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 IV Implications for Providers
8 New Technology Adoption in the Hospital Paul F. Griner Medical technology adoption decisions in hospitals may occur through planned acquisitions or through uncontrolled changes in medical practice. They reflect a complex set of dynamics, with method of reimbursement being only one, albeit an important one, of those forces. These dynamics must be understood before there can be any hope of developing optimal approaches to the adoption of new technology. This paper first presents a working definition of new technology, fol- lowed by a discussion of the various factors that influence decisions regard- ing its adoption. The reader should bear in mind that these factors are described from the perspective of a hospital physician administrator. This perspective reflects a mixture of knowledge of the benefits and limits of technology on the one hand and awareness of its drawing power on the other. The discussion includes selected data related to the effect of these factors on patterns of technology adoption. The paper closes with several predictions for the future as a way of focusing and stimulating debate. NEW TECHNOLOGY: WHAT ARE ITS COMPONENTS? New technology may be arbitrarily classified into five groups: 1. new diagnostic or therapeutic equipment, such as linear accelerators; 2. expensive procedures for example, transplantation; 123
124 PAUL F. GRINER 3. pharmaceuticals the fastest-growing segment of the hospital econ- omy, comprising such subareas as biologics developed through genetic en- gineering and expensive antibiotics; 4. "supportive" technology examples are hospital information sys- tems and intensive care unit monitors; and 5. "hidden" new technology- changes in medical practice as a result of new knowledge that causes a hospital's existing technology, both diagnostic and therapeutic, to be applied in new ways. The last category of new technology represents unplanned changes in the day-to-day life of hospital practice that often have significant cost im- plications but that generally do not come to light until after the change has been in place for a period of time. Some examples are the use of magnetic resonance imaging (MRI) to evaluate the internal structure of the knee and the use of gamma globulin to treat immune disorders of the blood. FACTORS INFLUENCING THE ADOPTION AND USE OF TECHNOLOGY Any strategy designed to promote discriminating approaches to the adoption and use of new technology must recognize and take into account the influ- ence of certain factors (Table 8-1~. These factors are discussed in the sections below. Method of Financing New equipment and hospital renovation (e.g., to accommodate a new program) require capital. The first factor to consider in new technology adoption is thus the hospital's pool of reserves. Does the hospital have the funds needed to renovate space and acquire new equipment? Its endow- ment, its other sources of nonoperating revenue, and its cash reserves are the primary components of its pool of reserves. The extent of that pool is largely a function of the hospital's ability to generate surpluses from its TABLE 8-1 Factors Influencing the Adoption of New Technology 1. Method of financing "up-front" costs 2. Method of recovering incremental operating costs, including depreciation 3. Regulation 4. Level of competition 5. Capacity 6. Evidence of effectiveness 7. Hospital/medical staff organizational relationships 8. Decision-making processes
NEW TECHNOLOGY ADOPTION IN THE HOSPITAL 125 day-to-day operations and to generate nonoperating revenue from gifts and other sources. Operating surpluses, in turn, are largely a function of the level of rate regulation to which a hospital is subject.) In the underregulat- ed states of the Midwest, hospital cash reserves tend to be high; reserves of up to $150 million in some of the large midwestern teaching hospitals are not unusual. Conversely, in the heavily regulated hospitals of the North- east, operating surpluses are uncommon, and for many hospitals (e.g., most New York state hospitals), reserves are nonexistent. Cash for capital ex- penditures must therefore come from borrowing, and the hospital's borrow- ing capacity is determined by its overall financial position. It is not surpris- ing, as the data presented later in this discussion show, that new technology is introduced and disseminated more slowly in states that limit capacity through rate regulation. Recovery of Operating Costs Financing the initial cost of new technology is one thing; obtaining reimbursement for its operating costs can be quite another. Reimbursement is more complex than financing because there are two important variables involved: the level of regulation and the payers. Among states requiring certificates of need,2 approval may apply equally to all payers, including health maintenance organizations (HMOs), or just to Medicaid. New York is a good example of a state that requires a certificate of need; as one might imagine, adoption of new technology in this state is more conservative than in any other. Where certification of need is not required, other factors may then assume greater importance. The ability of the payment system to adapt to changes in the cost of a service may also influence the rate of adoption of new technology. For example, the diagnosis-related group (DRG) system associated with Medi- care appears to be a disincentive to the introduction of new technology because the rate of recalibration of DRGs for the incremental cost of the new technology occurs quite slowly (Kane and Manoukian, 1989~. In an- other example, one might anticipate that the problem of fixed-price reim- bursement for inpatient care will slow the introduction of some pharmaceu- ticals, particularly those developed through genetic engineering for highly specific uses at considerable cost. iRate regulation is the payment for care of an individual patient that a regulatory entity allows a hospital to collect. It may be based on a per-diem figure or it may be a fixed payment per admission according to the diagnosis-related group. 2In some states, certificate of need regulation requires that hospitals justify the need for a capital project (e.g., device, facility, or program) on the basis of community or regional need.
126 PA UL F. GRINER An additional aspect of recovering the cost of new technology is recov- ery of the cost of the initial investment, that is, capital cost recovery through depreciation, which should be distinguished from recovery of operating costs. By and large, regardless of the method of reimbursement, capital costs traditionally have been passed on to payers and reimbursed fully. This is likely to change, however, given the proposal by the U.S. Health Care Financing Administration (HCFA) to incorporate capital expenditures into the DRG-based reimbursement per diagnostic category. Although this ap- proach may help control costs overall, it has no mechanism to account for the legitimate capital costs of individual hospitals and is causing much concern among hospital administrators. Data are available to support these observations bearing on the relation- ship of capital reserves, operating surpluses, and regulation on the one hand and adoption of new technology on the other. There are 30 teaching hospi- tals in the states of New York, Massachusetts, Connecticut, and New Jer- sey, where hospital reimbursement is heavily regulated. The average oper- ating margin of these hospitals in 1989 was -$2.4 million, compared with an operating surplus averaging $6 million for the country's remaining teaching hospitals (Fishman et al., 1990~. In addition, annual capital expenditures for building improvements and equipment replacement, reasonable surro- gates for overall capital outlays, were 25 percent less for the 30 Northeast hospitals than such expenditures for similar hospitals throughout the rest of the country. Comparing upstate versus downstate New York hospitals reveals anoth- er factor that may influence the adoption of new technology, namely, the hospital's share of service to the poor. New York metropolitan hospitals had operating deficits in 1989 (-3.2 percent) that were 60 percent higher than their upstate counterparts (-2.0 percent). This difference in part re- flected more charity services by downstate hospitals and bad debt levels that were twice as high as levels for upstate hospitals (6 percent versus 3 percent, respectively). It is little wonder that New York City hospitals frequently complain of their inability to compete for new technology (Fish- man et al., 1990~. Regulatory Equity Fueled by technological advances and stimulated by reimbursement in- centives, much health care that previously involved hospitalization is now provided in ambulatory settings. What one might refer to as unbalanced regulation less regulatory oversight than is required for hospitals-has provided a further stimulus for the rapid growth of out-of-hospital technolo- gy (e.g., ambulatory surgical centers, imaging centers, ambulatory rehabili- tation centers). In many areas of the country, hospitals that wish to expand
NEW TECHNOLOGY ADOPTION IN THE HOSPITAL 127 their ambulatory services are at a disadvantage because of unbalanced cer- tificate of need requirements compared with outpatient facilities. Such facilities contribute significantly to the rate of inflation in Part B Medicare, a rate that is rising more than twice as fast as that of hospitals. This is an area that demands more regulatory attention. Level of Competition The level of competition among hospitals in a region is an important dete~inant in the adoption of new technology, particularly in areas with excess hospital capacity. This phenomenon occurs most often in regions with little regulatory control of capacity but with substantial growth in the use of managed care. Examples of such areas include Minneapolis, Chica- go, and California. Where competition is great, the principal focus of the hospital adminis- trator becomes the hospital's market share. High-technology services tend to attract patients, and the more services a hospital can offer, the more competitive it will be for exclusive contracts with employers and other third-party payers. Such circumstances support a tendency to develop du- plicate services considerably in excess of regional needs. As a result, man- aged care models do not appear to be as effective as regulation in control- ling the adoption of new technology. (The closed-panel Kaiser system may be an exception to this generalization because, in this form of managed care, both the hospital and its medical staff have incentives to acquire and use technology sparingly.) Few data are available to prove or disprove this hypothesis, although some information bearing on the number of hospital FTEs (full-time-equivalent employees) according to the reimbursement en- vironment may be acceptable as a surrogate, given that even new technolo- gy tends to be labor intensive. Fishman and colleagues (1990) present figures for the number of FTEs per occupied hospital bed, after adjusting for case mix, in four heavily regulated northeastern states, four underregu- lated midwestern states, and two states with a sizable proportion of man- aged care. The number of FTEs per bed in these regions are 5.2, 6.1, and 6.9, respectively. ~ ~ health care, regulation is more effective than competition in controlling These data support the hypothesis that in the field of costs. Worth sharing in regard to this issue are the experiences of the author and his colleagues (Block et al., 1987) during the 1980s in Rochester, New York, under an experimental hospital payment system. This payment pro- gram encompassed two extremes. One featured a community cap on capital expenditures; that is, all major capital projects required review and approval by the hospital coalition before subsequent review by the local health sys- tems agency and the state. During the period 1980 through 1987, the oper
128 PA UL F. GRINER TABLE 8-2 Rochester Hospital's Experimental Program Under Two Systems Effect Global Budget All-Payer Diagnosis-Related (1985-1987) Group (1988-1990) Revenue Predictable (pro); variably Variable; determined by volume adequate (con) and case mix Community Controlled Unpredictable; inflationary health care costs Incentive Introduction of new technology Incentive for ambulatory alternatives to inpatient care Physician + satisfaction Manage services Controlled (pro); rationed (con) Yes Obtain market share Available to all (pro); inflationary (con) No Access to + ++ hospital services Quality of care Good Cooperative spirit Good Good Constrained ating expenses of the participating hospitals were guaranteed but capped under a global budget, providing a ceiling on both operating and capital costs. In the other extreme of the payment system (1988-1990) the hospi- tals functioned under an all-payer DRG system and capital costs were not limited. Table 8-2 displays the advantages and disadvantages of these two sys- tems. Under the system of capped operating and capital budgets (i.e., the Global Budget column), revenue was predictable and hospital costs were controlled. From 1980 through 1987, hospital costs increased by 168 per- cent nationwide; in Rochester, the increase was 108 percent. The average Blue Cross premium in the United States in 1989 was $3,500; in Rochester, it was $1,600. Managers spent most of their time managing, new technolo- gy was introduced in a controlled fashion, incentives for ambulatory alter- natives to hospitalization were great, and adequate markers existed to con- firm that quality of care was good. Some physicians felt that access to hospital or high-technology services was being constrained, and hospital boards were concerned about the loss of sovereignty. These observations, in combination with a year when demands (and costs) of hospital services outstripped the dollars available, led to an interest in exploring the opposite alternative. Consequently, from 1988 through 1990, the hospitals were reimbursed under an all-payer DRG system with a capital pass-through.
NEW TECHNOLOGY ADOPTION IN THE HOSPITAL 129 Physicians were more satisfied, access was probably better, hospital boards were happier, and revenues increased at a higher rate; but community health care costs were higher, managers spent most of their time trying to gain market share (despite the observation that the rate of bed occupancy was high for all hospitals), considerable duplication of new technology occurred, and incentives for ambulatory alternatives to inpatient care virtually disap peared. The experience in Rochester is a classic example of the conflict be- tween the need to address the collective interests of society on the one hand and the interests of the individual on the other (the individual in this case being the patient, the physician, and the hospital all together). One system favored the greater social good; the other favored the individual. The goal must be a balanced approach that recognizes both needs. Ultimately, some combination of local, state, and federal regulation, together with price-driv- en competition, should promote that balance. Table 8-3 presents some data that reflect major differences in the intro- duction of new technology throughout the country, according to whether regulation or competition prevails (Fishman et al., 1990~. The four heavily regulated states mentioned earlier (New York, New Jersey, Connecticut, and Massachusetts) are compared with the country at large in terms of the proportion of institutions that perform various transplant procedures and another procedure (cochlear implants) whose efficacy remains to be deter- mined. Striking differences are apparent. Two-thirds of the less regulated teaching hospitals perform bone marrow transplants, but only one-quarter of similar hospitals in the heavily regulated states undertake the procedure. The proportion of teaching hospitals that perform heart transplants in less regulated states is 4.5 times the proportion of hospitals perfoll.ling them in the more regulated states (78 percent versus 17 percent, respectively). If one focuses only on those less regulated states in which managed care has a TABLE 8-3 Proportion of Hospitals Performing Selected Procedures According to Level of Regulation Procedure More Regulated Less Regulated (% of hospitals) (% of hospitals) Transplant Adrenal 0 8 Bone marrow 27 66 Heart 17 78 Liver 17 45 Cochlear implant 7 40
130 PAUL F. GRINER significant effect on patterns of practice (i.e., California and Minnesota), these differences appear to hold up. These data support the conclusion that strategies designed to limit overall capacity are more effective than compet- itive strategies, including managed care, in controlling the introduction of new technology. Other Factors That Influence the Adoption of New Technology The four factors discussed above method of financing a technology's initial cost, method of recovering operating costs, level of regulation, and degree of competition are the principal factors influencing the decisions of hospital administrators regarding the adoption of new technology. There are a number of others as well, the hospital's capacity for technology as- sessment being one. Some hospitals (e.g., Johns Hopkins) have an office of technology assessment, which requires medical staff to support initiatives for new technology with evidence of scientific efficacy before other consid- erations are explored. At Strong Memorial Hospital in Rochester, New York, more than $1 million was saved over 4 years through the application of a protocol that restricted the use of tissue plasminogen activator (t-PA) to patients for whom streptokinase was not appropriate. The protocol was developed by a process of consensus among interested faculty after review- ing the relevant literature on the subject. Such decisions, which are not without political risk, require the medical staff to accept responsibility for prudent use of the hospital's resources. Hospital-medical staff organizational relationships are yet another fac- tor bearing on the adoption of new technology. Until recently, the hospital has been expected to provide all of the resources needed for its medical staff to deliver patient care. But as the availability of hospital capital decreases, interesting financial relationships are developing in some instances between hospitals and medical staff. Some examples are joint ventures for the acquisition of and operation of new technology, or arrangements where- by staff members purchase equipment and are responsible for its operating costs, while the hospital advances money for working capital, applying standard banking practices for its repayment. The locus of decision making in the hospital is becoming an important aspect of new technology adoption decisions. The ultimate responsibility for policy rests with the hospital board. In cases in which hospital boards exercise that authority, the adoption of new technology is determined large- ly by affordability. In instances in which hospital boards still serve in a pro forma mode, medical staff pressures may override financial reality testing and local or regional needs. Finally, in many teaching hospitals, the re- search activities of faculty lead to the development and introduction of new diagnostic and therapeutic technology.
NEW TECHNOLOGY ADOPTION IN THE HOSPITAL Hidden New Technology 131 So-called hidden new technology refers to changes in the application of technology in hospital practice, which occur almost daily and which often have significant cost implications. Such changes occur in an unplanned way and result from the generation and dissemination of new knowledge bearing on clinical diagnosis and management (e.g., new applications of MRI, new uses of existing expensive pharmaceuticals such as gammaglobu- lin for the treatment of immunologic disorders). The pharmacy budget of Strong Memorial Hospital in Rochester, New York, has increased by 75 percent in fewer than 5 years. Across the country, hospital pharmacy bud- gets are the fastest growing segment of the hospital economy. Because of these trends, hospitals are beginning to put in place more comprehensive drug surveillance systems to monitor changes in medical practice and to ensure that these new practices adhere to reasonable standards of cost- effectiveness. Hidden technology encompasses many other aspects of the day-to-day care of hospitalized patients. Changes in automatic infusion pumps, hospi- tal beds, cancer chemotherapy protocols, monitors/defibrillators, and pulse oximeters are but a few of the many enhancements to care that are intro- duced daily throughout the hospital, usually without a management master plan. SUMMARY The following sums up a combination of facts and the author's personal experience regarding the adoption of new technology. 1. New technology is introduced more slowly in areas of the country in which health facilities are heavily regulated than areas in which competitive strategies are promoted. 2. A regional capital expenditures cap tied to a global operating budget appears to be the most effective method of controlling capital costs. The Rochester experience bears this out, although this paper does not include all the supporting data. 3. The single most important determinant of a hospital's adoption of new technology is the hospital's operating margin. This statement holds true both when payment programs are designed to control capacity and when they are designed to stimulate competition. 4. Hospitals are beginning to recognize the many other factors that influence the adoption of new technology for example, changes in organi- zational and financial relationships with medical staff, development of in- house technology assessment capabilities, and the increasing role of the hospital's governing body in decisions about major hospital expenditures.
132 PA UL F. GRINER 5. Two predictions for the future: one that is already becoming clear is that throughout the 1990s, the adoption of new technology will occur more slowly as hospital operating surpluses continue to fall. The other prediction is that hospitals and their medical staffs will have to develop innovative organizational and financial relationships involving shared risk. These rela- tionships will ensure that the adoption of new technology is the result of a balance between the hospital's competitive desire for excellence and the region's interest in helping to achieve affordability by ensuring that services that are not actually needed are not made available. REFERENCES Block, J., Regenstreif, D., and Griner, P. 1987. A community hospital payment experiment outperforms national experience: The hospital experimental payment program in Roches- ter, NY. Journal of the American Medical Association 2:193-197. Fishman, L. E., Serrin, K. G., and Bigelow, J. S., eds. 1990. Council of Teaching Hospitals: Survey of Academic Medical Center Hospitals' Financial and General Operating Data, 1989. Washington, D.C.: Association of American Medical Colleges. Kane, N. M., and Manoukian, P. D. 1989. The effect of the Medicare prospective payment system on the adoption of new technology: The case for cochlear implants. New En- gland Journal of Medicine 20:1378-1383.
9 Physicians' Acquisition and Use of New Technology in an Era of Economic Constraints Bruce I. HilIman Physicians perceive a growing harshness in the health care environ- ment. They are concerned that the policy initiatives of the 1980s and the apparent increasing stringency of the coming decade will negatively affect both their ability to deliver what they believe to be optimal care for their patients and their own incomes and enjoyment of medical practice. Central to these concerns is the access of physicians to medical technol- ogy. In prospect, a physician's ability to acquire medical technology de- pends on the regulatory, reimbursement, and competitive milieus in which he or she practices, as well as on the perceived technical and clinical poten- tials of the technology. This paper discusses to what extent physicians might be influenced by current health policies and the expected medical organizational environment with respect to their decisions to acquire and use new technologies. The paper focuses specifically on the acquisition of major medical devices, because these technologies have been especially scrutinized by health policy researchers and policy decision makers. This discussion is composed of three parts. First, in the belief that recent past experience might presage what will occur in the immediate future, the paper presents earlier research describing how physicians re- sponded to environmental influences in deciding whether to acquire a new medical technology of the 1980s magnetic resonance imaging (MRI). The second part addresses the changes that have occurred since the early diffu- sion of MRI. The third section considers how these changes might selec 133
34 BRUCE J. HILLMAN lively influence physicians who are considering the acquisition of some future new technology. DIFFUSION OF MAGNETIC RESONANCE IMAGING INTO CLINICAL PRACTICE This section of the paper discusses research on nonmedical influences that have affected MRI acquisition (Hillman, 1986; Hillman et al., 1986, 1987a,b). During 1984-1985, 3 years after the first MRI scanner was intro- duced in the United States, investigators at the RAND Corporation studied the diffusion of MRI. Physicians at that time were, as now, concerned about the development of a much more restrictive budgeting environment than they had previously confronted. Prospective payment for hospital in- patient care had just been introduced for Medicare patients, and many states still had vigorous health planning bodies. The Food and Drug Administra- tion (FDA) had designated MRI as a class 3 technology-requiring exten- sive premarketing tests and approval before general diffusion- and there was virtually no third-party reimbursement for the procedure. The prevail- ing opinion was that these conditions, partly by the design of policymakers and partly as a result of circumstances, would reinforce the already signifi- cant uncertainties surrounding the technology and act as potent disincen- tives to physicians and other providers that might be considering MAT n quisition. ~ 4~ ~. ~ ~=~ A ~ _ ~ _ _ ~ ~ ~ V ~ ~ _ ^ ~ 4 4 ,5 ~ ~ ~- IVIK! appeared on the health care scene to proffer important benefits that were not supplied by other technologies: imaging without ionizing radiation or injected or ingested contrast material; improved differentiation of body tissues; and the promise of new types of information concerning metabolism and physiology. Physicians were excited by these possibilities and their potential to benefit patients' health. But there were also major uncertainties: how high the acquisition and operational costs of the tech- nology would be; which version of the technology would become ascen- dant; where the technology would best be located; and what MRI would add to what could be learned about patients' problems by using already well- understood technologies (e.g., x-ray, computed tomography ECT], ultrasound). Clearly, many of the uncertainties reflected, in part, a concern that influences external to characteristics of MRI in the health care milieu might restrict potential acquirers of MRI in their ability to successfully operate MRI facilities. To determine to what extent such restrictions were felt, the RAND investigators adopted a case study approach. They performed 83 extended, guided, open-ended interviews with individuals interested in MRI- related issues manufacturing, selling, regulating, acquiring, and reimburs- ing. Thirty-seven of the interviews were with potential acquirers of MRI; almost uniformly, these interviews represented physicians' views. This
PHYSICIANS ' ACQUISITION AND USE OF TECHNOLOGY 135 paper explores how regulatory, reimbursement, and competitive influences affected physicians' consideration of MRI acquisition. Regulation of Magnetic Resonance Imaging Acquisition Two major regulatory factors were extant during the early diffusion of MRI: the responsibilities of the FDA in approving new medical devices, and health planning, predominantly in the form of state certificate-of-need (CON) activities. The Food and Drug Administration MRI was the first major technology to be designated class 3 by the FDA under the 1976 Medical Device Amendments to the Food, Drug, and Cosmetics Act. This classification required MRI manufacturers to submit their devices to rigorous testing to ensure the safety and efficacy of their technology. Rather than evaluate MRI devices as a class of technologies, the FDA decided to consider each manufacturer's device and each alteration of each device separately, portending a very extended process. In principle, prior to obtaining FDA certification, a firm could neither advertise its de- vice nor sell its devices for profit. Facilities were also affected: they were prohibited from charging patients more than their actual costs. Despite these constraints, marketing proceeded with little interruption. At the time of the RAND study, the FDA had just issued its first premarket- ing approvals; the devices of most companies had not yet been approved. Nevertheless, the RAND investigations could discern no important differ- ences in the advertising of companies with and without approval. Manufac- turers without approval advertised MRI "generically" in medical journals and magazines but made clear their views on characteristics that reflected their own approach to the technology. Corporate "detail" personnel touted real or imagined advantages of their technology vis-a-vis others and prom- ised impending FDA approval and third-party reimbursement. In actuality, FDA approval processes and the decisions on reimbursement from a prolif- erating number of independent payment entities would require considerable time, and no one could actually predict when they might be concluded. To develop the data necessary to obtain FDA approval, each manufac- turer developed alliances with a number of clinical sites hospitals and radiologists' practices, for the most part. The manufacturers provided vari- ous incentives to physicians to participate in their data-gathering activities, including discounting the price of the MRI scanner or, in some cases, even providing the technology free of charge. Sometimes, because there was virtually no third-party reimbursement for MRI, manufacturers also paid a stipend for performing scans. In other instances, physicians and institutions
136 BRUCE J. HILLMAN were designated as test sites even though they provided little data; the RAND team believed these were actually "sales" disguised as premarketing approval activities to circumvent FDA regulations. Those interviewed by the RAND investigators generally responded quite positively to MRI. Physicians believed that although the FDA approval process clearly would be drawn out, eventually all the devices would achieve approval. The basis for this view related to the charge of the FDA itself. Physicians knew that MRI was safe there was already considerable experi- ence with the fundamentals of the technology and with patients, both in the United States and elsewhere. The FDA's responsibility for determining efficacy hinges on whether the device does what a manufacturer's labeling claims it will do, which in the case of MRI is to provide images of normal and pathologic anatomy. Again, more than sufficient information was available to potential acquirers that this was the case. Representatives of health maintenance organizations (HMOs) were the only providers interviewed who were more restrained in their enthusiasm for MRI. However, the views of these individuals were more a reflection of corporate fiscal philosophy than any specific concerns related to FDA regulation. As far as the RAND researchers could determine, concerns over FDA approval inhibited no physician or facility interested in MRI from acquiring the technology. Indeed, the need of manufacturers to establish relationships with providers to obtain premarketing approval may well have enhanced early diffusion. Certificate-of-Need Regulations State regulatory statutes, predominantly those involving CONs, had a quite different effect on MRI diffusion, depending on the stringency of a particular state's requirements. CON regulation was empowered by the Health Planning and Resources Development Act of 1974 and was directed particularly at providers to constrain their building of new facilities and their acquisition of major technologies. The act empowered states to establish health systems agencies to administer regulations that the states individually would enact. By the time of the RAND study, shortly after these agencies had been defended by the Reagan admin~s~abon, CON at minis~abon varied enormously among the states in terms of scope and effectiveness (Salkever and Bice, 1976; Bice and Urban, 1982; Brown, 1983; Steinberg, 1985~. ICON regulations most particularly affected the ability of hospitals to add facilities or acquire expensive technologies by setting a maximum amount that a hospital could spend without first requesting the permission of the state's health systems agency (by filing a CON application). In all but a very few states, outpatient facilities and physicians' offices were exempt from this process.
PHYSICIANS ' ACQUISITION AND USE OF TECHNOLOGY 137 To deal efficiently with the variation among the states, the RAND in- vestigators concentrated their efforts in five states that in their view spanned the regulatory spectrum: California and Arizona, which considered them- selves "free markets," providing no serious regulatory obstacles to technol- ogy acquisition; Illinois and New York, which tightly controlled hospital acquisition of technologies and planned assessments of MRI at a small number of hospitals chosen to receive the technology-but which did not attempt to control the acquisition of scanners in outpatient settings; and Massachusetts, which effectively constrained acquisition of MRI by all pro- viders through a combination of CON (called DON, or determination of need, in Massachusetts) and payment rate regulation. Massachusetts had decided to permit initial acquisition of MRI by only eight hospitals. In California and Arizona, physicians perceived or actually had no dif- ficulty in obtaining the technology none, that is, that could be laid to the door of state agencies. A telling observation in the Los Angeles area, where RAND is located, was that one MRI was planned for every mile of a 6-mile stretch of Wilshire Boulevard; by the end of 1985, there were be- lieved to be 25 MRI scanners planned for or operating in Los Angeles and its environs. The motivation for MRI acquisition in California and Arizona was largely competitive and is discussed below. In Illinois and New York, the intent was to plan hospital acquisition according to the results of assessments of the clinical utility of the device. Initially, relatively few hospitals would be allowed to acquire the technolo- gy, and it was to be these few installations that would provide the data for further CON approvals. The states, however, controlled acquisition of the technology by hospitals but not by outpatient facilities, and this quickly led to the hospitals becoming competitively disadvantaged. A further factor was that this development occurred while hospitals were busy accommodat- ing the revenue reductions that accompanied prospective payment under Medicare and the attempts of other payers to reduce cost shifting (i.e., increasing charges to some payers to compensate for reduced payments by others). The confluence of these phenomena put hospitals, the traditional leaders in new technology acquisition, at a serious disadvantage, thus open- ing a new niche for other prospective providers of MRI services. As a result, the RAND team reported that new entities not previously engaged in providing medical technology services became interested in MRI, and new types of financial arrangements were developed. With hospitals in states with active CON processes being placed at a disadvantage in terms of acquiring MRI, new sources of capital were need- ed to finance such acquisitions. The arrangements, which included venture capitalists, health care corporations, and, to a greater extent than in the past, physicians, seemed to be based more on the opportunity for financial gain than on a desire to deliver a medical service. Because of the considerable
38 BRUCE J. HILLMAN financial risk posed by acquisition of such an expensive technology, the participants in these arrangements sought to secure enough of a patient base to make the imaging venture successful. Physician investors were recruited on the basis of their ability to refer patients for MRI services, imaging specialists were excluded because they provided no such benefit to the partnership, and there were reports of facilities basing returns on investment on the number of referrals made by the investor. Providers in Massachusetts understood that to acquire MRI they would have to convince the state that they offered special advantages over other candidates. The state would permit the purchase of relatively few scanners, and the process of selecting those purchasers would be an extended one. Providers thus reasoned that their chances of access to MRI would be en- hanced if they went into the CON process along with other providers. Mas- sachusetts was the only state in the RAND study in which consortia were formed to acquire MRI. This pattern existed despite experience with com- puted tomography scanners suggesting that consortia! operations of high technology posed considerable difficulties (Brust et al., 1981~. Regardless of whether a hospital intended to actually operate an MRI facility, it nevertheless filed a CON application, expecting that by the time the process resulted in its selection, perhaps many years hence, it would be ready for the technology. (This trend was observed in hospitals with more than 200 beds.) In essence, filing a CON application was viewed as the means of holding a place in line. This practice, however, imposed an additional burden on an already ponderous and expensive bureaucracy, drawing out the selection process. Indeed, many argued that the cost of the bureau- cracy in Massachusetts outweighed the savings of limited diffusion, while providers decried poor patient access to the benefits of MRI. Incontrovert- ibly, however, whether or not it was to the benefit of the populace, CON was effective in limiting the diffusion of MRI in Massachusetts. Reimbursement for Magnetic Resonance Imaging MRI began to diffuse at a time when third-party payers were in the first stages of confronting increasing financial constraints. As noted earlier, the federal government had just instituted prospective payment for Medicare inpatients. In addition, policymakers were discussing whether to institute prospective payment for all physicians or, alternatively, if this proved infea- sible, at least for hospital-based physicians. Research was being conducted to evaluate these possibilities (Mitchell, 1985; Ginsburg et al., 1986~. Oth- er third-party payers, faced with rising health care expenditures, were con- sidering the adoption of the Medicare initiatives. They were also raising their rates to major benefits purchasers, but at the same time they were under pressure from those purchasers to identify ways to reduce their costs.
PHYSICIANS' ACQUISITION AND USE OF TECHNOLOGY 139 On another front, health maintenance organizations were becoming serious competitors, recruiting patients from traditional indemnity insurance plans. In general, reimbursement agencies viewed MRI suspiciously. It had been less than a decade since they had noted the increased expenditures associated with the introduction of CT scanning; although by this time it was acknowledged that CT had been a beneficial, possibly even cost-effec- tive advance (Banta, 1980; Evens, 1980), the feeling remained that a great deal of waste had accompanied the introduction of the technology. Pur- chasers seemed aware of the potential benefits of MRI, but most were not convinced that the advances demonstrated to that point were significant enough and different enough from CT to warrant coverage. At this stage in the diffusion of MRI, there was little scientific evalua- tion of the technology; what had been published, although considerable, was fraught with biases and dealt in large part with small, selected patient samples (Cooper et al., 1988; Kent and Larson, 1988~. A Blue Shield Association assessment based only on the peer-reviewed literature indi- cated that the benefits of MRI were insufficiently proven and recommended against general coverage; MRI proponents argued that the methodology of the study was biased and the findings outmoded. The Health Care Financ- ing Administration (HCFA) began an extended evaluation to determine its coverage policy. Uncertainty remained about how capital reimbursement would be handled under prospective payment. Interviews conducted by the RAND investigators with prospective ac- quirers of MRI elicited much concern and anxiety over coming alterations in the reimbursement environment. The technology was more expensive than any previously considered for acquisition--as much as $3 million to purchase and site and up to $800,000 per year to operate. Nonetheless, in no instance did the RAND team find that providers had been dissuaded from acquisition by considerations related to reimbursement. The general belief was that despite what most considered to be "brave talk," resistance to reimbursement on the part of third-party payers was uncoordinated. Pay- ers had never before been able to decline payment for the use of major new technologies, and they were unlikely to be able to do so with MRI. This was particularly true with respect to HCFA; prospective MRI acquirers felt that the agency had delayed too long for it to exercise any option but general coverage. Diffusion and experience with MRI were too extensive and the lay, professional, and corporate pressures favoring coverage for most uses were too powerful to resist. The RAND researchers found little evidence that either the reimburse- ment initiatives of the early 1980s or the threat of further impending reim- bursement innovations was retarding MRI diffusion. Most of the providers interviewed noted that in the absence of third-party reimbursement, MRI was likely to lose money, at least in the short term. To minimize such
140 BRUCE J. HILLMAN losses, acquirers were instituting new payment requirements that they con- ceded they had never considered in the past, including payment by the patient prior to scanning and reduced "charity" care. Certain elements of the reimbursement milieu-the uncertainty over how capital reimbursement would be handled under the diagnosis-related group (DRG) system and a reduced ability of hospitals to shift costs- probably promoted early acquisition decisions. (Providers hoped by early acquisition to be "grandfathered" under the old rules.) Concerns over reim- bursement probably affected some siting decisions, as well as who would own and operate the devices. As with CON regulation, reimbursement considerations produced a disadvantage with regard to acquisition of tech- nology by hospitals. In concert with the tax advantages favoring for-profit ownership, these issues promoted outpatient siting of MRI. Competition Over Magnetic Resonance Imaging The RAND investigators evaluated competition among providers in the same five states in which they had assessed the effects of regulation. In states other than Massachusetts (where, as noted earlier, regulation effec- tively constrained access to the technology), there was ardent competition over MRI. Three forms of competition can be identified: competition to provide MRI services, competition over patients, and competition among specialties for control of the technology. The health care policy initiatives of the 1980s were designed, in part, to introduce price competition into medicine. Yet the RAND study team found no instance in which providers considered reducing prices as a means of attracting patients to MRI services. Rather, providers expressed the inten- tion of competing on the basis of the service and developing their marketing efforts more fully than they had ever done before. The goal of such marketing was to use MRI as a competitive instrument to enhance utilization of a provider's other services. Experiences with CT scanning had given most providers a healthy respect for the power of a new technology to imbue them with a desirable image. Consequently, they viewed MRI as a tool in their battles with local competitors to attract referring physicians, and hence patients, a strategy that would eventually lead to their controlling a larger share of the local market. Their belief was that if they could use MRI to get patients "in the door," they might "capture" them for their other, more remunerative services. A number of early MRI purchas- ers, who had acquired the technology on this basis, were disappointed when they found that local physicians, concerned over the possible loss of their patients, were reluctant to refer them for scans. Competition among MRI facilities over referring physicians consequently took a peculiar turn. Given the very high financial risks associated with
PHYSICIANS' ACQUISITION AND USE OF TECHNOLOGY 141 such an expensive technology, coupled with expected coming financial con- straints, facilities vied for the patronage of the physicians most likely to have a high number of potential referrals for MRI scans: neurologists, neurosurgeons, and orthopedists. Frequently, providers offered these physi- cians special inducements to refer patients to their facilities. Most com- monly, the physicians became limited partners in facilities being operated as entrepreneurial ventures or even in joint ventures with hospitals. Despite concerns that such financial involvement might inappropriately influence referrals, lead to abuse of the technology, and infringe on traditional medi- cal ethics (Relman, 1985; Hyman and Williamson, 1989; Morreim, 1989), this practice appeared to be growing rapidly in popularity. (See also the series of letters to the editor and editor's response in the January 23, 1986, issue of The New England Journal of Medicine EVol. 314, No. 4, pp. 250- 2531.) The competition for physicians and their patients, which was intend- ed to secure a sufficient patient stream to enable financially stable MRI operation, in fact strained traditional relationships among providers and promoted early MRI acquisition. In addition to competition in providing services and for patients, the RAND researchers observed enhanced competition among specialties over the control of MRI. Turf battles over technology are not new in medicine; this one, however, appeared to be more contentious than previous struggles, possibly because of the high financial and scientific stakes associated with MRI, in concert with physicians' concerns that their ability to sustain their incomes by providing traditional services might be compromised by the more draconian coming milieu. Radiologists, the specialists traditionally responsible for medical imaging, were the earliest and most frequent pur- chasers of scanners. They were also the physicians who most often advised hospital administrators and were most instrumental in helping to foul ate administrators' views. However, radiologists do not primarily care for pa- tients; they depend for referrals on other physicians, which puts them at a disadvantage in the more entrepreneurial context of MRI. Many of the providers interviewed in the RAND study indicated that they were uncomfortable in proceeding with MRI acquisition but that the competitive environment dictated such an action. Their view was that they gained a competitive advantage by being the first MRI facility in their locale. (Being second meant that fewer of the most desirable referring physicians would be available.) Yet in no instance did the presence of an existing facility dissuade a provider from pursuing MRI acquisition; in these cases, providers cited the existence of other facilities as evidence that if they did not hurry their acquisition, they might be left out entirely. Overall, the RAND investigators concluded that competition among providers was a potent force encouraging the early diffusion of MRI technology.
42 BRUCE J. HILLMAN Summary: The Acquisition of Magnetic Resonance Imaging From the foregoing, it is evident that physicians- whether representing themselves or an organization-were largely unaffected by the intended thrust of regulatory, reimbursement, and competitive influences in what they themselves believed to be a harsh environment for technology acquisi- tion. Perhaps they simply did not understand the full implications of the policy initiatives opposing MRI purchase. It is likely, however, that physi- cians' decisions were guided by two considerations: first, although hostile, the forces opposing technology acquisition were still disorganized and were likely, in the end, to be ineffectual; and second, physicians in the past had always found a way to access new technology- MRI would be no different. The physicians were correct in these surmises. The technology was se- ductive, and the policies that had been intended to retard its acquisition were ineffective in the face of significant professional and lay demand for MRI. TECHNOLOGY ACQUISITION IN THE "NEW HEALTH CARE" ERA There is a general perception among physicians that the times have grown tougher, especially over the past few years. Despite the recent expe- rience with MRI, many observers believe that physicians might be more responsive now to environmental influences opposing technology acquisi- tion than they were in the 1980s. Certainly, imposing barriers have been erected that might well dissuade physicians from purchasing an expensive new technology similar to MRI. But there are also aspects of the milieu that might reasonably encourage acquisition. Table 9-1 lists various aspects of the health care environment that might promote or retard technology acquisition. As the table indicates, several might be expected to have both positive and negative effects. This section considers how particular influ- ences might be expected to affect technology acquisition by physicians in the coming years. TABLE 9-1 Influences in the Environment That Promote or Retard New Technology Acquisition . _ Promoting Retarding Nature of technology Hospitals and regulation Competition for patients Providers' turf Nature of technology Hospitals and regulation Competition for patients Providers' turf Reimbursement Managed care
PHYSICIANS' ACQUISITION AND USE OF TECHNOLOGY Nature of Technology 143 One important aspect of recent technological innovation has been an emphasis on developing technologies that are safer and less invasive. MRI is a good example, as are the development of digital subtraction angiogra- phy for outpatient arteriography, ambulatory dialysis, and technological modifications that permit more surgery to be performed on an outpatient basis. In part, the emergence of this trend is related to the realization that medical technologies have an important role in improving the comfort of care and the quality of life, as well as in extending its length. The develop- ment of less invasive technology is also a response to the policy environ- ment, which has favored more flexible siting, particularly siting outside of hospitals. That newer technologies are amenable to outpatient siting is likely to continue to encourage their acquisition. Outpatient use of technol- ogy is still relatively free of regulatory constraints and enjoys more favor- able reimbursement than inpatient use. The high cost of many new technol- ogies is a potential deterrent to their acquisition but, as with MRI, physicians and other entities still find significant incentives for financial investments in outpatient medical services. The sole disincentive in this regard is the recent congressional prohibition against physicians having ownership inter- ests in outpatient laboratory facilities to which they refer patients and the threat that in the future this ban might be extended to other types of medical facilities. Hospitals and Regulation Although some states still maintain CON policies, the overall sense is that health planning generally is weakening. Although this trend might have portended the reemergence of hospitals as aggressive early acquirers of new technology, there are few signs of such activities. Indeed, hospitals now appear more financially debilitated than they were even in the previous decade. They remain disadvantaged under current reimbursement policies and are less able to shift costs among payers, given the increasing volumes of contracted and managed care. As a result, there is a persistent, advanta- geous niche for entrepreneurial technology acquisition in outpatient settings by physicians and other entities, either jointly with or exclusive of hospi- tals. These potential providers of technologically advanced services still have access to capital, tax incentives, and mechanisms to ensure a sufficient volume of patients-most notably through self-referral and to encourage new technology acquisition (ECRI Technology Management Assessment, 1985~.
144 BRUCE J. HILLMAN Competition for Patients Physicians' incomes have kept pace with or exceeded inflation over the past decade, despite declining patient rolls (Starr, 1982~. This pattern has been largely attributed to increasing intensity of care, fueled in part by the assimilation of technologies into office practice (Holohan and Zuckerman, 1990~. Demand by patients for the use of innovative technologies and their appreciation of the convenience of receiving all of their care in their physi- cian's office have further encouraged these trends. New reimbursement policies now seek to reduce payment for technological care but do little to reduce utilization by individual physicians, who often have a financial in- centive to continue or even increase the use of technology. A significant counterbalancing influence, however, is that even independent physicians increasingly participate in managed care, either as members of preferred provider organizations or as contractors to HMOs. In this situation, physi- cians must compete for patients on the basis of price as well as on the services they offer. Presumably, they will have to consider whether the overhead they will assume in purchasing a technology will diminish their price competitiveness. Reimbursement A number of recent reimbursement initiatives might be expected, indi- vidually and cumulatively, to affect the ability of physicians to acquire and use new technologies profitably. Perhaps the most significant is HCFA's implementation of a resource-based fee schedule for physician payment under Medicare, which replaces the traditional "reasonable, customary, and usual" method. The major avowed intent of this change is to reduce com- pensation for what HCFA considers overcompensated, technology-based services in favor of more time-intensive activities, such as taking patient histories and performing physical examinations. Other payers also are currently, or are considering, implementing fee schedules with similar goals. In addi- tion, reliance on Recertification, fee ceilings, and limits on balance billing is growing and would be expected to influence technology acquisition and use adversely. Capitated payments for bundled caret are increasingly pop- ular with HMOs and clearly make the use of technology financially disad 2The provider assumes financial risk for the care of patients by accepting a lump sum payment to provide a "unit" of service. This "unit" is often a period of time, such as a year. If the provider spends less on care than the lump sum amount, he or she profits; if more, the provider experiences a financial loss. Thus, under capitation, there is a fiscal disincentive to employ medical technology, particularly if it is deemed to be of marginal benefit.
PHYSICIANS ' ACE UISITION AND USE OF TECHNOLOGY 145 vantageous. Many physicians fear that if capitation becomes the dominant payment mechanism, as many believe it might in the future, it would inap- propriately curtail modern technological medicine. Since the late 1970s, payers have voiced concerns about whether their outlays for services that require new high-technology devices return conso- nant benefits in improved health. A central concern is that little assessment of such technology occurs before its general diffusion. Certainly, this was the case with MRI: aside from the studies required for FDA certification, no scientific assessment occurred for at least 5 years after its introduction into the United States, and little has occurred as of today (Cooper et al., 1988; Kent and Larson, 1988~. The resultant uncertainties regarding appro- priate application generate unnecessary costs in excess capacity, duplicative studies, financially motivated abuse, and early obsolescence. Payers have threatened to tie reimbursement more closely to scientific evidence that using a new technology improves patient health. Efforts in this regard, however, have been uncoordinated, underfunded, and, as in the case of MRI, unsuccessful. In 1989, the federal government established yet another in a line of agencies intended to foster and coordinate technology assess- ment-the Agency for Health Care Policy and Research (AHCPR). Unlike its predecessors, however, a major focus of AHCPR's activities is the inves- tigation of the outcomes of medical practice; it is expected to advise HCFA on decisions concerning Medicare coverage. Although AHCPR has not received the level of funding that was originally expected or that would be sufficient to address this charge in a serious fashion, there remains an ex- pectation that, to reduce their costs, both private and public payers must soon consider patient outcomes to a greater extent in their coverage deliber- ations. Should this occur, the period necessary to assess the effectiveness of a new technology and physicians' concerns over eventual reimbursement might be sufficiently great that diffusion would be slowed. Managed Care In the article by Wennberg in this volume, a distinction is made be- tween "micromanaged" and "global limits" types of managed care. Cur- rently, in the United States, a continuum of forms of managed care reflect various combinations of aspects of the philosophies of micromanaged care and managed care with global limits. Managed care in some form, includ- ing HMOs, preferred provider organizations, and utilization controls, now accounts for nearly 40 percent of the health care market. It is expected to continue to grow perhaps to encompass as much as 70 percent of the U.S. population. One intent of managed care is to reduce the use of marginal and inappropriate technology. In staff-model HMOs, an example of a glo- bal limits approach, this reduction is accomplished in part by queuing and
46 BRUCE J. HILLMAN by providing financial and other incentives to member physicians to use the community's technology resources less frequently. The growing success of this type of managed care entity has allowed it to assimilate more technolo- gies into its own operations, which offers tighter internal control of utiliza- tion. Virtually all managed care agencies seek to reduce their technology- related outlays by encouraging potential contract providers to competitively discount their fees and by paying for bundled services, such as in capitation arrangements. Indemnity insurers, as well as HMOs, have increasingly relied on micromanagement, including precertification and the development of standardized approaches to clinical presentation, to control what they view as marginal technology utilization. The emphasis of managed care on cost containment presages a more cautious approach to new technology acquisition and utilization in the future. Providers' Turf The personal motivations of physicians for acquiring a new technology usually include intellectual enjoyment, the pride associated with being on the "cutting edge," the respect of one's peers, being able to apply the tech- nology to benefit patients, and the possibility of greater financial remunera- tion (Hillman et al., 19849. The confluence of a number of the factors described in the preceding sections impinges on the issue of which groups of physicians can most benefit from technology acquisition in the current organizational climate and in the expected climate of the future. Financial concerns are particularly acute and increasingly may be the motivation for a physician's acquisition of a particular technology. In 1990, an article in the Wall Street Journal entitled "Warm Bodies" described how physicians who previously had referred patients to radiologists for diagnostic imaging stud- ies were now attempting to coerce these same radiologists into "kicking back" a portion of the payments from these referrals. The threat that consti- tuted the coercion, which had materialized in some cases, was that failure to do so would result in the referring physicians' establishing their own facili- ties (Waldholz and Bogdanich, 1989~. There is ample evidence that ownership of technologies by physicians- particularly when the technologies are located in their offices engenders a capacity for technology utilization that is almost always exploited. Recent research indicates that office ownership of imaging technologies by nonra- diologist physicians results in four times the frequency of imaging utiliza- tion compared with physicians who refer their patients to radiologists for similar services (Hillman et al., 1990~. The Inspector General's office recently has published "safe harbor guidelines" that define how freestand- ing facilities may be organized and operated so as not to be legally culpable for conflict of interest. Congress is considering statutes that would further
PHYSICIANS ' A CQ UISITION AND USE OF TECHNOLOGY 147 restrict physicians' ownership of freestanding facilities to which they refer patients, but little consideration has been given to the same circumstance in physicians' offices. HOW MIGHT PHYSICIANS RESPOND TO THE "NEW HEALTH CARE" ENVIRONMENT? It should be evident from the foregoing that the current and expected future health care environments present physicians with a complex, often interrelated set of considerations that might affect their decisions concern- ing the acquisition of expensive medical devices. The focus of this paper has been on such devices because they have been highly visible technolo- gies with respect to the attention paid them in the scientific and lay media; they have also served as focuses for policy research and for issues related to reimbursement. It is important to note that the adoption of these technolo- gies probably differs in ways from the adoption of drugs, procedures, and even less expensive devices. Among these differences are the means by which physicians first learn of the new technology, what influences affect the acquisition decision, and characteristics of the key decision makers. How responsive physicians are to the considerations described in this paper depends, at least in part, on how well these factors coincide with the criteria physicians actually use in making technology acquisition decisions. In this regard, something can be learned from experiences with previous technologies (Greer, 1981; Hillman et al., 1986~. Physicians, in this author's opinion, first consider how the purchase of a new technology might benefit their patients and fulfill their patients' preferences. A related consideration is the specific attributes of the tech- nology and, to a lesser extent, what remains uncertain about it. The charac- teristics of a physician's practice (e.g., solo, small group, multispecialty group), the costs of the technology (to own and operate), and the kinds of difficulties the technology presents are also important. Note that all of these "primary considerations" are, for the most part, remote from concerns about the status of the general health care environment. Most physicians consider only secondarily (at least consciously) the financial rewards they might generate or how the regulatory, reimbursement, and competitive mi- lieu might affect their ability to operate the technology successfully. This lack of convergence between the intents of policymakers, payers, and regulatory agencies and physicians' considerations suggests that, at least in the near term, very little may change in how physicians approach acquisition of a new technology. Despite their anxiety over the harshness of the environment at the advent of MRI, physicians behaved as they have always behaved with respect to their acquisition decisions. This "business as usual" attitude is likely to continue until physicians are faced with a
148 BRUCE J. HILLMAN failure of this approach. Clearly, the stringency with which regulatory agencies and health care payers approach their cost-containment initiatives will be a major influence on physician responsiveness with regard to tech- nology acquisition and utilization. -- ~ ~ , providers in circumventing restrictions in the past, and the considerable ambiguity proffered by the current environment, suggest that physicians' technology-related practices may change to a lesser extent than many au- thorities expect. REFERENCES Nonetheless the inaenuitv shown hv Banta, H. D. 1980. Computed tomography: Cost containment misdirected. American Journal of Public Health 70:215-216. Bice, T. W., and Urban, N. 1982. Effects of Regulation on the Diffusion of Computed Tomography Scanners. Grant No. HS03750. Washington, D.C.: National Center for Health Services Research. Brown, L. D. 1983. Common sense meets implementation: Certificate of need regulation in the states. Journal of Health Politics, Policy and Law 8:480-494. Brust, J. C. M., Dickinson, P. C. T., and Healton, E. B. 1981. Failure of CT sharing in a large municipal hospital. New England Journal of Medicine 304:1388-1393. Cooper, L. S., Chalmers, T. C., McCallv. M Rerrie.r T ~nr1 Q~1~c ~ ~ __ ~ 1 :._ _ ~. 1 . _J7 __.7 ~ ~ _.~ ~ ~ by. I. 1988. The poor Halley o~ early evaluators of MR imaging. Journal of the American Medical Associa- tion 259:3277-3280. ECRI Technology Management Assessment. 1985. Meeting the challenge of free standing imaging centers. Options for hospitals and hospital based radiologists. Journal of Health Care Technology 1:257-278. Evens, R. G. 1980. The economics of computed tomography: Comparison with other health care costs. Radiology 136:509-510. Ginsburg, P. B., Newhouse, J. P., Mitchell, J., Palmer, A., Freeman, M., Hillman, B., et al. 1986. Planning a Demonstration of Per Case Reimbursement for Inpatient Services Under Medicare. Pub. No. R-378 HCFA. Santa Monica, Calif.: RAND Corporation. Greer, A. L. 1981. Medical technology: Assessment, adoption and utilization. Journal of Medical Systems 5: 129- 145. Hillman, B. J. 1986. Government health policy and the diffusion of new medical devices. Heal th Services Research 21 :681 -711. Hillman, B. J., Joseph, C. A., Mabry, M. R., Sunshine, J. H., Kennedy, S. D., and Noether, M. 1990. Frequency and costs of diagnostic imaging in office practice A comparison of self-referring and radiologist-referring physicians. New England Journal of Medicine 323: 1604-1608. Hillman, B. J., Neu, C. R., Winkler, J. D., Aroesty, J., Rettig, R. A., and Williams, A. P. 1987a. The diffusion of magnetic resonance imaging scanners in a changing U.S. health care environment. International Journal of Technology Assessment in Health Care 3:545- 553. Hillman, B. J., Neu, C. R., Winkler, J. D., Aroesty, J., Rettig, R. A., and Williams, A. P. 1987b. How experiences with x-ray computed tomography influenced providers' plans for magnetic resonance imaging scanners. International Journal of Technology Assess- ment in Health Care 3 :554-559. Hillman, B. J., Neu, C. R., Winkler, J. D., Aroesty, J., Rettig, R. A., and Williams, A. P. 1986.
PHYSICIANS ' ACQUISITION AND USE OF TECHNOLOGY 149 Diffusion of Magnetic Resonance Imaging into Clinical Practice. Pub. No. R3392-HHS. Santa Monica, Calif.: RAND Corporation. Hillman, B. J., Winkler, J. D., Phelps, C. E., Aroesty, J., and Williams, A. P. 1984. Adoption and diffusion of a new imaging technology: A magnetic resonance imaging perspective. American Journal of Radiology 143:913-917. Holohan, J., and Zuckerman, S. 1990. Understanding the recent growth in Medicare physician expenditures. Journal of the American Medical Association 263:1658-1661. Hyman, D. A., and Williamson, J. V. 1989. Setting the limits on physicians' entrepreneur- ship. New England Journal of Medicine 320:1275-1277. Kent, D. L., and Larson, E. B. 1988. Magnetic resonance imaging of the brain and spine. Annals of Internal Medicine 108:402-424. Mitchell, J. B. 1985. Physician DRGs. New England Journal of Medicine 313:670-675. Morreim, E. H. 1989. Conflicts of interest: Profits and problems in physician referrals. Journal of the American Medical Association 262:390-394. Relman, A. S. 1985. Dealing with conflicts of interest. New England Journal of Medicine 313 :749-751. Salkever, D. S., and Bice, T. W. 1976. The impact of certificate of need controls on hospital investments. Milbank Memorial Fund Quarterly 54: 185-214. Starr, P. 1982. The Social Transformation of American Medicine. New York: Basic Books, pp. 420-425. Steinberg, E. P. 1985. The impact of regulation and payment innovations on acquisition of new imaging technologies. Radiology Clinics of North America 23:381-389. Waldholz, M., and Bogdanich, W. 1989. Warm bodies: Doctor-owned labs earn lavish profits in a captive market. Wall Street Journal, March 1, pp. Al, A6.
