatment for a particular condition." Based on 13 technologies that completed the full

192 ANNETINE C. GELIJNS Medicare coverage process (including technology assessments by the Office for Health Technology Assessment) from the 1983-1988 period, it took 2.4 years from the time that HCFA received the initial inquiry to the final disposition date. 45. A condition of the approval for new Class m devices is that information received by manufacturers on device defects or adverse reactions should be reported to the FDA within 10 days. 46. In absolute terms, the United States invests heavily in biomedical research and development. Shepard and Durch (98), for example, indicate that the United States accounts for 45 percent of funds spent in the Organization for Economic Cooperation and Development countries, and the top five countries United States, Japan, The Federal Republic of Germany, France, The United Kingdom account for 84 percent of all biomedical R&D expenditures. If con- sidering per capita spending, however, Switzerland and Sweden head the list. 47. It is within this context that medical societies are increasingly issuing guidelines regarding the use of a particular new procedure; however, usually these guidelines emerge after a new procedure has already diffused more widely into clinical practice. The NIH consensus development conferences may issue similar recommendations regarding the appropriate use and effectiveness of a new procedures in clinical use. 48. The heated debate in the American Association of Neurological Surgeons and the New England Journal of Medicine illustrates the difficulties a number of promi- nent physicians had accepting the EC/IC bypass trial results (112,113), as well as the importance of ensuring "clear definition and relative homogeneity of the patients to be randomized." 49. Inherent in his proposal is a fluid protocol that allows incremental changes in techniques. 50. Alternatively, Buxton—in a three-year evaluation of heart transplants in the United Kingdom uses cross-sectional analyses to estimate changes in benefit and cost parameters over a longer time period than the study period directly allows (115~. 51. The few clinical trials using sham operations clearly demonstrated that a strong placebo effect can be associated with these surgical interventions, thus underlin- ing the importance of controls (119~. 52. The OECD in general defines industrial companies with 11 percent of their turnover in R&D already as "research intensive" (111~. 53. One furthermore should keep in mind that, whereas the success rates of NCEs are higher for 1970 cohorts than for earlier cohorts, at present 73 percent of NCEs initiating human testing are still discontinued before an NDA is submitted (63~. 54. The number of drugs approved for the U.S. market averaged 36 NCEs per year between 1950 and 1960. A decline of 54 percent occurred in the early 1960s, after which the numbers fluctuated, averaging 14 NCEs per year through the end of the 1970s. Since the end of the 1970s, approval rates recovered somewhat (26 in 1985, 20 in 1986), though this recovery did not specifically take place in U.S.- originated but in foreign-owned approvals (30,31,1251. 55. Notably, such benefits include the structural prevention of potentially unsafe and/or ineffective drugs; these basic premises on which the regulatory system is based are generally considered valuable. However, it is interesting that—in con- trast to the medical device amendments—there is no legal mandate to encourage development and innovation, but only to assure the marketing of"safe and effec- tive" drugs (124~.

COMPARING DEVELOPMENT OF TECHNOLOGIES 193 56. For example, getting a drug on the market one year earlier would reduce the aver- age break-even point economically (i.e., where R&D costs equal revenues) by three to four years (128~. 57. Grabowski (128) has determined that it would take 12 years of projected revenues at the present rate to achieve a real return on capital of 8 percent. A 10 percent real return would require 19 years of projected revenues at the present rate. 58. Furthermore, while the advantages for generic drugs can be reaped immediately, the advantages inherent in the law for innovative products can only be reaped fur- ther in the future (i.e., at the point where the patent term would have expired without the law). 59. Although a drug may continue to earn positive profits after the patent expiration date, under the pressure of generic competition the sales of a patent-expired prod- uct currently fall by 50 percent or more in the two or three years after patent expiry. 60. Furthermore, hospital formularies favor the lowest cost products, and the "Maximum Allowable Cost Program" reimburses Medicare patients only for the lowest cost product. In addition, international competition from Japan and Europe has increased. Recently the European Economic Community (EEC) intro- duced "the protection of the exclusive rights of the company that submits a file for regulatory approval. Files of new products, irrespective of the patent situa- tion, will remain inaccessible to others for up to 10 years from the time the first EEC approval has been granted" (129~. 61. "It is the purpose to encourage, to the extent consistent with the protection of the public health and safety and with ethical standards, the discovery and develop- ment of useful devices intended for human use and to that end to maintain opti- mum freedom for scientific investigators in their pursuit of that purpose" (Medical Device Ammendments 520, galls. 62. IDEs are devices under development which require FDA approval to initiate clin- ical evaluation in humans. 63. Consider, for example, the management of angina. The development of coronary artery bypass surgery and of beta-blockers were initiated at roughly the same time. The imbalance in assessment strategies, however, implies that the surgical option could undergo much more rapid diffusion than the pharmacological option, as beta-blockers were not as rapidly available to practicing physicians. 64. A subsequent study by Grabowski et al. (133) used a more sophisticated model, and found roughly similar results. As a measure of regulation they considered the average amount of NDA review time. Regarding research opportunities, they used changes in the productivity of pharmaceutical R&D in the United Kingdom during the 1960s as a control measure for changes in non-regulatory factors in the United States. 65. However, one should keep in mind that the four largest drug categories in the early 1960s anti-infectives, analgesics, cardiovasculars, and psychopharmacologic~still remained the largest therapeutic categories in the early 1980s. 66. While, as mentioned above, assessing efficacy and safety may be more complex with psychopharmacological products, this is certainly not the case with anti- infectives. Furthermore, the NDA review. times within the regulatory agency for these two categories were rather similar with regard to drugs in other therapeutic classes. In addition, the percentage of psychopharmacological drugs and anti- infectives first marketed abroad (under a different regulatory system) were also

194 ANNETINE C. GELIJNS roughly similar to the percentage first marketed abroad in other therapeutic class- es. According to these measures, it appears that neither psychopharmacological products nor anti-infectives were regulated more stringently. The additional decline in these two categories, above other categories, therefore should be relat- ed to a number of non-regulatory factors, such as a potential decrease in research opportunities, or a potential increase in perceived risk of developing drugs in a specific area (for instance, a relationship between the decrease in tranquilizers and the thalidomide tragedy). REFERENCES 1. Bernal ID. Science in History. Volume I-IV. Illustrated edition. Gretna, La.: Pelican Books, 1969. 2. Freeman C, Clark J. Soete L. Unemployment and Technical Innovation. London, 1982. 3. Rosenberg N. Inside the Black Box: Technology and Economics. New York: Cambridge University Press, 1982. 4. Landau R. Rosenberg N. The Positive Sum Strategy. Washington, D.C.: National Academy Press, 1986. 5. Casimir HBG. Haphazard Reality, Half a Century of Science. New York: Harper & Row Publishers, 1983. 6. Kline SJ, Rosenberg N. An overview of innovation. In Landau R. Rosenberg N feds). The Positive Sum Strategy. Washington, D.C.: National Academy Press, 1986:275-306. 7. Mowery D, Rosenberg N. The influence of market demand upon innovation: A critical review of some recent empirical studies. Research Policy 1979;8:102-153. 8. Hillman BJ. Government health policy and the diffusion of new medical devices. Health Services Research 1986;21:681-711. 9. Fox RC. The cultural shaping of biomedical science and technology. A preface. International Journal of Technology Assessment in Health Care 1986;2:189-194. 10. Nelson RR, Winter SG. In search of useful theory of innovation. Research Policy 1977;6:36-76. 11. Greer AL. Adoption of medical technology: The hospital's three decision-systems. International Journal of Technology Assessment in Health Care 1985;1:669-680. 12. Cody E. France orders sale of new pill. Washington Post, 1988. 13. Liebenau J. Innovation in pharmaceuticals: Industrial R&D in the early 20th centu- ry. Research Policy 1985;14:179-187. 14. Felig P. Biomedical research in the industrial setting: Contrasts and similarities to academia. Journal of the American Medical Association 1987;258:2407-2409. 15. Thier SO. Public policy and private sector innovation. Keynote Address at the Pharmaceutical Manufacturers Association Meeting on America's Medical Research, Washington, D.C., November 1987. 16. Mansfield E, Rapoport J. Wagner S. Hamburger M. Research and innovation in the modern corporation. New York: W.W. Norton, 1971. 17. Cuatrecasas P. Contemporary Drug Development - Dilemmas. The Center for the Study of Drug Development Publication Series. New York: University of Rochester, 1983.

COMPARING DEVELOPMENT OF TECHNOLOGY 195 18. Spilker B. Multinational Drug Companies: Issues in Drug Discovery and Development. New York: Raven Press, 1989. 19. Maxwell RA. The state of the art of the science of drug discovery an opinion. Drug Development Research 1984;4:380. 20. Wardell WM, Sheck LE. Is pharmaceutical innovation declining? In Lindgren B (ed). Pharmaceutical Economics. Stockholm: Swedish Institute for Health Economics, Liber Forlag, 1984:180-182. 21. Ahlquist RP. Study of adrenotropic receptors. American Journal of Physiology 1948;53:586-600. 22. Powell EE, Slater IH. Blocking of inhibitory adrenergic receptors by a dichloro analog of isoproterenol. Journal of Pharmacology and Experimental Therapeutics 1958:122:48~88. 23. Black JW, Stephenson IS. Pharmacology of a new adrenergic beta-receptor- blocking compound (nethalide). Lancet 1962;2:311-314. 24. Black JW, Crowther AF, Shanks RG, Smith LH, Dornhorst AC. A new adrenergic beta-receptor antagonist. Lancet 1964;1:1080-1081. 25. Nobel Committee, Stockholm, 1988. 26. Frishman WH. Clinical differences between beta-adrenergic agents: Implications for therapeutic substitution. American Heart Journal 1987; 113: 119~1198. van der Linden W. Pitfalls in randomized surgical trials. Surgery 1980;87:258-262. 28. Zbinden G. Flury-Roversi M. Significance of the LD-50 test for the toxicological evaluation of chemical substances. Archives of Toxicology 1981;47:77-99. 29. Zbinden G. Current trends in safety testing and toxicological research. Natur- wisenschaften 1982;69:255-259. 30. Kaitin N. Richard BW, Lasagna L. Trends in drug development: The 1985-86 new drug approvals. Journal of Clinical Pharmacology 1987;27:542-548. 31. Mattison N. Trimble AG, Lasagna L. New drug development in the United States, 1963 through 1984. Clinical Pharmacology and Therapeutics 1988;43:29~301. 32. Wardell WM, Sheck LE. Is pharmaceutical innovation declining? In Lindgren B (ed). Pharmaceutical Economics. Stockholm: Swedish Institute for Health Economics, Liber Forlag, 1984: 18~182. 33. Barton Hutt P. Regulation in the United States. International Journal of Technology Assessment in Health Care 1986;2:619-628. 34. FDA Off~ce of Planning and Evaluation. Assessment of the economic impact of revisions to regulations governing the submission of investigational new drug applications. Rockville, Md.: Food and Drug Administration, 1984. 35. Lasagna L, Bearn AG. Innovation and Acceleration in Clinical Drug Development. Roundtable 2: Early clinical trials Phase I-IIA: Optimal designs for maximum information. New York: Raven Press, 1987, pp. 95-103. 36. Tucker SA, Blozan C, Coppinger P. The outcome of research on new molecular entities commencing clinical research in the years 197~1978. Off~ce of Planning and Evaluation, study 77. Rockville, Md.: Food and Drug Administration, 1988. 37. Nightingale SL. Drug regulation and policy formulation. Milbank Memorial Fund Quarterly 1981;3:412 444. 38. Eschbach JW. Presentation at the Institute of Medicine. The workshop New Biology: Impact on Drug Development. Washington, D.C.: National Academy Press, 1989.

196 ANNETINE C. GELIJNS 39. Doyle AK. Biologic end-points as surrogates for clinical outcomes. In Lasagna L, Bearn AG (eds). Innovation and Acceleration in Clinical Drug Development. New York: Raven Press, 1987:117-123. 40. Bergner M. Quality of Life, Health Status, and Clinical Research. In Lohr KN (ed). Advances in Health Status Assessment: Conference Procedings. Medical Care 1989; 27:148-156. 41. General Accounting Office. FDA Drug Approval a lengthy process that delays the availability of important new drugs. Hearings before the Subcommittee on Science, Research, and Technology. May 1980. 42. Moses LE. Randomized Clinical Trials. In Assessing Medical Technologies. Institute of Medicine. Washington, D.C.: National Academy Press, 1985:73-80. 43. Chalmers TC. The clinical trial. Milbank Memorial Fund Quarterly 1981;59:324-339. 44. Young FE, Norris JA, Levitt JA et al. The FDA's new procedures for the use of investigational drugs in treatment. Journal of the American Medical Association 1988;259:2267-2270. Wardell WM, Tsianco MC, Anaveliar SN, Dans HT. Postmarketing surveillance of new drugs II. Case Studies. Journal of Clinical Pharmacology 1979;19:169-184. 46. Lasagna L. Techniques for ADR reporting. Detection and prevention of adverse drug reactions. Stockholm: Almquist G. Wichsell International, 1984:146-151. 47. Wiener H. Problems in the assessment of side effects of new drugs. New York: Pfizer Pharmaceuticals, internal document. 48. McKinlay SM. Experimentation in human populations. Milbank Memorial Fund Quarterly 1981;59:308-323. 49. FDA Consumer Special Report. From test tube to patient: New drug development in the U.S. Rockville, Md.: Food and Drug Administration, January 1988. 50. FDA Off~ce of Planning and Evaluation. Final regulatory impact analysis of changes to regulations governing the submission and review of new drug applica- tions. Part 314, Title 21. Rockville, Md.: Food and Drug Administration, August 1983. 51. Maynard A, Hartley K. The regulation of the pharmaceutical industry. In Lindgren B (ed). Pharmaceutical Economics. Stockholm: Swedish Institute for Health Economics,LiberForlag, 1984:123-137. 52. Medica-Pharmaceutical Forum. Clinical Trials. London: Royal Society of Medicine Services, 1987. 53. Jones JK. Regulatory use of adverse drug reactions. Stockholm: Skandia International Symposia, Almquist G. Wichsell International, 1984:203-214. 54. Joint Commission on Prescription Drug Use. Final Report. Washington, D.C.: Government Printing Off~ce, April 1980. 55. Borden EK, Lee JG. A methodological study of post-marketing drug evaluation using a pharmacy-based approach. Journal of Chronic Diseases 1982;35:803~16. 56. Inman WHW. Drug Surveillance Research Unit. Letter from the director. University of Southampton. PEM News 1983;N1. 57. Wennberg JE, Roos N. Sola L, Schori A, Jaffe R. Use of claims data systems to evaluate health care outcomes. Journal of the American Medical Association 1987:257:933-936. 58. Maffison N. Richard BW. Postapproval research requested by the FDA at the time

COMPARING DEVELOPMENT OF TECHNOLOGY of NCE approval, 1970-1984. Drug Information Journal 1987;21:309-329. 59. Levy RI, Sondik EJ. The management of biomedical research: An example for heart, lung, blood diseases. In Roberts EB, Levy Rl, Finestein SW, Moskowitz J. Sondik EJ. Biomedical Innovation. Cambridge: Massachusetts Institute of Technology Press, 1981. 60. Feinstein AR. Clinical Biostatistics. St. Louis: Mosby, 1977. 61. Eden M. The engineering-industrial accord: Inventing the technology of health care. In Reiser SJ, Anbar M (eds). The Machine at the Bedside. New York: Cambridge University Press, 1984. 62. Thomas LJ. Federal support of medical device innovation. In Ekelman K (ed). New Medical Devices: Innovation, Development, and Use. Washington, D.C., National Academy Press, 1988:59, 61. 63. National Research Council. Directions in Engineering Research. An Assessment of Opportunities and Needs. Washington, D.C.: National Academy Press, 1987: 8~85. 64. Office of Technology Assessment. Federal policies and the medical devices industry, October 1984. Washington, D.C., 1988. 65. Pollard MR, Persinger GS. Investment in health care innovation. Health Affairs 1987;Summer:93-100. 66. U.S. Department of Commerce, Bureau of Industrial Economics. U.S. Industrial Outlook. Washington, D.C., 1983. 67. Spilker B. Planning studies to evaluate medical devices. In Guide to Planning and Managing Multiple Clinical Studies. New York: Raven Press, 1987. 68. Roberts EB. Technological innovation and medical device innovation. In Ekelman K (ed). New Medical Devices: Innovation, Development, and Use. Washington, D.C.: National Academy Press, 1988. 69. van Hippel E. The dominant role of users in the scientific instrument innovation process. Research Policy 1976;3:212-239. 70. van Hippel E, Finkelstein S. Product designs which encourage or discourage related innovations by users: An analysis of innovation in automatic clinical chemistry analyzers. Working paper. Sloan School of Management. Cambridge: Massachusetts Institute of Technology, 1978. Whitehead EC. Development of technicians auto analyzer. In Ekelman K (ed). New Medical Devices: Innovation, Development and Use. Washington, D.C., National Academy Press, 1988. Shaw BF. The role of the interaction between the manufacturer and the user in the technological innovation process. Ph.D. dissertation. Sussex, U.K., 1987. Allen TJ. Communication Networks in R&D Laboratories. R&D Management 1970;1:1~21. 74. Roberts EB, Peters D. Commercial innovation from university faculty. Research Policy 1981;2:108-126. 75. Melmon KL. Strategic planning and the transfer of technology to clinical medicine. Presented to the Stanford Alumni Association, North Carolina, March 17, 1989. 76. Office of Planning and Evaluation. Pre-amendment medical device safety and efficacy testing. Rockville, Md.: Food and Drug Administration, October 1982. 77. Arthur D. Little. Medical device safety and efficacy testing PB 83-105-031; PB 83-105-049; PB 82-113-036. 197

198 ANNETINE C. GELIJNS 78. Food, Drug, and Cosmetics Act, Section 501 to 521. 79. Kennedy RS. Clinical investigations with medical devices: New rules. Journal of the American Medical Association 1981;245:2053-2054. 80. Food and Drug Administration. The Code of Federal Regulations. Section 860.3. Kessler DA, Pape SM, Sundwall DN. The federal regulation of medical devices. New England Journal of Medicine 1987:317:357-365. 82. General Accounting Office. Medical Devices FDA's 510(k) operations could be improved. August 1988. 83. IDE regulations. 45 Federal Register 3732, 1980. 84. Blozan CF, Gieser NC, Tucker SA. A profile of investigational device exemption applications. OPE study 76. Rockville, Md.: Food and Drug Administration, November 1987. 85. Fineberg HV, Bauman R. Susman M. Computerized cranial tomography: Effect on diagnostic and therapeutic plans. Journal of the American Medical Association 1977;238:224-230. 86. Horwitz RI, Feinstein AR, Crede WB, Clemens ID. Does technology work? Judging the validity of clinical evidence. In Reiser SJ, Anbar M feds). The Machine at the Bedside. New York: Cambridge University Press, 1984:193-208. 87. McNeil BJ, Keeler E, Adelstein SJ. Primer on certain elements of decision mak- ing. New England Journal of Medicine 1975;293:211-215. 88. Friedman PI. The early evaluation of MR imaging. American Journal of Radiology 1988;151:860-861. 89. Schwartz IS. Evaluating diagnostic technologies. In Assessing Medical Technologies. Institute of Medicine. Washington, D.C.: National Academy Press, 1985:80-89. 90. Barton Hutt P. Medical device regulation: Reasonable workable. Legal Times Washington, May 1980. 91. Russet LB. The Diffusion of New Hospital Technologies in the United States. International Journal of Health Services 1976;4:557-580. 92. Anderson OF, Steinberg E. To buy or not to buy technology: Acquisition under prospective payment. New England Journal of Medicine 1984;311:182-185. 93. Moxley JH. How trends will interact: The perspective of the hospital. In Ekelman K (ed). New Medical Devices: Innovation, Development, and Use. Washington, D.C.: National Academy Press, 1988:127-137. 94. Altman SM. Impact of the changing medical payment system in technological innovation and utilization. In Ekelman K (ed). New Medical Devices: Innovation, Development and Use. Washington, D.C.: National Academy Press, 1988: 93-103. 95. National Advisory Council on Health Care Technology Assessment. The Medicare Coverage Process. Office of the Assistant Secretary for Health, Department of Health and Human Services, September 1988. 96. Carr SW, Jones JH. Drug epidemiology data bases applicable to medical device monitoring. Medical Device Data Base Conference, Washington D.C., American Medical Association, October 1983. 97. Green SB, Byar DP. Using observational data from registries to compare treat- ments: The fallacy of omnimetrics. Statistics in Medicine 1984;3:361-370. 98. Shephard D, Durch IS. International comparison of resource allocation in health sciences: An analysis of expenditures on biomedical research in 19 industrialized

COMPARING DEVELOPMENT OF TECHNOLOGIES 199 countries. Supported by the Fogarty International Center, National Institutes of Health. Harvard School of Public Health, 1985. 99. Swazey, JP, Fox RC. The clinical moratorium. In Fox RC. Essays in Medical Sociology. New York: Wiley-Interscience, 1974:325-363. 100. Sladek JR, Shoulson I. Neural transplantation: A call for patience rather than patients. Science 1988;240:1386-1388. 101. Fox RC, Swazey J. The Courage to Fail: A Social View of Organ Transplants and Dialysis. Chicago: University of Chicago Press, 1974. 102. Comroe JH, Dripps RD. Scientific basis for the support of biomedical science. In Roberts ED, Levy RI, Finestein SW, Moskowitz J. Sondik EJ: Biomedical Innovation. Cambridge: MIT Press, 1981. 103. Moore FD. Therapeutic innovation: Ethical boundaries in the initial clinical trials of new drugs and surgical procedures. Daedalus 1969;Spring:502-522. 104. Bergkamp L. American IRBs and Dutch Research Ethics Committees: How they compare. IRB: A Review of Human Subjects Research. Hastings Center 1988;10:1-6. 105. Barnes BA. Discarded operations: Surgical innovation by trial and error. In Bunker JP, Barnes BA, Mosteller F (eds). Costs, Risks and Benefits of Surgery. Oxford University Press, 1977:109-123. 106. Bunker JP, Barnes BA, Mosteller F (eds). Costs, Risks and Benefits of Surgery. Oxford University Press, 1977. 107. Bunker JP, Hinkley D, McDermott WV. Surgical innovation and its evaluation. Science 1978;200:937-941. 108. Eddy DM, Billings J. The quality of medical evidence: Implications for quality of care. Health Affairs 1988;Spring:20-32. 109. Wennberg JE. Improving the medical decision making process. Health Affairs 1988;Spring:99-106. 110. Gilbert JP, McPeek B. Mosteller F. Statistics and ethics in surgery and anesthesia. Science 1977;198:684-689. 111. Office of Economic Cooperation and Development (OECD). Science and technology indicators. No.2: R&D, Invention and Competitiveness. Paris, March 1986. 112. EC/IC Bypass Study Group. Failure of extracranial-intracranial arterial bypass to reduce the risk of ischemic stroke. New England Journal of Medicine 1985;313:1191-1200. 113. EC/IC Bypass Study Group. Are the Results of the Extracranial-Intracranial Bypass Trial Generalizable? New England Journal of Medicine 1987;13:820-824. 114. Chalmers TC. Randomization and coronary arte~y surgery. Annals of Thoracic Surgery 1972;14:323-327. 115. Buxton M. Problems in the economic appra~sal of new health technology: The evaluation of heart transplants in the United Kingdom. In Drummond MF (ed). Economic Appraisal of Health Technology in the European Community. Oxford Medical Publications, 1987. 116. Feinstein AR. An additional basic science for clinical medicine: II. The limita- tions of randomized trials. Annals of Internal Medicine 1983;49:544 550. 117. Bonchek LI. The role of the randomized clinical trial in the evaluation of new operations. Surgical Clinics of North America 1982;62:761-769. 118. Bonchek LI. Are randomized trials appropriate for evaluating new operations? New England Journal of Medicine 1979;301:~4 45.

200 ANNETINE C. GELlJNS 119. Dimond KG, Kittle CF, Crocked JE. Comparison of internal mammary artery lig- ation and sham operation for angina pectoris. American Journal of Cardiology 1960;5:483-486. 120. Hlatky MA, Lee KL, Harrel FE, Califf RM, Pryor DB, Marck DB, Rosatti RA. Tying clinical research to patient care by use of an observational database. Statistics in Medicine 1984;3:375-384. 121. Roos LL et al. Computers to identify complications after surgery. American Journal of Public Health 1985;75:1288-1294. 122. Kaye HL. The biological revolution and its cultural context. International Journal of Technology Assessment in Health Care 1986;2:275-284. 123. Wiggins SN. The cost of developing a new drug. Pharmaceutical Manufacturers Association, Washington, D.C., June 1987. 124. Laubach GB. Federal regulation and pharmaceutical innovation. In Levin A (ed). Regulating Health Care: The Structure for Control. Proceedings of Political Science 1980;3:64-80. 125. May MS, Wardell WM, Lasagna L. New drug development during and after a period of regulatory change: Clinical research activity of major United States pharmaceutical firms, 1958 to 1979. Clinical Pharmacology and Therapeutics 1983;33:691-700. 126. Pharmaceutical Manufacturers Association. Update on biotechnology products in development. Washington, D.C., July 1988. 127. Temin P. Taking Your Medicine: Drug Regulation in the United States. Cambridge, Mass.: Harvard University Press, 1980. 128. Grabowski HG. Health Care Cost Containment and Pharmaceutical Innovation. Boston: Center for the Study of Drug Development, 1986. Reprint RS S707. 129. Wiedhaup K. View of contraceptive development. Presentation for the National Research Council, Washington, D.C., March 1988. 130. Read JL, Campbell PM. Health care innovation: A progress report. Health Affairs. 1988;Summer:174-185. 131. Peltzman S. An evaluation of consumer protection regulation: The 1962 drug amendment. Journal of Political Economy 1973;81:1049-91. 132. Baily MN. Research and development costs and returns: The U.S. pharmaceutical industry. Journal of Political Economy 1972;80:78-85. 133. Grabowski HG, Vernon J. Thomas L. Estimating the effects of regulation on inno- vation: An international comparative analysis of the pharmaceutical industry. Journal of Law and Economics 1978;21: 1-32. 134. Wiggins SN. The effect of U.S. pharmaceutical regulations on new introductions. In Lindgren B (ed). Pharmaceutical Economics. Stockholm: Swedish Institute for Health Economics,LiberForlag, 1984:191-205, 195-201. 135. Wardell WM. Introduction of new drugs in the United States and Great Britain: An international comparison. Clinical Pharmacology and Therapeutics 1973; 14:773-790. 136. Wardell WM. Therapeutic implications of the drug lag. Clinical Pharmacology and Therapeutics 1974;15:73-96. 137. Wardell WM. The drug lag revisited: Comparison by therapeutic area of drugs marketed in the United States and Great Britain from 1972 through 1976. Clinical Pharmacology and Therapeutics 1978;24:499-527. 138. Bakke OM, Wardell WM, Lasagna L. Drug discontinuations in the United

COMPARING DEVELOPMENT OF TECHNOLOGY 201 Kingdom and the U.S., 1964 to 1983: Issues of safety. Clinical Pharmacology and Therapeutics 1984;35:559-567. 139. Hass AK, Portable DP, Grossman RE. New Drug Introductions, Discontinuations and Safety Issues in the United States and the United Kingdom: 1960-1982. OPE (FDA) study 68. 140. Berlin H. Jonsson B. International dissemination of new drugs: A comparative study of six countries. Managerial and Decision Economics 1986;7:235-242.

Next: Appendix B: Workshop Agenda »
Modern Methods of Clinical Investigation Get This Book
×
Buy Hardback | $49.95
MyNAP members save 10% online.
Login or Register to save!
Download Free PDF

The very rapid pace of advances in biomedical research promises us a wide range of new drugs, medical devices, and clinical procedures. The extent to which these discoveries will benefit the public, however, depends in large part on the methods we choose for developing and testing them.

Modern Methods of Clinical Investigation focuses on strategies for clinical evaluation and their role in uncovering the actual benefits and risks of medical innovation.

Essays explore differences in our current systems for evaluating drugs, medical devices, and clinical procedures; health insurance databases as a tool for assessing treatment outcomes; the role of the medical profession, the Food and Drug Administration, and industry in stimulating the use of evaluative methods; and more.

This book will be of special interest to policymakers, regulators, executives in the medical industry, clinical researchers, and physicians.

  1. ×

    Welcome to OpenBook!

    You're looking at OpenBook, NAP.edu's online reading room since 1999. Based on feedback from you, our users, we've made some improvements that make it easier than ever to read thousands of publications on our website.

    Do you want to take a quick tour of the OpenBook's features?

    No Thanks Take a Tour »
  2. ×

    Show this book's table of contents, where you can jump to any chapter by name.

    « Back Next »
  3. ×

    ...or use these buttons to go back to the previous chapter or skip to the next one.

    « Back Next »
  4. ×

    Jump up to the previous page or down to the next one. Also, you can type in a page number and press Enter to go directly to that page in the book.

    « Back Next »
  5. ×

    To search the entire text of this book, type in your search term here and press Enter.

    « Back Next »
  6. ×

    Share a link to this book page on your preferred social network or via email.

    « Back Next »
  7. ×

    View our suggested citation for this chapter.

    « Back Next »
  8. ×

    Ready to take your reading offline? Click here to buy this book in print or download it as a free PDF, if available.

    « Back Next »
Stay Connected!