National Academies Press: OpenBook

Technology and Health Care in an Era of Limits (1992)

Chapter: PART I: SETTING THE STAGE

« Previous: FRONT MATTER
Suggested Citation:"PART I: SETTING THE STAGE." Institute of Medicine. 1992. Technology and Health Care in an Era of Limits. Washington, DC: The National Academies Press. doi: 10.17226/2024.
×
Page 1
Suggested Citation:"PART I: SETTING THE STAGE." Institute of Medicine. 1992. Technology and Health Care in an Era of Limits. Washington, DC: The National Academies Press. doi: 10.17226/2024.
×
Page 2
Suggested Citation:"PART I: SETTING THE STAGE." Institute of Medicine. 1992. Technology and Health Care in an Era of Limits. Washington, DC: The National Academies Press. doi: 10.17226/2024.
×
Page 3
Suggested Citation:"PART I: SETTING THE STAGE." Institute of Medicine. 1992. Technology and Health Care in an Era of Limits. Washington, DC: The National Academies Press. doi: 10.17226/2024.
×
Page 4
Suggested Citation:"PART I: SETTING THE STAGE." Institute of Medicine. 1992. Technology and Health Care in an Era of Limits. Washington, DC: The National Academies Press. doi: 10.17226/2024.
×
Page 5
Suggested Citation:"PART I: SETTING THE STAGE." Institute of Medicine. 1992. Technology and Health Care in an Era of Limits. Washington, DC: The National Academies Press. doi: 10.17226/2024.
×
Page 6
Suggested Citation:"PART I: SETTING THE STAGE." Institute of Medicine. 1992. Technology and Health Care in an Era of Limits. Washington, DC: The National Academies Press. doi: 10.17226/2024.
×
Page 7
Suggested Citation:"PART I: SETTING THE STAGE." Institute of Medicine. 1992. Technology and Health Care in an Era of Limits. Washington, DC: The National Academies Press. doi: 10.17226/2024.
×
Page 8
Suggested Citation:"PART I: SETTING THE STAGE." Institute of Medicine. 1992. Technology and Health Care in an Era of Limits. Washington, DC: The National Academies Press. doi: 10.17226/2024.
×
Page 9
Suggested Citation:"PART I: SETTING THE STAGE." Institute of Medicine. 1992. Technology and Health Care in an Era of Limits. Washington, DC: The National Academies Press. doi: 10.17226/2024.
×
Page 10
Suggested Citation:"PART I: SETTING THE STAGE." Institute of Medicine. 1992. Technology and Health Care in an Era of Limits. Washington, DC: The National Academies Press. doi: 10.17226/2024.
×
Page 11
Suggested Citation:"PART I: SETTING THE STAGE." Institute of Medicine. 1992. Technology and Health Care in an Era of Limits. Washington, DC: The National Academies Press. doi: 10.17226/2024.
×
Page 12
Suggested Citation:"PART I: SETTING THE STAGE." Institute of Medicine. 1992. Technology and Health Care in an Era of Limits. Washington, DC: The National Academies Press. doi: 10.17226/2024.
×
Page 13
Suggested Citation:"PART I: SETTING THE STAGE." Institute of Medicine. 1992. Technology and Health Care in an Era of Limits. Washington, DC: The National Academies Press. doi: 10.17226/2024.
×
Page 14
Suggested Citation:"PART I: SETTING THE STAGE." Institute of Medicine. 1992. Technology and Health Care in an Era of Limits. Washington, DC: The National Academies Press. doi: 10.17226/2024.
×
Page 15
Suggested Citation:"PART I: SETTING THE STAGE." Institute of Medicine. 1992. Technology and Health Care in an Era of Limits. Washington, DC: The National Academies Press. doi: 10.17226/2024.
×
Page 16
Suggested Citation:"PART I: SETTING THE STAGE." Institute of Medicine. 1992. Technology and Health Care in an Era of Limits. Washington, DC: The National Academies Press. doi: 10.17226/2024.
×
Page 17
Suggested Citation:"PART I: SETTING THE STAGE." Institute of Medicine. 1992. Technology and Health Care in an Era of Limits. Washington, DC: The National Academies Press. doi: 10.17226/2024.
×
Page 18
Suggested Citation:"PART I: SETTING THE STAGE." Institute of Medicine. 1992. Technology and Health Care in an Era of Limits. Washington, DC: The National Academies Press. doi: 10.17226/2024.
×
Page 19
Suggested Citation:"PART I: SETTING THE STAGE." Institute of Medicine. 1992. Technology and Health Care in an Era of Limits. Washington, DC: The National Academies Press. doi: 10.17226/2024.
×
Page 20
Suggested Citation:"PART I: SETTING THE STAGE." Institute of Medicine. 1992. Technology and Health Care in an Era of Limits. Washington, DC: The National Academies Press. doi: 10.17226/2024.
×
Page 21
Suggested Citation:"PART I: SETTING THE STAGE." Institute of Medicine. 1992. Technology and Health Care in an Era of Limits. Washington, DC: The National Academies Press. doi: 10.17226/2024.
×
Page 22
Suggested Citation:"PART I: SETTING THE STAGE." Institute of Medicine. 1992. Technology and Health Care in an Era of Limits. Washington, DC: The National Academies Press. doi: 10.17226/2024.
×
Page 23
Suggested Citation:"PART I: SETTING THE STAGE." Institute of Medicine. 1992. Technology and Health Care in an Era of Limits. Washington, DC: The National Academies Press. doi: 10.17226/2024.
×
Page 24
Suggested Citation:"PART I: SETTING THE STAGE." Institute of Medicine. 1992. Technology and Health Care in an Era of Limits. Washington, DC: The National Academies Press. doi: 10.17226/2024.
×
Page 25
Suggested Citation:"PART I: SETTING THE STAGE." Institute of Medicine. 1992. Technology and Health Care in an Era of Limits. Washington, DC: The National Academies Press. doi: 10.17226/2024.
×
Page 26
Suggested Citation:"PART I: SETTING THE STAGE." Institute of Medicine. 1992. Technology and Health Care in an Era of Limits. Washington, DC: The National Academies Press. doi: 10.17226/2024.
×
Page 27
Suggested Citation:"PART I: SETTING THE STAGE." Institute of Medicine. 1992. Technology and Health Care in an Era of Limits. Washington, DC: The National Academies Press. doi: 10.17226/2024.
×
Page 28
Suggested Citation:"PART I: SETTING THE STAGE." Institute of Medicine. 1992. Technology and Health Care in an Era of Limits. Washington, DC: The National Academies Press. doi: 10.17226/2024.
×
Page 29
Suggested Citation:"PART I: SETTING THE STAGE." Institute of Medicine. 1992. Technology and Health Care in an Era of Limits. Washington, DC: The National Academies Press. doi: 10.17226/2024.
×
Page 30
Suggested Citation:"PART I: SETTING THE STAGE." Institute of Medicine. 1992. Technology and Health Care in an Era of Limits. Washington, DC: The National Academies Press. doi: 10.17226/2024.
×
Page 31
Suggested Citation:"PART I: SETTING THE STAGE." Institute of Medicine. 1992. Technology and Health Care in an Era of Limits. Washington, DC: The National Academies Press. doi: 10.17226/2024.
×
Page 32
Suggested Citation:"PART I: SETTING THE STAGE." Institute of Medicine. 1992. Technology and Health Care in an Era of Limits. Washington, DC: The National Academies Press. doi: 10.17226/2024.
×
Page 33
Suggested Citation:"PART I: SETTING THE STAGE." Institute of Medicine. 1992. Technology and Health Care in an Era of Limits. Washington, DC: The National Academies Press. doi: 10.17226/2024.
×
Page 34

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.

Part I Setting the Stage

1 Introduction Gerald D. Laubach, John E. Wennberg, and Annetine C. Gelijns This volume summarizes the third of a series of Institute of Medicine (IOM) workshops whose intent is to critically examine medical innova- tion that is, the process by which medical research findings are translated into actual benefits in clinical practice. The raison d'etre for this third workshop (see Appendix A) is deftly captured in the opening sentence of the second chapter: "The U.S. health care sector is the target of a massive social struggle over its reform." Our purpose here is not to become addi- tional participants in that debate but rather to visualize the restructured U.S. health care system that might ultimately emerge from it. In that visualiza- tion, we are particularly interested in the likely implications of the new system for practitioners, patients, and the generators of new medical tech nology. Such an undertaking inevitably involves a good deal of speculation. On closer examination, however, the degree of speculation may not be nearly as great as one might anticipate. Although currently more than two dozen discrete proposals for restructuring health care in the United States have been put forward, many of them seem transitional in nature. Workshop participants seemed to agree that two alternative scenarios of health care financing and delivery stand out as likely candidates. The first mirrors an extension of current trends: a mature form of today's pluralistic public- private system that relies heavily on managed care to influence ("micro- manage") physician and patient decision making directly. The second sce 3

INTRODUCTION s cost and cost-effectiveness within the vast range of medical technologies (see also the papers by Griner and Hillman). This volume defines the term medical technology broadly to encompass medical devices, instrumentation, pharmaceuticals, biologicals, diagnostic and therapeutic procedures, and in- tegrated systems of care delivery. This definition is at variance with the stereotype of medical technology as being inherently and invariably cost inflating. Some medical interventions, such as organ transplants, are indeed quite expensive (albeit potentially cost-effective); yet the cost of one of the most cost-effective technologies, the poliomyelitis vaccine, is trivial-. Dis- crimination among existing alternative technologies to ensure their appro- priate use, as well as encouragement of the development of cost-effective new interventions, seems a logical goal of all health care systems. Beyond these relatively straightforward economic considerations, our earlier workshops have revealed other vulnerabilities of the innovation pro- cess that could become acute in cost-sensitive health care systems. One is the tendency to discourage experimentation, both with regard to the use of emerging technologies and the exploration of new uses for accepted tech- nologies. The consequences may be serious, because innovation takes place not only in academic and industrial laboratories but also at the bedside. Small departures from established practice have yielded several important advances, a process that has produced substantial progress in surgery. But the phenomenon applies as well to technologies that are the end result of elaborate, formal R&D processes, such as those employed in the develop- ment of pharmaceuticals. A case in point is the evolution of adrenergic beta-blocking drugs, one of the more significant medical innovations of our time. These compounds were introduced for the treatment of two cardio- vascular indications, arrhythmias and angina pectoris. Today, they are used in the treatment of more than 20 diverse conditions, largely as a result of clinical discoveries made after beta-blockers were in general use (Gelijns and Thier, 1990~. The interweaving of innovation with clinical practice applies not only to the generation or creation of medical interventions but also to their pro- duction and delivery. For example, the ways in which products and human processes are packaged into discrete "units" of care-such as a coronary bypass procedure or an intensive care unit (ICU) may vary substantially between institutions and even among individuals within one institution. In the case of an emerging technology, this is not surprising, given that its optimum manner of use, the clinical results it produces, and the resource costs associated with it may not have stabilized. Even in the case of estab- lished technologies, however, their safety, effectiveness, and appropriate- ness typically depend on who uses them, how skillfully, and in what clinical situation. In contrast to other economic sectors, the "production" processes of medicine have relatively little standardization. Because the evaluative

6 GERALD D. PREACH, JOHN E. WENNBERG, AND ANNETINE C. GELl]NS clinical sciences are underdeveloped, the optimum use of many technolo- gies is an unresolved clinical problem (Wenuberg, 1990~. This phenomenon underlies an important deficiency of medical technol ogy assessment as it is often practiced. The important contribution to the process of further development and definitive evaluation that can be made by studying the outcomes of a new intervention in actual practice is not often reflected in the technology policies of otherwise sophisticated health care systems. Not infrequently, decision-making processes seem to be based on the idea that a technology can be evaluated, once and for all, on the basis of research findings collected before its use in everyday practice. But such assessments conducted in carefully controlled clinical settings-often tell us much less than we need to know about the actual performance of a technology. These decision-making processes reflect an idealized, unrealis- tic understanding of both medical practice and medical innovation. In sum, previous work of the committee indicates that the net social benefit of some of the anticipated technology policies of a restructured American health care system may be positive. The predictable emphasis on cost-effective technologies, as exemplified by the trend toward minimally invasive procedures, may offer something for everyone: potential economic savings for payers, better-quality care for patients, professional growth for practitioners, and strong encouragement to particular kinds of innovation. The impact of other likely policies may be ambiguous or negative. For example, the systematic discouragement of experimental technologies may actually produce little or no savings in health care expenditures and could well reduce future benefits both economic and humane that might have flowed from successful innovation. Indeed, the overarching risk to society of today's technology policies may well be their inherent conservatism- their tendency to "freeze in place the status quo," as Neumann and Wein- stein (1991) observed in an earlier volume in this series. Because much of today's medical technology is both costly and limited in its effectiveness- the "halfway technologies" of Lewis Thomas-delaying the emergence of more definitive interventions may be unsound policy. It is not obvious that health care policies need to dampen innovation to achieve their intended purpose that is, encouraging cost-effective care. Instead, it may prove possible to devise policies that constructively merge a comprehensive, cost-conscious U.S. health care system with a more rational program in the evaluative clinical sciences-and at the same time preserve a strong commitment to medical research and development. What, then, are some of these policies? Our deliberations to date suggest a few. 1. Adopt the philosophy and techniques of modern industrial practice to foster a process of continuous quality improvement in health care. Such policies systematically and constructively alter the status quo. Moreover,

INTRODUCTION 7 U.S. private-sector health care firms might be expected to be strongly at- tracted to the concept, given the widespread interest in continuous improve- ment strategies in American industry today. Provocative evidence suggests that these techniques can work in medical settings (Berwick et al., 1990~. Britain's National Health Service is also exploring reforms along these lines (Berwick et al., 1992a, b). 2. Capitalize on the massive data bases required for accounting purposes in the U.S. health care system by making them an instrument for continuous evaluation of medical practice. Such a strategy would not discourage the new and experimental by denying reimbursement but instead would require partic- ipation in a systematic program of evaluation. It would utilize the same principles to reevaluate older, potentially obsolete technologies. 3. Adopt a universal policy of transparency for technology policies and reimbursement decision rules. In the pluralistic U.S. system, such an open policy would help optimize the social benefits of medical technology. 4. Adopt policies that more systematically engage patients in decision making regarding their treatment. This volume strongly supports the ethical case for including the patient as a full partner in clinical practice (see the papers by Mulley and Silberman). Less well recognized is the economic case some patients prefer more conservative, less costly options for the management of their medical problems than might be chosen by their physi- cians alone. 5. Finally, incorporate policies that explicitly encourage technological innovation, including adequate evaluation. Numerous countries have suc- cessfully implemented public policies to counterbalance the innovation-sup- pressing effects of health care regulation; for example, the British National Health Service and the French and Japanese governments retain explicit incentives for innovation in their pharmaceutical pricing formulas (Burstall, 1991; Neimeth, 19914. Thus, policies to compensate for unintended sup- pression of innovation in health care can be crafted and indications are that they do, indeed, work. In conclusion, both the United States and the majority of other industri- alized nations are actively debating how to reform their health care systems. These debates are generally couched in terms of consequences of changes in health care for costs, access, and quality. Often overlooked are the likely implications of such changes for medical innovation and the subsequent effect of innovation on costs, access, and quality. It is our hope that the discussions in this volume encourage consideration of these issues. REFERENCES Berwick, D. M., Godfrey, A. B., and Roessner, J. 1990. Curing Health Care. San Francisco: Jossey-Bass Publishers.

~ GERALD D. LA UBACH, JOHN E. WENNBERG, AND ANNETINE C. GELl]NS Berwick, D. M., Enthoven, A., and Bunker, J. P. 1992a. Quality management in the NHS: The doctor's role I. British Medical Journal 304:235-239. Berwick, D. M., Enthoven, A., and Bunker, J. P. 1992b. Quality management in the NHS: The doctor's role II. British Medical Journal 304:304-308. Burstall, M. L. 1991. European policies influencing pharmaceutical innovation. In: A. C. Gelijns and E. A. Halm, eds. Medical Innovation at the Crossroads. Vol. 2, The Chang- ing Economics of Medical Technology. Washington, D.C.: National Academy Press, pp. 123-140. Gelijns, A. C., and Thier, S. O. 1990. Medical technology development: An introduction to the innovation-evaluation nexus. In: A. C. Gelijns, ed. Medical Innovation at the Crossroads. Vol. 1, Modern Methods of Clinical Investigation. Washington, D.C.: Na- tional Academy Press, pp. 1-15. Neimeth, R. 1991. Japan's pharmaceutical industry postwar evolution. In: A. C. Gelijns and E. A. Halm, eds. Medical Innovation at the Crossroads. Vol. 2, The Changing Econom- ics of Medical Technology. Washington, D.C.: National Academy Press, pp. 155-167. Neumann, P. J., and Weinstein, M. C. 1991. The diffusion of new technology: Costs and benefits to health care. In: A. C. Gelijns and E. A. Halm, eds. Medical Innovation at the Crossroads. Vol. 2, The Changing Economics of Medical Technology. Washington, D.C.: National Academy Press, pp. 21-34. Wennberg, J. E. 1990. What is outcomes research? In: A. C. Gelijns, ed. Medical Innova- tion at the Crossroads. Vol. 1, Modern Methods of Clinical Investigation. Washington, D.C.: National Academy Press, pp. 33-46.

Innovation and the Policies of Limits in A Changing Health Care Economy John E. Wennberg The U.S. health care sector is the target of a massive social struggle over its reform. The strategies of the past have failed to establish access and contain costs. Indeed, the trend today is toward less access for the poor and many in the middle class, with more care for those who remain "enti- tled." Thirty-three million Americans are without health insurance. For those with insurance, the rates of utilization of physician services and ex- pensive diagnostic techniques and the number of invasive procedures being performed continue to spiral. If these trends are not altered, the United States will be spending more than 15 percent of its gross national product (GNP) on health by the year 2000. The message of the 1992 presidential debates is that this situation must change. The systems for financing care must be fixed; to do this costs must be contained. Improvements in access must be accompanied by policies of limits. The way the politics and policies of limits are fashioned will depend on assumptions about the relationship between the utilization of care and the benefits of care-that is, on the shape of the "benefit-utilization curve." One popular interpretation is that the shape of this curve is such that the nation needs to ration effective care. Patient demand and medical progress now make the health care system so expensive that it can no longer be available on equal terms to everyone; moreover, the nation simply cannot afford to pay for everything that works and that patients want. This predic- ament arises because of the successes of biomedical research, the resulting 9

10 JOHN E. WENNBERG efficacy of clinical science, and the efficiency of practicing physicians in translating medical knowledge into beneficial medical interventions. As utilization increases, benefits also increase, but at a declining rate of return as the level of invested resources increase. Somewhere along the curve, society finds itself in a zone in which the benefits can no longer be afford- ed: the costs of further transfers of GNP toward health care and away from national priorities such as education and housing are simply too great. As a consequence, society must learn to make explicit judgments about the value of specific services as they apply to an individual patient's case. Some experts advocate rationing by age; others recommend using detailed algo- nthms for specific patient subgroups defined on the basis of "cost-effective- ness."i The intent, however, is the same. Policies are needed to set limits on specific services, to develop explicit methods to ration effective care that brings less than socially acceptable marginal returns. The effect of such policies is to deny access to care that works and that patients want on the basis that it is not cost-effective. In the opinion of many, this denial of access inevitably produces a two-tiered system of care, one for the affluent and another for those whose access to care must be underwritten by policies of entitlement. An alternative interpretation emphasizes that the inadequacies of clini- cal science and flaws in the role of the physician as the decision-mal~ing "agent" of the patient make it impossible to determine the shape of the benefit-utilization curve in medicine. Although investments in basic bio- medical science have greatly increased the power of technology to intervene in the natural history of disease, efforts to evaluate the outcomes of these interventions the effects of medical technology and theories of efficacy in the specific situations of everyday practice have substantially failed. The risks and benefits of most medical care are poorly understood. Moreover, the agency role of the physician is flawed by professional dominance. This role, which depends on the capacity of physicians to make vicarious judg- ments about what patients want, has created a market in which the prefer- ences of patients are entangled with those of the physician. In short, in medicine, too little is known about what works and what patients want. This interpretation emphasizes the major role played by supplier-in- duced demand, in which the weaknesses in the scientific and ethical status of clinical medicine ensure that available resources are utilized without evidence that more is necessarily better or that patients necessarily want 1The most sophisticated articulation of this argument is made by Aaron and Schwartz (1984); for a recent update, see Aaron and Schwartz (1990). Among those advancing arguments for age-based rationing are Callahan (1987) and Lamm (1987). For a description of Oregon's approach to rationing, see Fox and Leichter (1991) and Brown (1991); both reports appear in an issue of Health Affairs that focuses on Oregon's priority setting.

INNOVATION AND THE POLICIES OF LIMITS 11 more. Medicine's untested and often conflicting theories of efficacy justi- fy and the dominance of professional preferences ensures the full de- ployment of resources, no matter how many or in what quantities. Indeed, medical theory is often implicit and closely associated with per-capita quan- tities of supply such as hospital beds and physician subspecialists. The crisis in costs is the inevitable consequence of the policies of growth that have prevailed in the U.S. health care sector: the open-ended financing of entitlement services based on funding of utilization; the accelerated produc- tion of manpower based on perceptions during the 1960s and 1970s of a "shortage" in medical manpower; the specialization of the physician work force and its division into technology-dnven subspecialties whose work- loads have uncertain impacts on the health status and satisfaction of pa- tients; easy access to capital markets for the construction of facilities and the purchase of technology; and a willingness of payers to reimburse for services involving underevaluated technology. The end result of these pol- icies has been quantities of supply that are well in excess of the amount required to produce and deliver services that are known to work or that patients are known to want. Under this interpretation, policies of health care limits should concentrate on global restrictions on growth and the promotion of strategies to learn what works and what patients want. As one familiar with the patterns of use of medical care and the strengths and weaknesses of the scientific status of clinical medicine, I find a good deal of evidence in favor of the supplier-induced theory of demand.2 This paper seeks to explain this point of view in more detail. It examines alter- native interpretations of the shape of the benefit-utilization curve to raise the "which rate is right?" question; that is, what is the rate of service use (and the amount of resources required) when supplier-induced demand is reduced-when patients are informed of the state of medical progress (what is known and not known about the results of care) and when patients are free to choose according to their own preferences? It then looks at the struggle between two competing models for reforming the doctor-patient relationship. One is based on the assumption that the agency role of the physician can be essentially replaced by the guardianship of the third-party payer through m~cromanaged care and that the delegated decision model inherent in the agency role can be preserved by prescriptive rules of prac- tir`~ `1~.v~.1on~.~1 hv competing health care organizations or the state. The r-- -a - ~ ~~~r ~ -= 2The theoretical basis of this argument was developed in Wenuberg and colleagues (1982). Archie Cochrane's Effectiveness and Efficiency (1972) provides a thorough introduction to the problems of physicians in understanding the outcomes of care; Eddy and Billings (1988) provide a more recent example. Much of the epidemiological evidence concerning the prob- lem of supplier-induced demand is summarized in "Small Area Analysis and the Medical Care Outcome Problem" (Wennberg, 1990a).

12 JOHN E. WENNBERG other is based on replacing the delegated decision model with shared deci- sion making, a new partnership between the patient and the physician and the profession and the public. The target in this approach is to reform the ethical status of the doctor-patient relationship so that what is known and not known are explicitly shared and so that patient preferences become dominant in the choice of treatment from among reasonable and available plans of care. The paper also examines why neither micromanaged care nor the shared decision model are sufficient to achieve the goal of rationalizing utilization and containing costs. The implicit nature of much of medical theory keeps most of clinical practice outside of their influences. The paper thus argues that to contain costs it is necessary to limit capacity directly, and it sets out several principles to guide debate about strategies for devel- oping limits. The paper concludes with a discussion of the implications for innovation of policies of limits in medicine. THE SHAPE OF MEDICINE'S BENEFIT-UTILIZATION CURVE Different assumptions about the relationship between the utilization of care and the benefits of care are reflected in different assumptions about the shape of the benefit-utilization curve in medicine. The assumption that it is necessary to ration specific services that work and that patients want but that society cannot afford is based on two ideas. One is the notion of "expected value" or benefit now obtained from the resources invested in health care. The doctor-patient relationship efficiently distributes care in such a way that patients benefit from the care they now use more than they would if the care were not received or if it were replaced by a less costly item or service. The other idea is that the "marginal returns" can be ratio- nalized; that is, the benefits are sufficiently well understood that they can be ranked in terms of expected benefit per unit of service (or dollars) uti- lized. The diminishing marginal returns that occur in an economy in which utilization is the result of patient demand and biomedical progress are ill~- ~ _ ~ ~ ~ _ _ 1 · ~ A- a- Lra~u in rlgure A-. lo achieve sucn a curve, physicians must be remark- ably successful in the sorting of medical problems and in diagnosing and ranking them according to the expected outcomes of treatments. Indeed, they must be perfect in the execution of their agency role to interpret clini- cal science and understand the values their patients assign to alternative treatments, including the value of no treatment at all. They must thus ~ . possess something that clinical science does not now provide: knowledge .. The problem the physician about the outcomes that matter to patients r ~ r~~~~~~ faces, however, is not simply to know the outcomes but to weight them according to the individual patient's attitudes toward them. Under delegat- ed decision making, which has been the dominant medical decision model

INNOVATION AND THE POLICIES OF LIMITS a ._ In ._ \ \ \ - - Units of Care Supplied 13 FIGURE 2-1 Example of benefit-utilization curve where medical progress and pa- tient demand drive utilization with diminishing marginal returns. The curve is adapted from Aaron and Schwartz The Painful Prescription: Rationing Hospital Care, p. 11. since the time of Hippocrates,3 physicians make the choices. They intuit what patients would choose if patients had full information about choices and could do the weighting themselves. In this model, prescriptions for treatment are based on clinical experience, that is, on the "case series" of patients that the individual physician has treated and whose values and preferences have somehow become known. At the level of the population, the situation is even more complex. For the curve to take its hypothesized shape of decreasing marginal social re- turn, physicians must, collectively, make ordered choices; that is, the pa- tient who receives the first unit of service is the one who gains the most benefit from it, the second patient the one who gains the second most bene- fit, and so on, with the patient who receives the last unit gaining the least. For this to occur, however, physicians must know and respond to decisions their colleagues have made. There is no feedback loop currently operating in health care markets that would make this feasible; indeed, it is difficult to conceive what such a mechanism would look like. 3See The Silent World of Doctor and Patient by Jay Katz (1984) for an excellent history of the doctor-patient relationship.

14 JOHN E. WENNBERG Outcomes and preference research provide empirical evidence that dem- onstrates the weaknesses of the assumption that the benefit-utilization curves in medicine correspond to Figure 2-1. Over the past few years, my col- leagues and I have been involved in an in-depth investigation of the fine structure of a clinical decision problem: the choices that face men with a common form of prostate disease called benign prostatic hyperplasia, or BPH. The inquiry was motivated by small area variation studies showing that the chances that a man would undergo a prostate operation by the time he reached age 85 varied from about 15 percent in some communities to more than 50 percent in others. We asked a group of Maine urologists, some of whom lived in areas with low rates, others of whom lived in high- rate communities, if they could explain these variations. The urologists differed in their assumptions about the benefits to be derived from prostate operations and about the shape of the curve relating benefits to utilization. Some held a pessimistic attitude toward prostate disease. These physicians believed that BPH usually progressed to a life- threatening obstruction of the bladder or kidney and that it was best to operate early in the course of the disease to prevent future bad outcomes. We called this the preventive theory of surgery. Their clinical decisions were dominated by a hypothesized benefit-utilization curve in which the chief benefit of surgery was improvement in life expectancy. Their reason- ing was that if surgery were postponed until evidence of life-threatening obstruction appeared, the patient would be older, sicker, and more likely to die when surgery finally became unavoidable. By operating early, one avoided the higher death rates that occurred when the operation was post- poned. Because most men who exhibit early disease will progress to the point where surgery is inevitable, and because the death rate from surgery increases with age, BPH patients will live longer if they have the operation earlier. Because most men eventually develop BPH symptoms, the popula- tion captured under the benefit curve of preventive theory (Figure 2-2) encompasses the majority of older men. Other urologists were more optimistic about untreated BPH and argued for the quality of life theory of surgery. This theory posits that the benefit of surgery for men without obstruction of the bladder or kidneys is its ability to reduce symptoms and improve the quality of life. In the opinion of these urologists, BPH does not usually progress to the point where it threatens life; accordingly, surgery does not play a preventive role in avoid- ing early death. They estimated that a patient's "utility" for surgery-that is, the "expected value" the patient would gain if the surgeon prescribed surgery was greatest for those with severe symptoms, whereas those with mild symptoms benefited little. The benefit-utilization curve suggested by these physicians (Figure 2-3) thus has a different parameter of benefit as well as a more rapid decline in value.

INNOVATION AND THE POLICIES OF LIMITS \ o ._ a) Q o a) Q ._ ~4 o o ._ a) X LL \ \ \ \ \ \ 11 1 2040 60 80 Percentage of Men Having a Prostatectomy 15 FIGURE 2-2 Benefit-utilization curve under the theory that prostate surgery in- creases life expectancy. The unresolved competition between the preventive and the quality of life theories (and their corresponding benefit-utilization curves) reflects in- determinacy rooted in poor clinical science. The outcomes research we undertook showed that the preventive theory was incorrect.4 Early surgery appeared to lead to a slight decrease in life expectancy because for most men BPH does not progress to life-threatening obstruction. Those without evidence of such obstruction were better off with watchful waiting if the expected value of treatment was an increase in life expectancy. The curve in Figure 2-2 thus is incorrect. If prostate surgery has a place for men with symptoms, it lies in accordance with the quality of life theory. But the uncertainty about the shape of the benefit-utilization curve is more profound than the failure to define and measure the outcomes that matter to patients. Most urologists who believed in the quality of life hypothesis also practiced within the delegated decision tradition. They understood that they bore a responsibility as the patient's agent to interpret 4The findings of this research project have been widely reported. For examples, see Wenn- berg et al. (1988); Fowler et al. (1988); and Barry et al. (1988).

16 JOHN E. WENNBERG o ._ Q o Q ._ o - ._ a t . E o - - 20 40 60 80 Percent of Men Having a Prostatectomy (Ranked According to Decreasing Severity of Symptoms) FIGURE 2-3 Benefit-utilization curve under the theory that prostate surgery im- proves quality of life. for him what he needed and to convince him, for reasons of his own best interest, to accept their prescription. Yet our preference research showed that what patients want cannot be predicted from objective information available to the physician that is, from data gained during the physical examination, from laboratory tests measuring such factors as urine flow, or even from answers to questions about the severity of symptoms or impairment of qual- ity of life. Patients who by all such objective measures are similar may still differ in their preferences for treatment. Indeed, as it turned out, when they were informed about the alternatives and offered a choice, nearly 80 percent of men with severe symptoms choose watchful waiting, at least initially. Preferences for outcomes and level of aversion to risk cannot be intuited reliably by physicians based on objective knowledge; to know what patients want, physicians must ask them. When patient preferences are neglected or misunderstood, the benefit- utilization curve is erratic, without evidence of rational sorting, and the net expected value can actually be negative. For example, if 16 severely symp- tomatic men were ranked according to impairment of urine flow (a common

~ 4~ ~ fO[~ OF [~ +~ .E E o ~ a_ {l o -----/-e-\---/---\ o~ I {1 ~ \/ o o m O A - I I T ! T T ! i -\ \ \ / / 1 @ 3 0 5 6 77 7 8 9 @ 11 12 ~14 15 16 Waled Hen P^r 3u~e~ FIGURE 2-4 BeneOt-utHizabon curve for 16 meD limb severe prostate symptoms hobo were prescribed surgery under Me delegated decision mama model agnosUc 1001), our studies predict that only 4 would want surgery and tab 1be 4 choosing surgery would not be compelled with urine now imp~rmenL If surgery were preschhed for an 16 men on the basis of the delegated decision model tab is, without informing patients bout their options and asking them abut they wanted most patients would receive cage they did not want. Whereas the patients ranked second, Pugh, tenth, and tbi~eenth may want surgery, the Thorny do not. For these patients, the expected value of surgery is actuaHy negative, compared with the benefit they would have obtained Tom the w~cb~1 mating option they wanted (Figure 2-4~. WHICH RATE IS RIGHT? What is the rue of service (and resource) use when papists me in- ~rmed Tout the Age of medical progress Tout Rabat is knotted and not MOOD Tout the relationship between Segment options and the outcomes that mater to patients and He Bee to choose among options according to their own preferences? What He their Deludes towed the beneAts and asks of the expected outcomes? Preference rese~cb gives a tentative answer. Ruben patients with BPH He Ally indeed Tout tube treatment options and asked to p~icip~e in

8 JOHN E. WENNBERG clinical decisions, they choose less invasive treatments more often than they do when decision making follows the delegated model; their choice of treat- ment is strongly influenced by the degree to which they axe bothered by their symptoms and their fear of impotence. Moreover, the population- based rates of surgery decline. For this condition, the trend toward conser- vative (nonsurgical) treatment choice is evident even in prepaid group prac- tices in which the rates of surgery were already relatively low (and where patients faced no cost barriers at the point of delivery). The results of our preference research thus indicate the likelihood of significant negative re- turns on current patterns of resource deployment under the delegated deci- sion model. For the vast majority of illnesses and conditions, the pattern of varia- tion in treatment choice is similar to or even greater than the variation seen for prostate disease. Most conditions do not have a single best treatment. For most conditions for which there is one or more "appropnate" surgical treatments there is also one or more nonsurgical options that are feasible within current scientific understanding. Surgical options for nine condi- tions-angina, arthritis of the hip and knee, silent gallstones, menopausal conditions affecting the uterus, peripheral vascular disease, back pain due to disc disease, atherosclerosis of the arteries of the neck, and BPH ac- count for well over half of the major surgery performed in the United States (Table 2-1~. For these, the shape of the benefit-utilization curve is un TABLE 2-1 Common Conditions for Which the Shape of the Benefit Utilization Curve Is Unknown Condition Major Treatment Controversies Noncancerous condition of the uterus Angina pectoris Gallstones Peripheral vascular disease Cataracts Arthritis of hip and knee Prostatism (BPH benign prostatic hyperplasia) Herniated disc Atherosclerosis of carotid artery with threat of stroke Surgery (by type) vs. hormone treatment vs. drugs vs. watchful waiting Bypass surgery vs. angioplasty vs. drugs Surgery vs. stone crushing vs. medical management vs. watchful waiting Bypass surgery vs. angioplasty vs. medical management Lens extraction (by type) vs. watchful waiting Surgery (by type) vs. medical management Surgery (by type) vs. balloon dilation vs. drugs vs. microwave diathermy vs. watchful waiting Surgery (by type) vs. various medical management strategies Carotid endarterectomy vs. aspirin

INNOVATION AND THE POLICIES OF LIMITS o ._ Q U) ._ m 19 do;) The cube cannot be specified because of uncertainty overthe nature of the benefits and the preferences of patients. Units of Care Supplied FIGURE 2-5 Status of the current understanding of the benefit-utilization relation- ship for most medical treatment theories. known (see Figure 2-5~. Through outcomes research, however, and the distinguishing of patient preferences from those of the physician, it will be possible to create islands of rationality in a sea of uncertainty and supplier- induced demand. REFORM OF THE DOCTOR-PATIENT RELATIONSEIIP The flaws in the role of the physician as a decision-making agent for the patient are now widely apparent, and reform of the doctor-patient rela- tionship is under way. The question now is, which model will replace it? In the United States, ambitious programs have been implemented to check the autonomy of the physician by imposing the use of clinical policy man- agers to micromanage the choices doctors make for their patients. Unlike the classic staff-model health maintenance organization (HMO) in which cost containment is achieved through global restrictions on the quantity of supply, the strategy of micromanaged care is to force efficiency by setting parameters of practice to define the available options and to guide a myriad of everyday clinical decisions. Acting as agents for third-party payers, the micromanagers develop rules of practice and use them to patrol the deci

20 JOHN E. WENNBERG signs of physicians. Virtually all major insurance companies offer micro- managed care programs that invade the traditional decision-making authori- ty of physicians, subjecting their prescribing authority to the discipline of rules developed and administered by third parties. The strategy is not limit- ed to private payers: the state of Oregon, while setting an overall budget, administers constraints not through global budgeting or overall limits on resources but through an ambitious program to micromanage benefits (ser- vices) based on estimates of relative cost-effectiveness. This author's criticism of the micromanaged care model is based in part on its failure to understand the limitations of clinical science. The strategy assumes that the shape of the benefit-utilization curve is known and that this knowledge can be codified by the committees of experts convened to write the rules of practice. Inevitably, because of defects in the knowledge base of medicine which were discussed earlier, virtually none of the rules that emerge from this process can be based on an understanding of the structure of the decision problem over which they claim authority. Because the process is decentralized there seem to be well over a hundred parties engaged in rule setting- the results are a plethora of varying prescriptive or advisory codifications of the rules of practice, some that directly conflict with one another. In an unfortunate number of examples, the rules are secret or proprietary, the property of the managed care company. This critique is also based on the ethical weakness of micromanaged care-that is, its tendency to preserve the delegated decision-making model by substituting the guardianship of the third party for the benevolent author- itarianism of the physician. There is little pretense that this model of re- form concerns itself with the preferences of individual patients. The impo- sition of third-party rule often occurs through telephone conversations over an 800 number involving the physician and the agent of the third party. Patients are not involved. The irony is obvious: if physicians do not know what works and what patients want, how can the third party claim such knowledge? The substitution, however, is not simply that of one imperfect agent for another. The micromanagers are agents of the payer, not the patient. The intrusion opens the doctor-patient relationship to the free play of various interests in setting and enforcing clinical rules, a circumstance that offers numerous opportunities for irrationally and arbitrarily rationing care and otherwise perverting the struggle to base clinical choice on the preferences of patients. The opposing model for reform centers on what would be required to facilitate the sharing of decision making by patients and physicians. It depends on a program of outcomes and preference research and on philo- sophical inquiry to build the scientific and ethical base for helping patients make decisions that reflect their preferences. It depends on a new relation- ship between doctors and patients, based on open communication and the

INNOVATION AND THE POLICIES OF LIMITS 21 development of new styles of discourse and trust. Shared decision making competes for authority with micromanaged care. Each model seeks to in- fluence clinical choice in specific nonemergency situations such as those listed in Table 2-1, situations in which options exist and time is available to reach a decision. Most commonly, these choices involve the elective use of surgery, but the participants in the conversation are different from those in the managed care model. There, the conversation takes place between the payer and the physician; in shared decision making, discourse is between the physician and the patient. Because of its explicitly ethical basis, the shared decision-making mod- el is preferable when fateful choices must be made between elective treat- ment options that entail differing risks and benefits. In these situations, rational choice depends on finding out what the patient wants. As noted earlier, preferences cannot be diagnosed by physical examination, laborato- ry tests, or questionnaires about symptoms or quality of life. A patient who meets the managed care criteria of eligibility based on these objective fea- tures of "appropriate care" may be prescribed a treatment he or she does not want. The benefit-utilization curves that result under micromanaged care look like those shown in Figures 2-4 and 2-5. It is doubtful that effective communication with a patient can occur over an 800 number, no matter how interested the micromanager may be in patient preferences. The shared decision-making model is therefore appropriate for choosing treatments for patients with such common conditions as angina, gallstones, uterine bleed- ing, benign hypertrophy of the prostate, cataracts, back pain caused by herniated discs, and arthritis of the hip. LINKAGE BETWEEN THE SUPPLY OF RESOURCES AND MEDICINE'S IMPLICIT THEORIES OF EFFICACY The critique of third-party micromanaged care presented here also rests on the likely futility of such an approach in effectively limiting undisci- plined growth of the health care system. This is a weakness of the shared decision-making model as well. Neither approach is likely to produce sys- temwide containment of costs, which requires a macroeconomic policy; reforms targeted at the microeconomy of the doctor-patient relationship are insufficient tools for disciplining the macroeconomy. The reason is that most of medicine's resources are not used to execute discrete treatment options specified by well-developed medical discourse. Condition-treat- ment options such as those listed in Table 2-1 are the exception; most medical resources are allocated implicitly in varying patterns or cascades of acts undertaken to solve medical problems. The theoretical reasons for one pattern of allocation compared with another are often implicit and inacces- sible to precise rules of practice. Indeed, medical theories are often so

22 JOHN E. WENNBERG closely associated with the supply of medical resources that physicians do not even recognize them as explicit theories. The effect of the supply of beds on the clinical thresholds for hospital- izing patients is a good example. The supply of hospital resources varies remarkably across geographic areas, and the amount allocated is unrelated to illness rates or to explicit theories about the numbers of beds required to treat most diseases. When more beds become available, they are allocated across a broad range of medical conditions for which the clinical policies governing admission and readmission rates are correlated with the supply of beds. In areas with fewer beds, patients with these conditions are more often treated outside of the hospital. For only a few medical conditions are the clinical thresholds such that illness rates determine the probability of hospitalization. Let us consider the example of Boston, Massachusetts, and New Haven, Connecticut. Residents of Boston have about 4.5 hospital beds per 1,000 persons whereas New Havenites have only 2.9. Explicit rules of practice, codified in medical texts, dictate hospitalization for all patients with heart attacks or strokes, for those who need major surgery to treat their cancers, and for those who suffer from major trauma such as a hip fracture. Diag- nostic criteria for these conditions are explicit, and the technologies for distinguishing the presence or absence of the condition are well advanced. But hospitalization for patients with these conditions requires less than 20 percent of the available medical beds, even in low-bed-rate areas. Most of the beds are used for conditions that exhibit highly variable patterns of admission. Hospitalizations for low back pain are the single most important reason for the difference in bed use between Boston and New Haven. Gas- troenteritis is the second, followed by chronic bronchitis and pneumonia. The implicit rules of practice in Boston, which have been adapted to the greater supply of beds, result in a larger proportion of the population being admitted, with more frequent readmissions and shorter intervals between admissions. The mechanism underneath the association between beds and admission and readmission thresholds centers on the decision making of physicians who must decide what to do with sick people. Imagine someone who is faced with the task of watching a conveyor belt that presents, in seemingly random order, a series of balls-some black, some white, some in varying shades of grey. The rules require that all black balls be picked up and put on a shelf and that white balls be left on the conveyor belt. The sorting task is to examine each of the grey balls and decide whether they should be put on the shelf but at the same time to save room for all the black balls that must be put there. The conveyor belt simulates the flow of patients through the emergency room or the doctor's office. The black balls are the conditions that all

INNOVATION AND THE POLICIES OF LIMITS 23 physicians agree require hospitalization; regardless of how sick the patient may appear, he or she must be admitted. The white balls are those condi- tions that are treated outside of the hospital. The task is to decide which grey balls should be "admitted" placed on the shelf. Each must be evalu- ated as to its degree of grayness (relative illness). There are always more grey balls (more sick people) than there is room on the shelf. The sorter must find a way to deal with the problems presented by all of the grey balls. Making judgments about degrees of illness and finding safe alternatives for patients are difficult tasks and involve a great deal of uncertainty. If beds are available, it is much easier to rely on others the ward physicians and nurses to sort out the problems. It is far easier to admit a patient to the hospital (put a grey ball on the shelf) than to search out alternatives. Small area analysis reveals something that the sorters do not know. There are fewer beds per capita in New Haven; consequently, the shelf there is much smaller than the shelf in Boston about half its size. In New Haven, the sorter must spend more time evaluating shades of grey and finding safe alternatives by making calls to the patient's private physician, eliciting the help of home health agencies, or arranging transportation to nursing homes, hospices, or departments of social services. Small area studies reveal another curious feature: about the same amount of space on the shelf is held in reserve in the low-rate areas as in the high- rate ones. Despite the greatly reduced size of their shelves, there is no evidence that physicians in areas with low per-capita bed rates believe that they are rationing effective services. They do not even use all of the beds available to them. Clinicians in New Haven, like their counterparts in Boston, tend to hold about 15 percent of beds in reserve. They seem un- aware of scarcity and are satisfied that their theories of how hospital re- sources should be deployed are appropriate. The chiefs of service at Yale's teaching hospital in New Haven, when asked whether they were aware that treatments were being withheld because of the area's per-capita bed supply, could identify no examples of explicit rationing. Indeed, as is so often the case when the facts of variation are presented, the chiefs were not even aware that hospital resources were relatively scarce in their community. In the upside-down economy of medical care, supply comes first and theory follows, in virtual equilibrium as practice style adjusts to ensure the utilization of available resources. Thus, treatment theories governing the use of hospital beds are sufficiently flexible to allow the use of available beds, no matter what the per-capita level of supply; theories that establish the legitimacy of the use of particular procedures justify professional work- loads, virtually without regard for the number of specialists; and undereval- uated medical treatment theory is sufficiently rich to permit the employ- ment of internists and family practitioners virtually without regard to how many there may be per capita. As clinical problem solvers, it is in the

24 JOHN E. WENNBERG nature of physicians to deploy available resources, including themselves, close to the point of scarcity. They do this in pursuit of treatment theories that seem reasonable and that might just prove to be effective. This behav- ior is not the result of simple self-interest; it arises from physicians' percep- tions of their role as healers, their faith in plausible theories of efficacy, and their willingness to work to find solutions to the endless stream of problems their patients present. It is quite possible that higher per-capita rates of investment in health care produce no net benefit over what is achieved in areas with lower per- capita rates. Consider the evidence regarding bed supply. Much of the additional pool of resources in Boston is invested in more frequent readmis- sions of the chronically ill and in the care of terminal patients. In spite of the 70 percent greater per-capita expenditure for Bostonians compared with New Havenites, the mortality rates are the same, as predicted by the simi- larities in demographic characteristics. In addition, the more than twofold variation in expenditures for hospitalization among the 185 hospital service areas of New England is not correlated with mortality rates. Indeed, why should greater spending bring better results? In formulat- ing an expectation about whether more should be better, it is well to recall the contingencies that determine the capacity of the health care system. Capacity is not fashioned according to explicit theories about what works in medicine. The optimal number of beds is unknown, and the number that is actually built or supplied has no theoretical or empirical basis. (One looks in vain to medical texts to learn how many beds are needed for treating a population's burden of illness for such conditions as back pain, pneumonia, and gastroenteritis.) The number of beds is the result of the way the hospi- tal industry has grown. Per-capita rates are arbitrary, the product of the opportunities and desires of institutions and communities not of the needs or preferences of patients, shaped by the possibilities articulated by medical science. This is easily seen in case studies that reveal the history of the planning and construction of hospital beds. The populations of Waterville and Au- gusta in Maine are about the\same in size, but Waterville has nearly twice as many beds per capita. The reason is that Waterville hospitals were constructed according to the dictates of religious and professional orthodox- ies, a set of dynamics that resulted in three hospitals: a Catholic and a Protestant hospital, each used by allopathic physicians, and a third hospital reserved for osteopaths. In Augusta, only one hospital was built, an ecu- menical institution shared by all religious and professional persuasions. The medical care landscape in the United States is contoured by the jagged profiles of resource allocation exemplified by Boston, New Haven, Water- ville, and Augusta. In each example, the intensity of construction is deter- mined by dynamics that are indifferent to theories of efficacy or even to

INNOVATION AND THE POLICIES OF LIMITS 25 simple rules about the necessary numbers of beds in relation to the size of the population. The number of physicians who are trained is governed by equally arbi- trary policies, many of which were formulated in the 1960s, a period of great concern about medical scarcity. The number of physicians trained for each specialty is the product of administrative and political choices rather than a response to the resources required to produce services dictated by an answer to the "which rate is right?" question. In the case of procedure- oriented specialties, supply is well in excess of the number of practitioners needed to produce treatments that physicians agree are efficacious. For example, when neurosurgeons enter medical markets, they almost invari- ably find that the available supply has already taken care of the demand for surgical management of brain tumors and head trauma, which are the proce- dures that all physicians agree are needed. Neurosurgeons thus must invest most of their efforts in treating conditions for which there are valid nonsur- gical options. The most common are two condition-treatment options listed in Table 2-1: back operations and carotid artery surgery. Although it is reasonable to conjecture that more of such surgeries might produce some benefit, the studies noted earlier suggest that the amount of neurosurgery now being supplied under the delegated decision model could well exceed the amount patients want when they choose according to well-informed preferences. SEEKING LIMITS It is quite possible that the current crisis in health care in this country may stem from the excesses of an economic sector dominated by supplier- induced demand and professional uncertainty about the value of medical care and not from patient demand based on medical progress. The excesses arise because of errors in the assumption of neoclassic economic policy that capacity would be limited and the quality of care maintained by medical efficacy and patient demand, mediated through the physician who serves as the rational agent for patient and society. The effects of these errors are . . . now increasingly apparent: · Quality is poor. Patient values are not paramount in the decision to use care; information on options (and on the state of medical progress) is not freely communicated; and services (whether wanted or not) are pro- duced with varying efficiency in regard to outcomes. · Costs are out of control. The supplies of resources are created (in increasing amounts) without regard to explicit theories of efficacy and with- out knowledge of the shape of the benefit-utilization curves for medical interventions and of the amount of resources needed to produce the services patients want.

26 JOHN E. WENNBERG · Access is diminished. The decentralized structures that have fi- nanced care through insurance are increasingly unable to provide products that are affordable to business or to individuals without employment-based insurance. It seems very likely that the l990s will bring policy decisions that place explicit limits on the medical care system in the United States, although no model of governance has emerged as an odds-on favorite. It may well be impossible to reach a national consensus on what to do, in which case the initiative will fall to the states. If so, several models may evolve, but their shapes should be governed by certain principles and guidelines that find their empirical justification in the epidemiology of medical care and in ethics. The first principle concerns the general welfare: it is safe for patients and in the public interest to place global restrictions on growth in the capacity to provide medical care. Studies of geographic variation in services in this country provide solid evidence that the capacity of the hospital in- dustry and of the physician work force is now well in excess of that re- quired to provide services that are efficacious and that patients actually want. Most medical resources are allocated to treatments for which the theoretical basis for allocation is implicitly associated with the supply of resources and for which there is no empirical evidence that more is better. The nation can and should deal directly with the forces of inflationary growth in the health care sector with the policies that determine the num- bers and distribution of manpower, the size of the hospital industry, and the quantities of technology. The excess in capacity means that the amount spent on health care (as a percentage of GNP) can be directly limited and a health care system achieved that is in equilibrium with other sectors of the national economy without fear that valuable services must necessarily be rationed. The second principle concerns the welfare of those who do not now have access to care because they lack insurance: full entitlement of all Americans to health care can be instituted without increases in the propor- tion of GNP invested in health and without a loss of welfare to those now insured. The fear that policies that extend health care entitlement to all citizens will exacerbate the cost crisis is unwarranted; the dynamics that determine the capacity and costs of health care markets are to a large extent independent of illness rates and the demands of patients. To see why this is so, let us return to the analogy of the person sorting the black, white, and grey balls. Physicians are unaware of the relative size of the resource shelf-that is, of the per-capita quantities of "supply" invested in their markets; put another way, they are unaware of the relative size of the popu- lation they are serving. For example, two-thirds of the population of Ver- mont could move to Boston before the relative size of the Boston hospital

INNOVATION AND THE POLICIES OF LIMITS 27 resource shelf approaches that of New Haven. The testimony of physicians in New Haven and the statistical evidence that resources are held in re- serve equally in all medical markets, regardless of the relative per-capita rate of resource investment tells us that if the increase in population size occurred gradually, no one in the medical care industry serving Bostonians would notice the difference. The major change in practice style would be a change in the threshold for hospitalization-a more careful sorting of grey balls. The biggest change in the rates of use of hospitals would be for back pain, gastroenteritis, and chronic bronchitis. Fewer than 15 percent of Americans are completely uninsured. An understanding of the epidemiology of medical care leads to the prediction that their entitlement would permit them to be absorbed into the health care system without loss of benefit to those now in the system and without any special increase in aggregate expenditures. The capacity to treat the unin- sured is already there; what is needed is to make it possible for them to compete for the attention of the health care system on an equal basis with the insured. In a steady-state situation, the increases in costs for treating the uninsured will be offset by the savings realized by reducing utilization among those now insured. The third principle concerns the interests of patients for whom expen- sive medical care is effective in a system characterized by excess capacity: the resources required to meet unmet needs (e.g., prenatal care, bone mar- row transplants, long-term care) should be obtained by reallocation of ex- cess capacity and not by rationing effective care. From the point of view of patients with costly diseases, the reallocation of excess capacity is a more humane way to meet unmet needs than is the deliberate withholding of expensive, effective care on the grounds that the benefits are too costly. If the people of Oregon decide that total resources for health care should be limited, then resources to meet unmet needs should be reallocated from areas of excess capacity. Oregon has its own Bostons and New Havens. Rather than withholding specific treatments such as bone marrow trans- plants, which are known to increase the expectation for life (and that pa- tients are known to want), this principle recommends the reallocation of resources now invested in excess supplies of hospital beds. Large quanti- ties of resources are thus available for reallocation. If the practice patterns of Boston were more like those of New Haven, 700 hospital beds would be unused, and in 1982 dollars, $300 million would be available for realloca- tion to other medical needs (Culp et al., 1987~. INNOVATION AND THE POLICIES OF LIMITS Policies of limits that emphasize the rationing of care through prescrip- tive rules of practice, that is, through the micromanagement of the doctor

28 JOHN E. WENNBERG patient relationship, have very different implications for innovation than policies that set limits that have been developed in accordance with the principles set out in the previous section. The differences are key. At the level of the microeconomy the doctor-patient relationship- the latter poli- cies emphasize the underdevelopment of clinical science, the entanglement of preferences, and the implicit nature of much of medical theory. At the level of the macroeconomy, they emphasize the opportunities for meeting unmet needs that global limits and strategies for reallocating excess capaci- ty open up. These opportunities also include the development of the neces- sary professional infrastructure to deal with the weaknesses in the scientific and ethical status of the doctor-patient relationship. A successful policy of global limits has the immediate consequence of buying time to learn what works in medicine and to sort out the many conflicting, explicit theories governing resource deployment in the treatment of discrete conditions such as those listed in Table 2-1. But the several European and the Canadian models for managing the macroeconomy show clearly that setting global limits does not of itself lead to improvement in clinical science or to the development of models for clinical decision making that emphasize patient preferences. For innovation along these lines to prosper, policies to achieve global limits must be linked to a science policy that builds the infrastructure for evaluating medical theory and promotes new models of the doctor-pa- tient relationship. The introduction suggests some of the characteristics of a science poli- cy that would promote "rational" innovation under both policies of global limits and of managed care. Such policies would encourage the develop- ment of new ideas and technologies and their systematic evaluation in a context that fosters the progressive growth of a more fully rationalized microeconomy, namely, a doctor-patient relationship in which decisions are based on information about outcomes and on the preferences of patients. This chapter also draws attention to the sources of medical ideas and the current processes of evaluation to highlight the importance of problem solv- ing in everyday practice as a source of medical theory. In addition, it emphasizes the lack of standardization in innovative processes when they occur in the context of the daily practice of medicine. The varying sources of medical ideas and the complexity of the innova- tive processes of medicine have an important implication for science policy: evaluative research must be closely linked to daily practice. They also suggest two goals and two processes of evaluation: 1. the goal of theory evaluation, by which alternative treatments for common conditions are tested in a comprehensive, systematic approach; and 2. the goal of process evaluation, by which the various configurations for packaging technologies and organizing human resources and levels of skill are explicated and evaluated.

INNOVATION AND THE POLICIES OF LIMITS 29 The first goal involves outcomes research, the second involves quality man- agement, and their linkage is the relationship between their ends and the means for achieving them. Let us briefly consider the requirements for building an infrastructure for the evaluative sciences in medicine. The first volume in this series discussed the various disciplines that constitute the evaluative sciences and the rationale for their introduction into mainstream thinking in medical schools (Wennberg, 1990b). It also discussed the policy basis for outcomes re- search and an organizational strategy, the patient outcome research team approach, for meeting the ongoing requirements for evaluation of estab- lished treatment theories, as well as innovations as they emerge. These teams PORTs, as they are becoming known in the United States-are part of the infrastructure being developed by the Agency for Health Care Policy and Research, a new federal agency that represents the first explicit effort on the part of government to articulate science policy for the evaluative sciences. There is a certain irony that public policy to rationalize health care should develop in the one nation among Western democracies that Brian Abel-Smith (1985) labels the "odd man out," the single example of a nation that has failed to establish policies of global limits on expenditures. The need for rational reallocation is most acute in systems of care in which marginal spending on innovation is inhibited by policies of global limits. Strategies for avoiding explicit health care rationing by reallocating excess capacity to meet unmet needs for effective medical care depends on the successes of the evaluative sciences in identifying examples of excess ca- pacity and establishing evidence that care is, indeed, effective. It should be much easier to build the necessary infrastructure in systems of care in which the societal commitment to set limits is in place once the problem of professional uncertainty and excess capacity is understood by policymakers. At least in principle, systems of care governed by policies that rationalize the deployment of manpower and budgets can redefine professional tasks much more easily than is now possible in the United States, with the impor- tant exception of prepaid staff-model HMOs such as the Kaiser Permanente Plan or Group Health Cooperative of Puget Sound. The reallocation of professional time and talent toward the two goals of outcomes research and quality management is a clear example of a potential advantage that Canadian or European models have over the United States. For the sake of argument, let us assume that the health care industry, like any other high-technology industry, should allocate 10 percent of its earn- ings to the development and testing of its products. In the health care field, investments of this order of magnitude are now made only by well-capital- ized pharmaceutical and medical device industries, and as the example of off-label uses of drugs such as prazosin shows, this does not lead to full

30 JOHN E. WENNBERG rationalization of even drug-related clinical theory. The problem is the vast undercapitalization of the evaluative function within the ongoing practice of medicine. Resources to evaluate innovations arising in clinical practice or to undertake the quality management tasks that in most industries are part of the production process have not traditionally been made available by government or the private sector. One of the most constructive steps con- sistent with the principles outlined in the previous section would be to allocate a substantial proportion of the health care budget to the task of building the necessary professional infrastructure in the evaluative sciences. Canada offers an example of what could be done in a system with global limits in place. Canadian physicians are currently paid much less on a per-procedure basis than their counterparts in the United States, resulting in longer work hours and greater productivity in terms of the numbers of procedures performed per physician; as a consequence, the rates of use of many common surgical procedures (and probably a good number of diag- nostic procedures) in Canada rival those in the United States, even though the supply of specialists is considerably less. Some of the excess capacity that Canadians now allocate to such procedures could be safely allocated to conduct outcomes research and build the professional infrastructure for man- agement of quality. Leading physicians who are interested in these tasks could be safely recruited from active practice without fear that the reduction in services would harm patients. (This could be done on a half-time basis to allow these physicians to remain clinically active.) Such an effort would not require a reorganization of the fee-for-service financial structure but only the willingness of the provincial government to negotiate salaries for physicians who chose to invest part of their professional time in this man- ner. Networks of recruited, professional talent linked to centers for the evaluative sciences, would form the infrastructure for a variety of evalua- tive tasks as well as the dissemination of results. Presumably, such a strat- egy would also be cost saving, because the total cost of care per active physician the stream of medical acts he or she initiates or sustains-is very likely to be much greater than the total cost of research per physician- investigator. A commitment to evaluation along these lines holds the prom- ise of rationalizing a spectrum of current inefficiencies, particularly in the management of quality and the explication of as yet unrecognized variations in the processes of production. An all-payer or single-payer model also offers another opportunity for rapidly increasing the level of sophistication of practicing physicians in the evaluative sciences, in particular their understanding of the relevance of evaluation for the everyday practice of medicine. Medical education has been primarily geared to the production of medical students and the training of medical residents; in situations of perceived manpower scarcity, this focus is quite natural. The current situation of excess capacity presents a

INNOVATION AND THE POLICIES OF LIMITS 31 new challenge and opportunity for medical education: to pay attention to ongoing learning requirements in a field with rapid technological change and to commit to a mission of lifetime learning in which skills are reshaped, knowledge rebuilt, and careers refashioned to meet changing needs. Again, in Canada, these policies are within reach. Just as the existence of excess capacity justifies the redeployment of professional talent to build capacity for the evaluative sciences, it also justifies periodic salaried sabbaticals from clinical practice for all professionals, including physicians, nurses, administrators and others. In theory, the British National Health Service provides similar flexibil- ity for the reallocation of professional workloads. Moreover, the special role of the British general practitioner, the unique responsibility he or she bears for initiating referrals, offers a splendid opportunity for development of the new model for the doctor-patient relationship based on shared deci- sion making. Rationalization of treatment patterns for specific conditions such as BPH ultimately depends on rationalization of referrals from primary care to specialty care on the development of what in the United States is called the cognitive role of the physician. It is no coincidence that governmental policy encouraging the evalua- tive sciences developed first in the United States. The issue of practice variations and the need to improve the scientific basis of clinical decision making have been prominently discussed in professional journals as well as in the lay press. The linkage of practice guidelines to outcomes research and the growth of the idea that micromanaged care will contain costs brought together the critical support needed for a new federal initiative, the Agency for Health Care Policy and Research, at a time of budget deficit and reluc- tance by Congress to take on new tasks. The tensions between the trend toward cost containment based on micromanaged care and the needs and requirements for rational innovation continue to grow. The implications of micromanaged care for the innovative processes of the pharmaceutical and medical device industry, as well as for surgical innovation, are now being widely discussed. In some cases, the emphasis on cost-effectiveness and reallocation will seem commensurate with the goals outlined here. But in other cases, the restrictions operate in the other direction and affect the weakest link in the evaluative process: the assessment of innovation within the context of everyday practice. At a time when the expansion of practice- based infrastructure to support innovation is needed, rules that restrict the funding of "experimental" technologies are being more rigidly enforced. Moreover, the increasing sensitivity to cost shifting, that is, the effort on the part of the purchaser to get the~"right price," penalizes most the academic medical centers that traditionally have been the most productive sources of medical innovation. This is unfortunate, given that the health care system's situation of excess and professional uncertainty requires just the opposite.

32 JOHN E. WENNBERG Nevertheless, things are changing. Whatever the shape of the new American health care economy, the policies of reform, if they are to pro- mote rather than retard medical innovation, must assume the obligation to build the scientific and ethical basis of clinical medicine and contain re- source consumption within limits acceptable to the wider society. The obligation to reform the scientific and ethical basis of clinical medicine can be summarized in four guiding principles: 1. knowledge about relevant treatment options should be freely com- municated to patients; 2. the choice of intervention from among options that work and that society is willing to provide should be based on the patient's preference; 3. the production of treatments should be continuously improved; and, 4. new as well as conventional treatment theories should be continu- ously assessed and reassessed. The opportunity to build a productive microeconomy, to keep the doc- tor-patient relationship free from intrusions by the state or by third-party micromanagement, depends, in turn, on a public policy for health that deals with the problem of limits and innovation. The challenge to the policies and politics of reform is to (a) set limits on the growth of supply; (b) reallocate excess capacity to productive purposes; (c) support the lifetime learning requirements of the profession; and (d) build the professional infra- structure required to learn what works in medicine and to produce services efficiently, free of supplier-induced demand. REFERENCES Aaron, H. J., and Schwartz, W. B. 1984. The Painful Prescription: Rationing Hospital Care. Washington, D.C.: The Brookings Institution. Aaron, H. J., and Schwartz, W. B. 1990. Rationing health care: The choice before us. Science 247:418-422. Abel-Smith, B. 1985. Who is the odd man out: The experience of Western Europe in containing the costs of health care. Milbank Memorial Fund Quarterly 63:1-17. Barry, M. J., Mulley, A. G. Jr., Fowler, F. J., and Wennberg, J. E. 1988. Watchful waiting vs. immediate transurethral resection for symptomatic prostatism: The importance of pa- tients' preferences. Journal of the American Medical Association 259:3010-3017. Brown, L. 1991. The national politics of Oregon's rationing plan. Health Affairs 10:28-51. Callahan, D. 1987. Setting Limits: Medical Goals in an Aging Society. New York: Simon and Schuster. Cochrane, A. 1972. Effectiveness and Efficiency. London: Nuffield Provincial Hospital Trust. Culp, W. J., Freeman, J. L., and Wenuberg, J. E. 1987. Are hospital services rationed in New Haven or overutilized in Boston? Lancet 1:1185- 1188. Eddy, D., and Billings, J. 1988. The quality of medical evidence: Implications for quality of care. Health Affairs 7:19-32.

INNOVATION AND THE POLICIES OF LIMITS 33 Fowler, F. J., Wennberg, J. E., Timothy, R. P., et al. 1988. Symptom status and quality of life following prostatectomy. Journal of the American Medical Association 259:3018-3022. Fox, D. M., and Leichter, H. M. 1991. Rationing care in Oregon: The new accountability. Health Affairs 10:7-27. Katz, J. 1984. The Silent World of Doctor and Patient. New York: Free Press. Lamm, R. D. 1987. Ethical care for the elderly. In: Smeeding, T. M., ed. Should Medical Care Be Rationed by Age? Totowa, N.J.: Rowman and Littlefield, pp. xi-xv. Wenuberg, J. E. 1990a. Small area analysis and the medical care outcome problem. In: AHCPR (Agency for Health Care Policy and Research) Conference Proceedings, Re- search Methodology: Strengthening Causal Interpretations of Nonexperimental Data. DHHS Pub. No. (PHS) 90-3454. Washington, D.C.: Department of Health and Human Services, pp. 177-206. Wennberg, J. E. 1990b. What is outcomes research? In: A. C. Gelijns, ed. Medical Innova- tion at the Crossroads. Vol. 1, Modern Methods of Clinical Investigation. Washington, D.C.: National Academy Press, pp. 33-46. Wennberg, J. E., Barnes, B., and Zubkoff, M. 1982. Professional uncertainty and the problem of supplier-induced demand. Social Science and Medicine 16:811-824. Wennberg, J. E., Mulley, A. G. Jr., Hanley, D., Timothy, R. P., Fowler, F. J., Roos, N. P. Jr., et al. 1988. An assessment of prostatectomy for benign urinary tract obstruction. Journal of the American Medical Association 259:3027-3030. ;

Next: PART II: MANAGING CARE IN THE UNITED STATES »
Technology and Health Care in an Era of Limits Get This Book
×
Buy Paperback | $55.00
MyNAP members save 10% online.
Login or Register to save!
Download Free PDF

The U.S. health care system is in a state of flux, and changes currently under way seem capable of exerting sizable effects on medical innovation.

This volume explores how the rapid transition to managed care might affect the rate and direction of medical innovation. The experience with technological change in medicine in other nations whose health care systems have "single-payer" characteristics is thoroughly examined.

Technology and Health Care in an Era of Limits examines how financing and care delivery strategies affect the decisions made by hospital administrators and physicians to adopt medical technologies. It also considers the patient's stake in the changing health care economy and the need for a stronger independent contribution of patients to the choice of technology used in their care.

Finally, the volume explores the impact of changes in the demand for medical technology in pharmaceutical, medical device, and surgical procedure innovation.

  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!