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.
ALTERNATIVE APPROACHES TO TECHNOLOGY MANAGEMENT The previous chapter contends that the prevailing system for re- imbursing health care providers is singularly responsible for persistent biases in the adoption and use of equipment-embodied medical technology. The reimbursement system discourages ap- plication of economic criteria in adoption and use decisions. Third-party payers virtually underwrite hospital capital invest- ments; institutions bear little or no risk for poor decisions. Moreover, the system provides no incentives for regionalization of services and does not discourage wasteful competition amona hospitals for patients or prestige. The reimbursement system also hinders the introduction of system-wide coordinative tech- nology. To some extent, the reimbursement system creates in- centives for physician specialization, and the system for reimbursing physicians encourages overutilization of equipment- embodied procedures. Solutions to the problems of technology adoption and use need not include reimbursement reform. Other policiesâfor example, direct regulation of the use of medical servicesâcould con- ceivably redress the imbalance. But as long as reimbursement policies provide incentives (or at least no disincentives) to overadopt hospital clinical and ancillary technology and provide inadequate incentives to adopt coordinative technology, all other policies to manage the introduction of new technology must coun- teract these effects of the reimbursement system. Even with re- form of the reimbursement system, regulatory or enabling policies designed specifically to improve adoption and use decisions might be desirable. Consequently, this chapter reviews some policies that this committee or others have identified as possible ave- nues for dealing with the technology problem. It analyzes the 46
47 expected direction, and, in some cases, magnitude of impact that these policies would have on the adoption and use of new equipment-embodied technology. Policies to influence decisions to adopt and use new tech- nology involve different costs and problems of practicality. Perhaps the most important problem is the level of detail at which a policy must be implemented. Policies requiring knowledge of many individual transactions are more costly and difficult to administer than those providing a structure of incentives which individual transactions follow. But direct regulation is often the only politically feasible alternative, because changes in incentives often result in fundamental changes in the organiza- tion and delivery of services. Regulations often impose a layer of control over a system that remains essentially intact. The participants in the system will adjust to the controls. If in- centives are radically restructured, the entire system can be seriously disrupted, a prospect that most affected parties would vigorously oppose. Arguments against regulation have been elo- quently stated elsewhere (see, for example, Noll90 and Schultze108). This committee remains committed to seeking out policy solutions that minimize the need for detailed knowledge of many transactions on the part of regulatory bodies. The committee also recognizes that the technology problem is part of a larger problem of health resources allocation. Many of the policy alternatives reviewed in this chapter have been widely suggested as more general cost-containment strategies. We have analyzed the impact of such alternatives only on the problem at handâthe adoption and use of equipment-embodied tech- nologyâand have not attempted to enumerate the many administra- tive, ethical, and political aspects of each strategy. The policy options fall into six categories: (i) funding of development; (ii) direct regulation of sellers of technology; (iii) direct regulation of adoption and use; (iv) payment or reimbursement policies; (v) health manpower policies; and (vi) information strategies. Particular policy options within each area are discussed below. FUNDING OF DEVELOPMENTAL EFFORTS As Galbraith34 has emphasized, the development of large-scale technological systems requires extraordinarily large capital in- vestments, usually over a period of up to l0 years, before a return can be expected. If, as in the case of coordinative technology such as computerized medical information systems, the ultimate market for the technology is highly uncertain, venture capital certainly will not be forthcoming. The development of
48 the technology may need a boost, either through the creation of a more certain market environment or through subsidization of development costs, or both. The relative advantages of develop- mental subsidies versus creation of a ready market must be judged in terms of the present state of development of the technology (the earlier in development, the more likely are subsidies to be relatively advantageous) and the degree to which private develop- ment funds could be reasonably expected to respond to market sig- nals. Little is known about where private firms place their research and development efforts. The need for more information on the extent to which public support of research and develop- ment augments or merely substitutes for private sector commit- ments has been recognized by others.1^2 In the view of this committee, funding of large-scale tech- nological development projects by the federal government is a reasonable approach, particularly for coordinative technologies. In the past, the National Center for Health Services Research funded large-scale demonstration projects for development of medical information systems and telemedicine. Currently, the National Center funds a special-emphasis Center for Technology at the University of Missouri, but decreases in federal funding and a shift in emphasis from development to evaluation have left a void in this area. Lindberg's paper on medical information systems (Appendix E) documents the catastrophic effects of on- again off-again federal commitments to development in that area. DIRECT REGULATION OF INTRODUCTION OF NEW EQUIPMENT-EMBODIED TECHNOLOGY It has been suggested that new medical technology should be con- strained from diffusing until adequate evaluation takes place.103 A technology would be allowed to diffuse only upon meeting de- fined evaluative criteria in experimental or demonstration settings. Such controls already exist for new "medical devices" in the form of premarket clearance requirements pursuant to the Medical Devices Amendments of l976 (P.L. 94-295). Third-party payers who refuse to pay for new procedures unless they have been approved are also engaged in this kind of control process. Nonpayment policies lack the force of law, however, and can merely inhibit, not stop, diffusion; a new procedure can always be performed at direct patient expense. The medical devices law requires manu- facturers to demonstrate the safety and effectiveness of sub- stantially new medical devices prior to commercial distribution.* *This law is presently in its earliest stages of implementation; consequently, procedures outlined do not represent actual operation.
49 P.L. 94-295 specifies that: Safety and effectiveness of a device are to be determined (l) with respect to the persons for whose use the device is represented or intended; (2) with respect to the conditions of use prescribed, recommended, or suggested in the device's labeling; and (3) weighing any probable benefit to health from the use of the device against any probable risk of illness or injury from such use. The amendments provide for classification of devices by the Food and Drug Administration (FDA) into three categories: Class I de- vices need only comply with the general regulatory controls of the law; Class II devices are additionally subject to performance standards promulgated by FDA; and Class III devices are subject to premarket approval requirements, including demonstration of effectiveness by well-controlled investigations, including clin- ical investigation "where appropriate."9 The regulatory scope of the medical devices law is wide. The federal government holds the authority to prevent introduction of virtually any new technology that depends on a new medical device. This power is somewhat limited with respect to equipment-embodied technology, however, by two factors: the legal definition of "medical device" is not entirely coincident with equipment- embodied technology; and the interpretation by FDA of "effective- ness" is likely to be narrow. â¢ Limited coverage. The law defines "medical device" as: An instrument, apparatus, implement, machine, contriv- ance, implant, in vitro reagent, or other similar or related article, including any component, part or ac- cessory which is (l) recognized in the official National Formulary, or the United States Pharmacopoeia, or any supple- ment to them, (2) intended for use in the diagnosis of disease or other conditions, or in the cure, mitigation, treat- ment, or prevention of disease in man or other animals, or (3) intended to affect the structure or any function of the body of man or other animals, and which does not achieve any of its principal intended purposes through chemical action within or on the body of man or other animals and which does not depend on being metabolized for the achievement of any of its princi- pal intended purposes.
50 This definition includes most clinical and ancillary equipment- embodied technology, but it does not cover new configurations of existing equipment (for example, intensive care units) and co- ordinative technology (for example, medical information systems). â¢ Limited definition of effectiveness. Although the law reads broadly, its major concern is with issues of safety. Con- sequently, the FDA is likely to interpret effectiveness in the narrowest sense possible consistent with patient safety. Where a device poses negligible safety hazards to patient and user, it is unlikely that the FDA will require evidence of improved pa- tient outcomes as a condition for premarket clearance. Moreover, the law includes no provisions for considerations of economic or societal criteria in evaluating new devices. The cost-effectiveness of a new device cannot be considered by the FDA under the terms of the law. Policies to control the introduction of new technology suffer from inherent limitations. The most important is the "all-or- nothing" nature of their effect. Although approval of a new technology can be limited to specified conditions of use (as in the medical devices law), it is difficult to enforce such limits. Indeed, malpractice litigation is the only available sanction against inappropriate use of technology short of direct utiliza- tion controls, and it is applicable only if significant injury has been sustained. A hypothetical example illustrates the weakness of preintro- duction regulations in affecting health care costs. A new, con- venient, and relatively inexpensive test could be proven cost-effective when used in diagnosis for a particular set of presenting symptoms. It would therefore be allowed to diffuse. If it subsequently became overused, its introduction could actu- ally have precipitated an increase in total diagnostic expendi- tures. Therefore, even if cost-effectiveness were a criterion for premarket approval, decreased health care costs could not be assured in the absence of utilization controls. Premarket testing cannot always detect rarely occurring but serious hazards. It is also difficult to anticipate adverse reactions that appear only after a long time interval. Such effects can be more consistently detected by long-term postmar- ket monitoring of technology usage. Premarket control of the introduction of new technology may have unintended side effects, the costs of which outweigh the intended benefits. Studies of the regulation of new drugs in the United States have identified and attempted to measure the mag- nitude of such effects, particularly since the enactment of amendments in l962 requiring drug manufacturers to demonstrate efficacy as well as safety. Although it is invalid to draw direct
5l inferences about the effects of controlling the introduction of new equipment-embodied technology from studies of drug regula- tion, the drug studies do provide insights into relevant areas of concern. Major findings of studies of the impact of premarket drug regulation are set forth below: â¢ Impact on new product development time and cost. Several studies of the drug industry document an increase in development time and cost following enactment in l962 of amendments adding efficacy requirements to the l938 Food, Drug, and Cosmetic Act. Grabowski and Thomas,42 using cross-national data, attribute half of the postamendment decline of productivity of the U.S. drug research dollar to the premarket approval process; the other half is attributed to a worldwide "depletion of research oppor- tunities. " â¢ Impact on the producing industry. The increased cost and time duration of new drug development following the l962 amend- ments led to the hypothesis that regulation induces a shift in the composition of the industry favoring large, multinational companies. These companies can afford and compensate for in- creased costs and risks of product development that might cripple smaller domestic firms. New drugs introduced in the postamend- ment era were developed by fewer and larger companies than were drugs developed prior to l962.43 â¢ Impact on innovation. Several critics of the premarket approval process for drugs contend that it forces manufacturers to divert technical and financial resources to the generation of large amounts of preintroductory evidence while foregoing re- search that might lead to significant new discoveries.10' In- creased development cost per new product decreases the number or scope of research and development efforts that each company can afford if research and development budgets remain constant. A regulation-induced reduction in the number of companies produc- ing new drugs43 may also reduce the diversity of research ef- forts undertaken in the drug industry. It is not necessarily true, however, that innovation suffers with a reduction of re- search funding. A study of drug companies by Caglarcan et al.15 finds that companies that concentrate research and development funds in relatively few areas of research tend to produce more new chemical entities in the 4 succeeding years than do companies with more diversified research efforts. This finding suggests that even if premarket approval does increase development costs per new product, innovation rates need not be affected. â¢ Impact on patient outcome. Wardell and Lasagna135 and others107 suggest that existing premarket regulation of drugs in the United States costs patients more than it benefits them.
52 Costs include those imposed by delayed introduction of beneficial drugs. Benefits result from fewer adverse patient reactions to new drugs and reduced use of ineffective drugs. Methodological difficulties in measuring a "drug lag" between the United States and other countries and in assessing its significance in terms of ultimate patient outcome hamper research on this issue. Federal legislators have learned from the drug regulation ex- perience. The medical devices amendments were written to avoid or reduce many of the problems outlined above. Medical devices regulation will differ substantially from drug regulation, offer- ing greater flexibility and efficiency. Only devices for which general regulations and performance standards are inadequate will undergo the premarket approval process. The amendments provide for use of experts from the scientific and industrial communities on panels to classify devices and on advisory committees to re- view proposed regulations of performance standards, proposed prod- uct development protocols, and applications for premarket approval. Manufacturers of Class III medical devices can submit a product development protocol to obtain FDA assurance of the adequacy of its product evaluation procedure before embarking upon it. Manufacturers will thus be able to avoid the risk of committing resources to a testing program that is inadequate for approval or to one that generates superfluous data. In addition, the medical devices amendments require the establishment of an office within HEW to provide technical assistance for small manu- facturers in complying with the law. These innovations in regu- latory procedure were intended to prevent much of the delay, duplication of effort, strain on FDA manpower, and inadequate communication, which have plagued drug regulation efforts. How- ever, it is too early to determine the effectiveness of these and other provisions in avoiding the unfavorable impacts de- scribed above. The limitations and side effects inherent in a premarket ap- proval process argue against significant expansion of this regulatory approach beyond the limited evaluative criteria of safety and effectiveness currently covered under the medical de- vices law. Because premarket regulation offers little effective control over the actual use of technology that has been approved, consideration of economic criteria such as cost-effectiveness or net social benefit is inappropriate. DIRECT REGULATION OF THE ADOPTION AND USE OF EQUIPMENT-EMBODIED TECHNOLOGY A commonly suggested mechanism for altering patterns of adoption and use is the exercise of direct control over these decisions
53 by a regulatory body. Two regulatory mechanisms currently in use exert this control directly: capital expenditures regula- tion and utilization review. Capital Expenditures Regulation Two different programs currently exist to regulate hospital capi- tal expenditure decisions. They are the state-level certificate of need (CON) laws, which exist in approximately 30 states, and the regulation of capital expenditures under Section ll22 of Public Law 92-603 (l972). The programs are essentially the same; they differ only in the definition of what constitutes a regulat- able capital expenditure and in the sanctions employed. Both programs call for approval of hospitals' proposals for large (usually at least $l00,000) capital expenditures by a state health planning agency, with local health systems agencies acting as the first line of review. The Section ll22 program ties ap- proval to reimbursement by Medicare for the associated capital costs; CON laws are frequently tied to facility licensure or cer- tification. The National Health Planning and Resources Develop- ment Act of l974 (P.L. 93-64l) requires universal adoption of CON laws by l980. At present, some Blue Cross plans refuse to reimburse health care providers for services that have been dis- approved by the CON agency. Capital expenditures review is essentially a negative instru- ment; it can be used to slow the diffusion of new equipment- embodied technology, but it cannot easily be used to speed the adoption of underdiffused technology. Except where health plan- ning agencies have conditioned approval on the fulfillment of stipulations unrelated to the application* (as in Massachu- setts100) , the regulatory agency is powerless to encourage the development of coordinative technology. The history of CON and capital expenditures review programs is discouraging. Not only have few projects been disallowed,70 but CON appears to have no effect on the overall level of spend- ing by hospitals.106 However, one cannot judge the potential value of this regulatory strategy from experience to date. Ex- isting programs suffer from serious structural weaknesses and implementation problems that may be surmountable. Among the more obvious problems are the loopholes in existing laws. At present, capital expenditures review does not cover relatively small investments by hospitals, nor does it cover *The legality of conditioned approvals of projects has been questioned.100
54 noninstitutional providers (for example, physician groups or independent clinical laboratories) in most states. It has been suggested that CON be extended to noninstitutional settings of care60 and to a broader array of equipment purchases (under $l00,000). These changes would purportedly close loopholes in the laws and provide more influence over technology adoption decisions. A problem of equal importance is the lack of budget disci- pline within the program structure. The agencies are generally free to approve as many projects as they consider appropriate without considering trade-offs among alternative uses of capital funds. Indeed, in most areas, the boundaries of the health ser- vice area represented by the CON agency do not coincide with the boundaries of collective payment (which in the case of Medicare is the entire country). Thus, the local agency can spread the costs of a new service across the state or the nation while obtaining the benefits of improved access or quality of care and often increases in employment as well. One solution to this problem is the adoption of an annual area- wide limitation on capital expenditures subject to CON as has been proposed under the present Administration's cost contain- ment bill121 and as under consideration in the State of Maryland.95 The Health Systems Agency (HSA) would be responsible for allocat- ing these funds among competing projects. This approach would materially increase the power and responsibility of the USA's and would require them to make choices they have up to now appeared unwilling to make. 0 However, it is possible that such a program would induce hospitals to merely shift capital expenditures to the purchase of less expensive capital equipment not subject to the limit. It is also important that the HSA or local hospitals be able to accumulate capital spending allocations over multiple years to enable the implementation of especially large projects that may be needed. Apart from the structural problems described above, the agen- cies suffer from a lack of timely information to assist in deci- sion making, particularly in regard to new equipment-embodied technology.58 Assuming that the structural and implementation problems are overcome, capital expenditures review nevertheless presents serious generic problems. First, this regulatory approach re- quires detailed knowledge of individual equipment adoption de- cisions by review agencies, and the number of decisions probably increases exponentially as the dollar threshold for CON cover- age is decreased. That is, there are many more different pieces of equipment offered in the $50,000-$l00,000 price range than there are in the over-$l00,000 range. Thus, as loopholes are closed, implementation problems increase dramatically.
55 Second, agency decisions are likely to be politicized through coalitions or compromises. The net effect may be expenditure decisions that are no improvement over those made by hospitals. Third, certificate of need unquestionably adds to the cost of adopting new technology, both good and bad, and thus further bi- ases capital expenditure decisions against the introduction of potentially cost-reducing coordinative technology such as medical information systems. The question of whether direct regulation of capital invest- ments by providers is a good approach to control the tendency to overadopt clinical and ancillary hospital technology ultimately rests on one's philosophy toward regulation. The committee rec- ognizes the critical need to make such adoption decisions more consistent with the interests of society, but if similar results can be achieved through policies that alter incentives instead of counteracting them, then we believe that these other avenues are preferable to capital expenditures regulation. Utilization Review The review of the utilization of equipment-embodied technology is a mechanism that could conceivably reduce the tendency to overuse clinical and ancillary technologies. The term "utiliza- tion review" refers to several methods for controlling the use of services. These include prior authorization of service de- livery, retrospective evaluation of services rendered, and re- view of claims for payment. Various programs have been conducted by hospitals, third- party payers, foundations for medical care, and professional standards review organizations (PSRO's). For the most part, utilization review has been applied to hospital admission and length of stay and to claims for services rendered by physicians. Except where hospital admission has been expressly for the per- formance of a particular procedure, review of specific procedures has generally not been attempted. This is particularly true of diagnostic procedures in hospitals. The professional standards review organization program is currently encouraging its PSRO's to begin programs to review the utilization of ancillary ser- vices. 125 The effectiveness of utilization review programs as they have been implemented has not been unequivocally demonstrated.59 Some inherent weaknesses in the approach are: the need for de- tailed criteria to differentiate justified from unjustified util- ization, the costs of implementation, and the potential rigidity of such a system over time. More research is needed on the
56 effectiveness of utilization review in controlling the use of technology-intensive services, particularly ancillary hospital services such as laboratory and diagnostic radiology, relative to more indirect methods of influencing the utilization of these services. Both direct regulation of capital expenditures and utiliza- tion review require that regulators have detailed knowledge of individual transactions in the health care system. The commit- tee remains cautious about the long-run effectiveness of this type of approach. PAYMENT POLICIES Virtually any change in reimbursement strategies will alter in- centives to adopt and use new equipment-embodied technology. Desired strategies are those that reward providers for cost- effective adoption and use decisions and penalize them for de- cisions that clearly are not cost-effective from society's standpoint. As a corollary to this principle, providers should bear risk for their adoption decisions, either as a financial loss for a poor decision or as a loss of opportunity to invest in more effective technology. Several alternatives for changing reimbursement methods in this direction have been suggested. Coverage Limitations Perhaps the most obvious type of reimbursement reform is the re- fusal to cover hospital and physician expenses incurred for those procedures that have not been proven effective. The delay im- posed by the Medicare program in reimbursing providers for CT scanning of the body is an example of such a strategy, as is the decision by the National Association of Blue Shield Plans not to pay for 26 surgical and diagnostic procedures that have been found to be largely without value.8^ The primary problem with this approach is that it is a gross discriminator between "good" and "bad" technology. The decision of whether or not to pay for a procedure can be used successfully only for the few procedures found to be generally without value, but it is relatively easy to show that a technology has significant value for some patients. Consequently, a comprehensive program of noncoverage would have to be augmented by detailed utilization review. Third-party payers could be more circumspect in paying for new clinical and ancillary procedures. By requiring more proof of efficacy prior to payment, third-party payers could encourage the development of information on effectiveness. However, once a procedure is approved for payment, even for a limited patient
57 population, some control over the utilization of such services would be required if the tendency toward overadoption and over- use of clinical and ancillary technology is to be checked. The committee is concerned about the implications of adding an additional layer of control over a process of technological change already heavily burdened with regulation. It would re- quire most new equipment-embodied clinical and ancillary pro- cedures not only to meet the safety and efficacy requirements of the medical devices law, but also to meet more or less stringent effectiveness or cost-effectiveness standards imposed by dif- ferent payers under a coverage limit policy. The delay, uncer- tainty, and administrative costs entailed by the addition of a second regulatory structure in the prediffusion stage of techno- logical change must be considered in evaluating this alternative. Reimbursement of Hospitals The reimbursement of hospitals on a prospective basis has been considered by many to be an important and needed change. Pros- pective reimbursement (PR) merely implies the determination in advance of the payment period of a rate of payment for services rendered by hospitals. There are many variations on this general approach. Third-party payers have established rates on the basis of negotiated hospital budgets and on cost-based formulas. A few have established flat rates of payment unrelated to services ren- dered. In some cases, hospitals bear the risk of deficits and reap the benefits of surpluses over the payment period. In oth- ers, savings and deficits are shared by the hospital and the third-party payers. The essential character of PR is lost, how- ever, if levels of reimbursement are routinely adjusted retro- actively to eliminate all accumulated surplus or deficit of the hospital. In January l976, 22 of the nation's 74 Blue Cross plans en- rolling from 24 to 80 percent of their area populations were negotiating or establishing prospective rates or charges for their member hospitals.12 (Nine of these plans in two states were operating in cooperation with state rate-setting agencies.) Nine state governments were administering rate-setting programs affecting from 8 to 90 percent of state hospital revenues. These state programs impose binding rates of payment under various pay- ers, usually including Medicare, Medicaid, and Blue Cross.12 The proposed hospital revenue limitations under the Administra- tion's hospital cost-containment legislation121 is a PR system. Each hospital, with certain exceptions, will be limited to a predetermined rate of increase of total revenues from one period to another.
58 A major problem in the implementation of PR systems is the need to treat different hospitals differently. Because hospitals vary widely in their mix of patients, and because the prospective rate sets up incentives to "dump" complex cases and to change the length of stay, it is imperative that PR systems adjust rates for, hospital complexity or case mix. This greatly complicates the rate-setting process, and it has not been shown that an adequate adjustment method exists. A critical attribute of any prospec- tive reimbursement system is the notion of regulatory lag, the time interval between successive rate decisions. The rates ap- plied during this interval are generally based upon previously recorded accounting costs, although some systems use a flat rate determined not by previous costs but by administrative fiat. Where rates are cost-based, the interval between rate decisions is critical to the method's success in inducing cost-saving in- novations. If costs fall during the period, hospital surplus increases, whereas, if costs rise, as is likely in inflationary times, deficits may occur. Many have argued that the regulatory lag offers a major and perhaps the only incentive for efficiency in regulated organizationsâparticularly to apply new cost-saving technology.7,11" Most PR systems operate on a l-year interval. The committee debated the merits of lengthening the interval to increase incentives to introduce new cost-saving technology. But in periods of rapid inflation, the regulatory lag may threaten the fiscal integrity of institutions. Thus, PR necessitates a built-in adjustment mechanism for price inflation. Apart from this requirement, however, increasing the regulatory lag would en- courage the introduction of cost-saving technology, but it would also discourage the introduction of technology that is both cost- increasing and quality-enhancing. Unlike other regulated indus- tries, hospitals do not produce a uniform product. An implicit cost is incurred in increasing the time between rate decisions. That cost is the loss of ability to adopt new quality-improving technologies. In the opinion of this committee, the potential for increasing the regulatory lag as an incentive to efficiency has not been fully explored by rate-setting agencies. More study of this ap- proach to prospective rate setting is needed. Theoretically, PR should have a salutary effect on the adop- tion and use of new clinical and ancillary technology, since it requires hospitals to weigh alternative uses of funds. It is not clear that this does, in fact, happen. Bauer12 has observed that the primary weakness of prospectively set hospital rates is that they do not control the volume of services offered; these are controlled by individual physicians. However, if over time rate limitations cause decreasing adoption of new technology by hos- pitals, limited capacity can be expected to control utilization.
59 Although there is little definitive empirical evidence on the impact of PR on adoption and use of technology, studies of six prospective reimbursement systems sponsored by the Social Security Administration provide some preliminary and indirect information. Prospective reimbursement as it was applied in the six situations did not unequivocally reduce hospital cost infla- tion.1,23,55 Moreover, one study showed that PR had not signif- icantly curtailed the use of ancillary services;55 and in another study, the PR system appeared to have a negative impact on the number of clinical laboratory tests performed per case and a positive impact on the volume of radiology procedures per case.2' It appears that hospital administrators attempt to control costs of the services they control, while the use of physician- controlled services does not change. However, these programs may not have been in effect long enough when studied to induce desirable change in adoption behavior that would be reflected in later utilization statistics. Reimbursement of Physicians Changes in methods of reimbursing physicians have also been sug- gested. Hospital-based physicians (anesthesiologists, radiolo- gists, etc.) who control the majority of clinical and ancillary equipment-embodied technology are now paid predominantly by methods that reward for high volume. It has been suggested that requiring all hospital-based physicians to enter into salary agreements as a condition for third-party payment is a solution to the perverse incentives operating in hospitals.142 But it is not clear that this would significantly affect adoption and use of new technology. Certainly, for example, the impetus to the use of anesthesia is not the anesthesiologist, but the sur- geon. And ordering physicians largely determine the use of clinical laboratory and radiological procedures in the hospital. A salary system might reduce the productivity of hospital-based physicians, and thus increase costs without decreasing utiliza- tion. It may appear plausible that if, for example, radiologists' incomes did not vary with the introduction of new equipment, such as the CT scanner, there would be less pressure on hospitals to adopt this technology. But experience has demonstrated that, at least for the head scanner, neurologists and neurosurgeons have favored adoption just as actively as have radiologists. The argument for putting hospital-based physicians on salary may be justified on the grounds of equity and income distribution; it is not supportable on 'the grounds of improving incentives to adopt and use new clinical and ancillary equipment-embodied technology.
60 The fee-for-service system, which reqards physicians for the use of procedures, might be altered. Changes in the system could range from the institution of a salary system for all physicians to changes in the way fees are constructed. At present, fees for new procedures are based on historical precedent, not on prospective analysis of the resource costs to the physician for performing the procedure. The "usual, customary, and reasonable" approach to the establishment of physician fee schedules has been shown to be inflationary, and, although limits have been placed on fee increases in recent years, new procedures are generally not affected. A new procedure can be introduced at a high rate of payment by innovative physicians. This reimbursement rate can always be lowered, but the method for determining fees makes it difficult to raise a fee substantially once it has been estab- lished. Consequently, fees for new procedures are likely to be set at high levels. Third-party payers could establish fee sched- ules where payment for a procedure is linked to its effectiveness relative to other procedures or to its status as an experimental procedure. These incentive-based fee schedules would offer the physician higher economic returns per time period for more "de- sirable" procedures than for less desirable procedures. A major problem with this approach is the need to discriminate between appropriate and inappropriate use of a procedure. The most valuable procedure can be misapplied. By establishing high fees for desirable procedures, the policy is likely to induce too much of a good thing. Consequently, the success of incentive- based fee schedules is inextricably linked to the detailed con- trol of utilization of procedures. Capitation Payment Capitation payment methods reimburse for an agreed upon set of patient services with fixed periodic payments regardless of the value of services actually rendered. In principle, capitation rewards providers for efficient use of resources in producing the services covered by the capitation rate and penalizes inef- ficient providers. The most common form of capitation payment is the health maintenance organization (HMO), which includes all inpatient and outpatient services in the services covered by the capitation rate. Two competing forms of HMO are the prepaid group practice, which directly employs physicians and other health care personnel, and the independent practice association (IPA), a federation of independent practitioners who agree to participate in the capita- tion plan but who maintain their individual practices and are often individually reimbursed on a fee-for-service basis. The degree to which the IPA approaches the prepaid group practice HMO in organization depends on the amount of financial risk that
6l participating physicians take. In some plans, the contract pay- ers like Blue Cross and Medicare require that the third-party payers absorb the additional expenses if costs exceed the pre- mium base. Here, the participating physician takes very little fiscal responsibility for his actions, and the IPA more closely approximates the fee-for-service system. There is substantial evidence that prepaid group practices and IPA's do result in some economies of operation. Hospitalization rates are lower and lengths of stay in hospitals shorter than in the fee-for-service sector.52 The prepaid group practice appears to perform more efficiently in terms of hospitalization rates than do IPA's.37 There is also tentative empirical evidence that rates of surgery are lower at some HMO's than in the rest of the health care system.71 However, evidence on the comparative rates of diffusion of new equipment-embodied technologies does not exist except by way of anecdotes. A major confounding problem is that relatively few HMO's own their own hospitals. Most contract for hospital services with institutions serving a wider population. Thus, equipment adoption decisions are generally made by hospitals with reference to the larger population, and valid inferences about HMO adoption behavior are not possible. Much of the impetus for the development and demonstration of preventive technologies such as automated multiphasic health testing (Appendix C) and mammography screening (Appendix B) has come from the larger and older HMO's, whose resources have al- lowed these efforts. Much of the earliest and most successful developmental work on medical information systems was performed at the Northern California Kaiser Health Plan (Appendix E). At present HMO's provide health care for about 3 percent of the U.S. population. Numerous hypotheses have been promulgated to explain such negligible penetration of the health care de- livery market, including barriers arising from current health insurance arrangements and from the very law that was designed to encourage HMO's (the Health Maintenance Organization Act of l973, P.L. 93-222).61 Although recent amendments to P.L. 93-222 have reduced some of these barriers, HMO's still face significant obstacles. Unfortunately, HMO's are subject to some counterproductive incentives that might produce behavior that compromises the qual- ity of care, underserves subscribers, or selects subscribers to eliminate high-risk members. Thus, surveillance and reporting of HMO performance is mandated, possibly leading to regulation of HMO behavior.50 This may compromise the administrative sim- plicity of the concept. Nevertheless, HMO's appear to hold promise for instilling appropriate incentives in health care providers both to adopt cost-effective technology and to resist
62 the overadoption and overuse of clinical and ancillary hospital technology. HEALTH MANPOWER POLICIES Because of the hypothesized effect of manpower (particularly physician) specialization on the adoption and use of new clinical and ancillary equipment-embodied technology, it is often sug- gested that the absolute number and specialty distribution of physicians be controlled. Training of new physicians in the clinical specialties and subspecialties could be decreased, while training in areas such as computer applications or medi- cal information systems could be increased. Three policy instru- ments are available to alter the specialty distribution: (i) regulation of the number of residency positions offered by teach- ing hospitals; (ii) third-party reimbursements of residents' sti- pends; and (iii) training grants to teaching institutions. The regulation of the number of residency positions offered in any specialty or subspecialty is in the hands of the Liaison Commit- tee on Graduate Medical Education, which represents five medical professional organizations. Self regulation by the profession appears to be a viable alternative. In l972, the American College of Surgeons conducted a study of surgical services in the United States and concluded that the number of approved surgical resi- dencies should be sharply curtailed.4 Since then the number of approved residencies has indeed decreased. Whether limitation of physician specialization will signifi- cantly affect adoption of new technology is conjectural at this time. It is not clear that by limiting the number of physicians in the technology intensive specialties there will be less use of equipment-embodied clinical and ancillary technology. The rapid increase in the use of radiology procedures in the past l0 years has not been met with a proportional increase in the num- ber of radiologists. Between l964 and l970, the number of radi- ologists increased by approximately l0 percent; the number of diagnostic x-rays performed in the United States increased by 20 percent. Increased efficiency in the use of radiologists' time, substitution of nonphysician manpower for radiologists' time, and increased performance of simple radiologic procedures by non- specialists may account for the difference. There is a substan- tial but unknown quantity of "trickling down" of procedures from specialist to nonspecialist performance. As specialists become busier with more sophisticated, newer procedures, the use of existing, simple procedures is taken up by nonspecialists. The performance of simple surgeries by nonsurgeons and the performance of simple laboratory tests outside of the laboratory might be- come commonplace.
63 INFORMATION DISSEMINATION STRATEGIES In the previous chapter it is argued that perceptions of "qual- ity" influence the adoption and use of equipment-embodied tech- nology. If patients, physicians, and hospital administrators all view technological sophistication as good in and of itself and ignore cost in their decisions, then there is likely to be a bias toward the adoption and use of new clinical and ancil- lary technology. Several information strategies are possible. These can be directed either at consumers (or their representatives) or at providers of health care services. Consumers (or consumer rep- resentatives) include individual patients, third-party payers, employees or unions who contribute to group health insurance plans, or public bodies such as health planning or rate-setting agencies. Providers include practitioners, institutions, and profes- sional associations. Provider education may be directed at the physician as gatekeeper of the use of new technology, at the hospital as primary adopter of new clinical and ancillary tech- nology, or at organizations of professionals, including such quasi-public agencies as professional standards review organiza- tions (PSRO's). Suggestions for information strategies directed at physicians are based on the premise that medical education inadequately pre- pares the physician to consider cost in decisions to use ser- vices,96 particularly diagnostic procedures. Policies to include concepts of -statistical decision theory and cost-effectiveness in medical education have been suggested as solutions.yt> Recent research documents that educating physicians to consider costs when making clinical decisions can significantly affect the num- ber and cost of diagnostic tests used.24 Although it seems self- evident that medical students should be provided with a basis for rational clinical decision making, the effect of such a strategy on utilization would only appear after many years of such training, if at all. Moreover, adding required subjects to the medical education curriculum involves sacrifices in other areas of learning. The education of practicing physicians through PSRO's and utilization review programs has also been attempted. Because these programs are of recent origin, their effectiveness in chang- ing physicians' utilization behavior is unknown. Hospitals could also use information on the costs and effec- tiveness of new equipment-embodied technology. Hospital adminis- trators claim to have severe problems in evaluating the technical merits and cost implications of equipment and instrumentation.123 The lack of standardization and the complexity of equipment used in clinical and ancillary services renders hospitals relatively
64 ignorant of the potential hidden costs and technical problems of equipment they purchase. Efforts have been made to provide hospitals with methods for evaluating equipment prior to pur- chase,3 but more fundamental information about the effectiveness of new technology is badly needed by hospitals. Education of individual patients can effectively improve pat- terns of use of new technology, particularly of screening tech- nology. However, it is unrealistic to assume that patients can or should be gatekeepers for their own use of diagnostic and therapeutic services. Patients can be encouraged to seek "sec- ond opinions" prior to acceding to their physicians' advice, but the widespread use of second-opinion strategies has not yet been attempted, and their cost-effectiveness in curtailing utiliza- tion is unknown. The information needs of consumers and providers depend, of course, upon other policies chosen to manage the introduction of new technology. If, for example, adoption decisions were taken away from hospitals and put in the hands of regulators, then in- formation strategies would be best directed at the regulatory agencies. The next chapter of this report discusses the process for generating and disseminating evaluative information, and ana- lyzes the extent to which changes in that system are warranted. A CAVEAT ON THE EXPANSION OF REGULATORY PROGRAMS In attempting to array and analyze public policies that can be used to influence the process of development and diffusion of equipment-embodied technology, the committee was impressed by the plethora of public policies, particularly regulatory programs, already in force governing every stage in the process. A com- plex regulatory structure has evolved that at least nominally con- trols many aspects of development and diffusion. In fact, with few exceptions noted in previous sections, federal legislation has already authorized direct control of many decisions regarding the diffusion of equipment-embodied technology through three major lawsâthe Medical Devices Amendments of l975 (P.L. 94-295) , the National Health Planning and Resources Development Act (P.L. 93-64l), and the professional standards review organization (PSRO) provisions of the Social Security Amendments of l972 (P.L. 92-603). These laws are administered at various levels of government; the medical devices law will be administered by the federal govern- ment, while the planning act and the PSRO program are administered at state and areawide levels by public and quasi-public agencies and organizations. Leaving aside the question of each program's effectiveness in accomplishing its legislative or social objectives (a question
65 discussed above), it is important to consider the cumulative impact of these essentially independent regulatory programs on the process of technical change. The three mandates constitute a regulatory maze through which new technologies must wend their way. An equipment-embodied technology must first be approved as safe and effective under the Medical Devices Amendments. Those whose initial cost is $l00,000 or more and intended for hospitals must usually be granted a certificate of need. Fi- nally, the use of equipment-embodied technology may ultimately be controlled by utilization review criteria developed and ad- ministered by PSRO's. These decisions are made at several dif- ferent points in the process of technological change and, for the most part, independent of one another. The multiple layers of regulation certainly increase uncertainty about the market- ability of new technology and may require redundant administra- tive costs. The problem of regulatory burden goes beyond these three pro- grams. The development and diffusion of equipment-embodied medical technology occur in a complex regulatory environment involving federal, state, and local governments as well as pri- vate organizations. Table l shows the extent of direct regula- tory authority that impacts on various participants in the process of technical change. (The table does not include the indirect impact of funding policies such as manpower training, biomedical research, and health care financing programs.) The significant potential for redundancy and inconsistency in regula- tions relating to a new technology as it works its way through development and diffusion is clearly demonstrated by the table. While few would dispute the value of some regulation, two re- sults seem evident. First, on balance, regulation has or is likely to slow the development and diffusion process. Unfor- tunately, the extent of such a slowdown and the resulting bene- fits and costs are difficult to evaluate, due partly to the newness of the regulatory machinery and to the inadequacy of our knowledge about benefits and costs of new technology. Second, the present body of regulations is in many respects duplicative and conflicting. The same resultsâhowever valuableâcould be achieved in a streamlined system at lower cost. Among the potentials for redundant regulations is the plethora of state and federal agencies involved in licensure, certifica- tion, and accreditation of health care facilities illustrated in Table l. A study conducted in l975 by the Task Force on Hospital Regulation of the Hospital Association of New York State57 identi- fied a total of l64 regulatory bodies involved with hospitals in New York. Of these, 40 are federal and 96 state agencies. Four federal and 23 state agencies are involved in licensure; l0 fed- eral and 26 state agencies are involved in accreditation of the
) -1 â¢a c 4J IH 3 0) ra E 0) 2 â¢H â¢D C On s- . 8 Â§Â£ o O O M 01 4J H Â§ i â¢H 0 '" ^ t^ 2 msÂ§ 0) W rH 0) & C H M id 0> C 0) -H O O 4J â¢H O 0 g i a nl t n 4J O c Â° 5 >d u & Â§ â¢H 'O â¢H -H M t n p -n â¢rl 0) â¢H J) jj a s w C - â¢O O â¢ 0) IM 5-3 â¢0 0> i 0) TJ H 0) V 4-i .rl M it rH C â¢H o rH 'O O rl â¢ U M rd O | P C 4J H 0) Â§ pi e1) eg u a.â < 0) O 0) -H X 0) c & i I T H i ) i i i id 4J 0) M C s o ^ &> C . EL â¢H â¢9 0) M e 1 0) â¢H rH C p 0) ^1 0 3 ^H s* H rH O id -H C c ai 41 H rd 0 4J O 0) rH IH O 0) -rl id â¢H â¢H ) ~ i0 O 3 S -HE H >, 4J 0) 0) C 01 rH rH & id 4J rH O rH 4J C ffi 0 id ro -H 0) id C O e >, rl H IH â¢H 0) g â¢Â§ c 4J JS 4J 0) rH â¢H 0 ) a, oi a i2 1o & â¢r4 4J 4J SUH - a 0) H I4H 0) rH p id Â« S4J A S ^ C > & io to w 4J EM CO ^_- cu Ul rH gs i i i i | i id i 1 l i u â¢O Â§ â¢H 4J O Â§ â¢H JJ o" jj O H 0 ^ O c |< CN rHS & | 4J 0) 0) a O C X nI 3 o â¢a -H 0 Si 4J â¢H UH id -H i44 id H 04 id 4J 0 4J rl O 0) 4) UH c 0 rt 5 'â¢a o! rH IH s o 0 â¢ â¢3 jj J C rH a o U J ra 0 - r- rH i 0) tJ id rH C e e 0) 0) S id O E -H rH 1 * \ rl 0 â¢H 0 d) 0) 4J o o â¢rl 4J 4J -H id 0) crl M T3 o â¢ H -P TJ O 0 4J â 0 > S'2 â¢'3 8 c as viro 8 s rH < 0) ai id co â¢H O C h o n) id 4J X O â¢n -H vD JQ TJ Ov â¢H *> P rH rH â¢H ft O rO J3 â¢$ 3 cn H U D Id rH 0) S â rH e ft o) cu id u 82 c 4) 2 0. U H M X 4J 0) i i I T i T I 4J ) i | i i i i c 3 rQ ^^ ^^ o 0) 0) 0) 0) ro O U3 | 0) â¢H 0) c c c o Cn M | 0 â¢H 0) 0) 4J CN 2 â¢r4 4J 4J c o id -^ Q "* id e o â¢j â¢P â¢H in 0) â¢H j & 0) CN rO VD id â¢H i/) â¢0 I H i o 8^ Oi 4J S c CL ro â¢H fO IH o rH 01 cn D*"? N â¢H S â¢ X â¢H a â¢H rH o rH 4J 4J â¢ JJ Â«J 0) Â§" â¢H id id id â¢rl .1 rH lJ iH â¢ H rl iH â¢O â¢ id â¢ ID 4J o 1 O 0) S Si i 0) & O Â§ Â£ S 3 4) <H Si | i Â« in ro 4J O rH z l ) O M a C i*H -H 0) S ro O tT> E 3 0) 0 T3 pj 0 0) 4J H -H rH -rH 0) rH 5 '8 4J â¢H H O 0) 0) O 0) M _) â¢H X 0) C 4J ti 0 X C H 3 $ 4J O S o H 00 1 C 0 nj i id C M 0) 0. -H ft H g i rl a. ft 66
67 hospitals. The study clearly documents the duplication of inspections, reports, and administrative burdens of the many federal and state regulatory agencies in New York. Similarly, a report on the health care regulatory control system in Massachusetts67 identifies more than 40 discrete ave- nues of health regulation currently in force, an estimate be- lieved to be conservative. The report also documents that Massachusetts hospitals undergo at least 5 inspections yearly; nursing homes undergo at least l4 such inspections. Although hospitals in New York and Massachusetts may be subject to more regulation than facilities in other states, the redundancies illustrated by these examples are indicative of similar patterns in most other states. In sum, the health care industry is highly regulated by all levels of government and many private associations with varied objectives. In assessing the relative attractiveness of alter- native approaches to improving the adoption and use of new equipment-embodied technology by the health care system, policy- makers should be extremely cautious about the wisdom of develop- ing new layers of regulatory authority over those already existing. Moreover, much could be gained by reassessing the linkages among existing regulatory programs in an effort to un- cover conflicts, duplication, and inconsistencies ripe for re- form. SUMMARY Two conclusions are clear: First, it is difficult to predict the magnitude of the impact of most policies on the adoption and use of new equipment-embodied technology due to a dearth of empirical evidence; and second, no single policy applied in iso- lation appears to be a viable solution to the problem of adoption and use of new equipment-embodied technology. Nevertheless, certain policies appear more promising than others. Reform of the reimbursement system to promote appropri- ate incentives relative to the adoption and use of equipment- embodied technology is preferable to direct regulation of such decisions. In particular, limitations on third-party payer cov- erage of unproven clinical and ancillary technology, prospective reimbursement of hospitals, and especially capitation payment merit further exploration. Coordinative and preventive technology has lagged behind clin- ical and ancillary technology in development and diffusion, be- cause they are enormously costly to develop and difficult to integrate into the current system of care. Therefore, subsidiza- tion of the development of this class of equipment-embodied tech- nology is justified.