The Roles of Academic Health Centers in the 21st Century: A Workshop Summary (2002)

industry? Ralph Snyderman noted that AHCs have relied on philanthropy to fund the institution’s research contribution of approximately 15 percent. It is also hoped that technology transfer to private industry will be a source of revenue. Gerald Fischbach responded that although there is likely to be greater support of biomedical research from private sources, nothing will match the government’s support. Over time, Ralph Snyderman believes that there will be a limited number of research-intensive institutions that will thrive and grow but that the resources may not be distributed evenly across all AHCs. His concern is with institutions that do not have a track record in research but are trying to obtain sufficient funding.

Paul Ramsey noted that one of the issues brought about by the changing science, especially genomics, is the need for people with training in computational skills. However, often the people with these skills do not necessarily have an interest in health care; they are more oriented toward mathematics, for example. He asked how people with a core set of skills in other disciplines could be attracted to applying those skills to health care. In response, Gerald Fischbach responded that there are indeed serious issues on how integrate genome sciences, genetics and computational skills. He saw two possible approaches. One is to create integrated institutes that bring together different disciplines in one center. Another approach is to rely on the university and the department of computer science and applied mathematics to make joint appointments. Samuel Broder believes that computational biology is an area where the private sector has excelled beyond the academic sector. In fact, the scientists trained in the private sector (e.g., Celera Genomics) are now moving to the public sector (e.g. NIH). It is a training model that moves in the opposite direction of what is typically assumed.

John Porter noted that AAMC recently published a guidance piece on financial conflicts of interest in research, which recommends resolution of possible conflicts of interest at the organizational level. He asked if resolution at the local level is a good approach to resolving issues of disclosure or whether there should be national standards or oversight on disclosure. In response, Dr. Broder commented that the AAMC proposal is logical, but the key element is disclosure. There should be a structure throughout the university or medical school that requires peer and community input, similar to what is done on institutional review boards. Conflicts of interest are not necessarily negative, but need to be disclosed. In his view, it is not clear that a national agency is needed to oversee such a process, although national standards could be helpful.

David Blumenthal, M.D., Executive Director, Commonwealth Task Force on Academic Health Centers

David Blumenthal briefly described the Commonwealth Task Force on Academic Health Centers. This task force is completing its seventh and final year. It has published 6 task force reports with recommendations, 11 task force papers, and 44 publications in peer-reviewed journals. They have conducted surveys of faculty, residents, and department chairs and research administrators. They have also conducted analyses using secondary data from a number of sources.

The task force spent a fair amount of time looking at how market forces and financial stress affect the missions of AHCs. The "missions" were defined to include research, indigent care, education, and delivery of specialized and high technology services. The task force generally found that competition and financial stress do affect the AHC missions, both positively

and negatively. Research is negatively affected by market and other pressures. Indigent care has become more concentrated in AHCs, which may be interpreted as either good or bad for patients, but is not good generally for the AHC. There are anecdotal accounts of adverse impacts on teaching, but none were found empirically from the survey the task force conducted of residents. Specialized services also have become more concentrated in AHCs over the past five years. This may benefit patients to the extent there is a relationship between volume and outcomes of care.

The task force also found that mission-related activities do account for some of the increased costs at AHCs. These higher costs are not only related to teaching, but also to clinical research and the maintenance of standby services. In that sense, “indirect medical education” is misnamed because the higher costs found at AHCs relate to more than just medical education.

Dr. Blumenthal emphasized his willingness to share the task force data with the IOM Committee on the Roles of Academic Health Centers. Although answers to many questions remain elusive, he can help the committee identify important questions and identifyied the following ten issues that are likely to be confronted in studying the roles of academic health centers:

One, what is an AHC? An AHC can be defined as a single entity or, alternatively, as anywhere the missions are conducted, suggesting a different definition for each role. The Commonwealth Task Force on Academic Health Centers defined an AHC as the nation's 121 four-year medical schools and their owned or affiliated clinical and educational entities. One of the common things that all AHCs do is teach medical students, at a minimum. Regardless of the definition used, diversity is an abiding characteristic of AHCs. As the environment changes, this diversity may be a strength as they determine how to optimally accomplish the missions that society has entrusted to them.

Two, what is unique about AHCs? According to data prepared by the task force, AHCs conduct approximately 30 percent of the nation’s health care “R&D,” about 50 percent of which is federally funded. AHCs train all undergraduate medical students and nearly one-half of medical residents. AHCs provide a greater proportion of specialty services, such as burn, trauma, neonatal intensive care and organ transplant than would be expected. AHCs provide a greater proportion of charity care than would be expected, with the public AHCs providing more charity care than the private AHCs.

Three, are AHCs necessary? Dr. Blumenthal noted that in theory, each of the AHC functions can be done elsewhere. The functions need not automatically reside within the AHC and it may be useful to experiment with re-allocating some of the missions. For example, can some of the research done at AHCs (e.g., Harvard) be done at a basic science university (e.g., Massachusetts Institute of Technology)? In Dr. Blumenthal’s view, if the AHC did not exist, something like it would have to be established to provide a center for clinical innovation and to translate basic knowledge into practice.

Four, o AHCs deserve public support? The missions performed by AHCs have characteristics that predispose them to market failure. The arguments for public support differ by mission, but there are powerful arguments for a public contribution to support the missions. One argument is that AHCs produce goods that have significant externalities, such as research. In addition, the conduct of the missions incurs clinical costs that are not fully covered in the marketplace. As a result, although AHCs merit public support for conducting their mission, other institutions conducting the same activities may also merit the same support.

Five, how much support should AHCs receive? The level of support should depend on the quantity of mission-related activity that society chooses to supply and how efficiently those activities can be produced. A number of trends are affecting the answers to both of those questions, including the biological revolution and growth in NIH budgets, the aging of the population and its demand for health services, and the persistence and growth of the uninsured.

Six, how many AHCs are needed and where? The answer to this depends on the response to the previous question. According to Dr. Blumenthal, it seems unlikely that less capacity will be needed in the future, although it is possible that current capacity may need to be redistributed. If capacity expands, it should be in places where local needs for the AHC missions are also growing.

Seven, if AHCs are supported, how should that support be provided? Dr. Blumenthal believes that support should be targeted to the unique things that AHCs do and that support the special missions valued by the public. There are two general mechanisms for doing this: directly or indirectly. Approaches for direct support could include an annual appropriation for the full and reasonable costs of each mission (including the associated clinical expenses that are not covered in the marketplace). A federal agency could allocate a specified amount to each institution in proportion to the amount or quality or efficiency of mission-related activities performed. Examples of direct support to AHCs include funding by the National Institutes of Health or the Health Resources and Services Administration (HRSA). Advantages of direct support are that it is transparent, rational, can be planned and is subject to accountability. Concerns with the direct approach include the unpredictability of needed funding levels and difficulty in measuring not only quantity, but also quality and efficiency. It can also require a bureaucracy to administer the distribution of funds.

Indirect support takes the form of augmented payment for clinical services. Examples are indirect medical education payments under Medicare and some Medicaid programs, the extra payments collected by AHCs in private markets, or state regulations that allocate higher clinical payments to AHCs. The advantages of indirect support are its flexibility, continuity of past ways of doing business and reward for entrepreneurship, e.g. f an organization attracts more clinical business, it can increase its revenues. The concerns include hidden payments that make accountability difficult.

Eight, how should AHCs be held accountable? The performance of mission-related activities should be measured along multiple dimensions, such as the quantity of the activities performed, as well as their quality and efficiency. Although it may be desirable to relate payments to performance, there are no accepted measures that are valid and reliable. Additionally, AHCs are often unable to collect the type of data that would be needed on individual activities because the activities are jointly produced. Additional research and development in this area is required.

Nine, what should AHCs do differently in the future? Dr. Blumenthal believes that AHCs will need to improve their management, culture and infrastructure to make the production and organization of the missions more efficient. They will need to measure what they do and disclose the measures to the public and policy-makers.

Ten, what should government do differently? In Dr. Blumenthal’s view, the government should commit to supporting the real costs of providing the desired amounts of education, research, specialized services and indigent care. It should invest in the infrastructure necessary for AHCs to manage these costs appropriately, including measurement of the missions. Finally, it

should hold accountable all those who conduct mission-related activities, recognizing the difficulty of measuring what organizations do on a day-to-day basis.

Marla Salmon asked about the value of the interface between the academic health center and the broader university, how it can be optimized and how it might influence the type of research that is conducted. Dr. Blumenthal indicated that the Commonwealth Task Force did not analyze the interface between the AHC and the university, nor is he aware of anyone who has, although he would assume there is value.

Darrell Kirch asked if Dr. Blumenthal preferred direct funding approaches over indirect approaches. Nancy-Ann DeParle asked about direct medical education funding. Dr. Blumenthal indicated that proposals to create a medical education trust fund move in the direction of direct support, but are risky if the AHCs cannot generate sufficient income. He remains ambivalent on whether direct or indirect funding is the preferred approach for funding medical education. From a policy standpoint, direct support is more desirable, but from a political standpoint, “it is much more murky.”

Jordan Cohen asked if the task force perceived a need for fundamental change in the way society pays for and establishes accountability for the AHC missions and if there is strong evidence that current policies are "broken." Brian Biles suggested the need to consider what might ‘break’ in the future and the implications of that. David Blumenthal commented that because the health care environment is constantly changing and creating new challenges, it is not realistic to think that AHCs will not change as well. Larry Lewin noted prior discussions at the workshop that called for the redesign of the platform for the delivery of health care. Linda Aiken asked how AHCs, which train the next generation of clinicians, can create an acceptable working environment for education so people do not burn out and leave the profession. In her view, work and care processes that are redesigned in tertiary centers could serve as models for other providers. While Dr. Blumenthal agreed in principle that service delivery needed redesign however in practice, AHCs have not done this well, as demonstrated by the recent attempts at building networks. While AHCs should take a leadership role in demonstrating efficiency and being at the core of a revitalized and restructured health system, Dr. Blumenthal noted that may be asking a lot of organizations that have other missions. In order to develop and provide models, AHCs may need to reform their own environments.

Nicole Lurie asked if the proportion of R&D conducted at AHCs has changed over time. David Blumenthal indicated that the Commonwealth Task Force did not look at changes over time in this. Although private industry has expanded its research, the NIH budgets have also increased, so it is difficult to assess the change over time.

Christine Seidman asked what is known about the causes of inefficiencies in each mission activity. David Blumenthal responded that there are many causes of inefficiencies, but one is the lack of accountability. If an activity is not measured and its productivity is not examined, it is not as likely to be done efficiently. Not all inefficiencies are externally imposed. Managerial flaws exist that have lasted over time and have been difficult to address in the complicated institutions represented by the AHC.

Roy Wilson asked if medical education is a public good. In Dr. Blumenthal’s view, medical education merits public support, but he would not classify it as a public good. Education does indeed have value for the individuals who obtain it, but it also imposes clinical expenses on

the organizations that conduct education programs that cannot be captured in the private market. Ultimately, medical education provides consumer protection in that if patients cannot judge the qualifications of physicians themselves, medical education ensures that practitioners have received adequate training.

James Curran asked if it would be "simpler" to define support for the medical school and its affiliations rather than for academic health centers, so that the emphasis is on supporting the academic mission rather than “rescuing hospitals and inefficient health centers.” Dr. Blumenthal responded that from a political standpoint, there could be advantages to defining the AHC from its medical school, but that is not a definition of where the work gets done. Additionally, AHCs vary in what they produce. Some primarily train medical students. Others provide large amounts of indigent care or specialized services. Some do a great deal of research and some do very little. If a single definition is used, it should be inclusive of that variation.

Robert Dickler noted the appeal of examining each mission and its own bottom line, but questioned whether each mission can actually be viewed separately given the joint production function. Furthermore, if each mission is supported through a direct mechanism, what benefits of the joint production function might be lost? Dr. Blumenthal responded that some institutions do not produce joint products. For example, some AHCs focus almost entirely on education, but do little in the other missions. The joint production function is important and will make it difficult to hold AHCs accountable on a mission-by-mission basis, but we should still try to push forward and measure the productivity of each mission.

Claude Bennett suggested that accountability implies an ability to measure AHC performance in every mission and in every institution. Although we can measure research performance, it is less clear how that can be applied to other missions. David Blumenthal agreed that measurement can be improved. In education, the Accreditation Council for Graduate Medical Education (ACGME) has done extensive work. Other existing sources, such as AAMC’s survey of residents, could be used more effectively. But it will be difficult to come to any precise, defined measures that everyone can accept.

Bruce Vladeck, Ph.D., Senior Vice President for Policy, Mount Sinai/NYU Health

Dr. Vladeck addressed five points in his presentation—joint products, the role of Medicare, public support for AHCs, capital, and why he believes AHCs are so expensive.

The first issue related to the joint production function in AHCs. From a technical perspective, joint production means that costs cannot be allocated rationally across the functions. Revenues can be allocated, but allocating costs is intrinsically arbitrary. The joint production function also alters the meaning of cross-subsidies. Purchasing a clinical service from an AHC automatically provides some level of support for teaching and research. The same service may be purchased from another non-academic health institutions at a lower price, but if one tries to determine how much of the payment is for the clinical care and how much is for other missions, it is an arbitrary process.

Dr. Vladeck’s second point related to how public subsidies are provided to AHCs. Medicare is a poor mechanism for supporting AHCs, but it is the only lever the federal government can use. The goal of Medicare is to provide its beneficiaries access to high-quality care in a way that is at least equivalent to that of the average citizen. If Medicare beneficiaries are going to have access to AHCs, it is clear that the care will be more expensive than other

institutions. This was not a problem when public policy directed that costs be reimbursed. But when the prospective payment system (PPS) was implemented, the reasons for the greater expense could not be disaggregated and was built into PPS through indirect medical education. Now, the indirect payment mechanism is used as the tool to meet other goals, such as workforce goals, because it is the only tool available.

Dr. Vladeck’s third point related to how public support is provided to AHCs. In his view, the thinking about support for AHCs has been “fixated on Medicare indirect medical education and direct medical education payments and the hospital prospective payments. In fact .… the principal driver of behavior, on a day-to-day basis, of life in academic medical centers is the income of the faculty, not the income of the residents. Faculty income is a Part B issue.” He conjectured that AHCs experienced more turmoil with the introduction of the resource-based relative value scale for physician payment than from all the changes made in medical education payments. The significant growth in clinical faculty over the last decades has fueled the growth in clinical service revenues, which affects decision-making in AHCs more than hospital revenues. If public policy is to affect AHCs, there needs to be much greater consideration of how physicians get paid.

Dr. Vladeck’s fourth point related to capital. “AHCs have an enormous appetite for capital.” Research is growing at AHCs, which requires capital to support new space and equipment. New technology comes onto the market in steady waves and the entire health care industry is in a catch-up process relative to information technology. In the late 1980s, the transition to PPS led Medicare to remove the pass-through of capital expenses. Prior to PPS, hospitals that were operating on a break-even basis could issue bonds and obtain reasonable ratings on the debt because the debt service was a pass-through under Medicare. After PPS and elimination of the capital pass-through, AHCs are now evaluated by the rating agencies in the same manner as any other organization. That means having a margin of at least 3-4 percent to obtain a reasonable rate on the bonds. It also means that AHCs may not be able to acquire more debt at a time when the demands for capital are increasing. The issue of capital for AHCs has not received adequate attention and will slow the growth of important improvements.

Dr. Vladeck’s fifth point addressed the issue of why AHCs are so much more expensive in the first place. All costs associated with clinical services at AHCs tend to be higher. The emphasis on Medicare payment policy and trying to measure costs avoids the core question. Dr. Vladeck believes that the real reason AHCs are so expensive is because they are academic, and academic institutions are inherently inefficient in their governance, organization and management. From his view, the real public policy question is to what extent it is good to have lots of AHCs in this country. Over time, we have realized they cannot be self-sustaining because of the costs of their clinical enterprise. We have to find ways to support them, but if we get too sophisticated in thinking about what they do or how to subsidize them, we are fooling ourselves in how precisely we can identify and divide costs across the missions or measure the products and output of the AHC.

In closing, Dr. Vladeck asked that if the capital subsidies were addressed and if there were adequate subsidies for uninsured and under-insured patients (or there were no uninsured people), would that be sufficient support for AHCs to conduct their missions?

Jeff Goldsmith asked why higher costs should be paid to AHCs for their clinical services, when Medicare has an obligation to function as a prudent purchaser for its beneficiaries and

health plans have to function as prudent purchasers for their employee populations. Along a similar line, Al Dobson noted that although AHCs are more expensive, approximately 80 percent of the care provided at AHCs is “bread and butter” care that can also be obtained at community hospitals. In response, Dr. Vladeck agreed that although about 80 percent of the care provided at AHCs can be obtained at community hospitals, the premium paid to AHCs supports the other approximately 20 percent that represents the high-end monopoly services provided by AHCs, such as the neonatal intensive care unit. The costs of the high-end monopoly services are spread over the more conventional services, which are the ones that produce the revenue for AHCs. One approach would be identify those expensive, specialized services that are truly community resources, and provide subsidies for those.

Robert Reischauer asked how a resource-based, relative value scale might be adjusted for physicians working in AHCs and to what extent it might be geared to support desired public goods as opposed to other activities. In response, Dr. Vladeck suggested the need for a systematic look at physician payment.

Robert Dickler asked if the clinical role of AHCs should be constrained. For example, AHCs should not conduct certain activities that are better performed in the community, with the understanding that the AHC has a “franchise” on the subset of activities they do perform. This would require a mechanism for meeting the compensation requirements of faculty on a reduced case load. Dr. Vladeck responded that such an approach may make sense from an economic and education perspective, but is complicated by where patients want to receive their care. It is also possible that such an approach might increase costs because third- and fourth-year medical students (and their extra costs) would be “sprinkled” throughout the community, which could raise costs in the aggregate.

Larry Lewin asked to what extent AHCs can successfully compete in a managed care environment, given their joint production. Dr. Vladeck responded that some have done well and others could have done better if there was a different interpretation of antitrust legislation in terms of the way AHCs work with their physicians to negotiate with insurers. Although price is a driver in the market, it is not the only factor.

Marla Salmon commented that, in the past, nursing schools have been immune to some of the pressures facing AHCs and medical schools because they did not have clinical revenue. However, that is no longer the case and nursing is increasingly finding itself in a precarious position of financing education. Dr. Vladeck responded that adequate financing of baccalaureate and masters nursing education will never be effectively addressed until the issue of non-baccalaureate trained nurses is resolved.

Paul Ramsey asked how financial issues have affected the ability to deal with rural workforce issues and the role of AHCs for large rural areas. Dr. Vladeck responded that in his view, it is uneconomic in the modern world to provide state-of-the art, high-quality medical care to residents in isolated rural communities. It will not be possible to have a hospital or specialty practice or even a primary care group in many parts of rural America without some form of subsidization.

Christine Seidman asked if the joint production function included only the costs of clinical care and education, or if the research function is also included. Dr. Vladeck stated that the assumption is that the AHC will absorb a certain amount of the costs for research, especially start-up costs. However, it may be possible to separately identify more of the expenses

specifically related to research than can be done in education, for example. Additional research is needed in this area.

Darrell G. Kirch, M.D., Senior Vice President for Health Affairs; Dean, College of Medicine; and CEO, Penn State Milton S. Hershey Medical Center

The focus of Dr. Kirch’s comments were on the interactions between the medical school, the faculty practice plan and the hospital. Between 1960 and 2000, the population grew 56 percent, the gross domestic product grew almost 400 percent, the number of physicians grew by about 175 percent, the number of medical schools increased by just over 50 percent and the number of medical graduates grew by about 124 percent. This represents the growth in the education role of AHCs. On the research side, basic science faculty grew by over 330 percent, faster than the growth in the number of students. The most dramatic growth has been in the growth of the clinical enterprise, which experienced growth over 1,000 percent in the same time period. Most of this growth in faculty can be attributed to the desire to increase revenues.

The AHC receives revenues from a variety of sources, including tuition and appropriations, grants and contracts, and physician and hospital revenues. These revenues go into a single pool from which these diverse revenue sources are mingled together and are used to support the AHC missions in clinical care, research and education. In Dr. Kirch’s view, this pool of mingled funds needs to be made clearer and more rational. The lack of clarity in that process creates assumptions by people working in AHCs that are not necessarily supported by facts. For example, people will say, “we’re working harder for less.” In reality, faculty compensation has been leading inflation by a significant margin. Another common statement is, "No one pays us to teach.” However, evidence shows that tuition for medical school has increased and that 55 percent of medical students graduate with debts of over $75,000. The average tuition nationally now is over $22,000 per year. If all the tuition and state appropriation dollars were divided by the number of current students, it would come to about $67,000 per student, suggesting that in most state schools, revenues are probably sufficient to support the teaching mission. Another common statement is, “Our research makes a profit for the university and they don’t share it with us.” In reality, institutions do not recover their full costs from NIH grants and, additionally, fund the start-up costs of investigators who occupy laboratory space, but do not obtain grants.

Dr. Kirch identified the keys, in his view, to designing a better system. The first part is understanding the funds flow. Dr. Kirch has been in two AHCs where efforts were undertaken to understand each mission’s financial performance, so it is indeed possible to accomplish. At Pennsylvania State, both the research mission and the education mission have deficits, but the clinical enterprise is profitable. They have also explicitly identified an administrative function that technically nobody pays for, so that also represents a deficit. The clinical enterprise profits are not sufficient to fund the combination of deficits, so the balance is made up through endowment and other income. Other schools may experience different results. When Dr. Kirch worked at the Medical College of Georgia, where there was strong state support, the education mission was profitable and subsidized the research and indigent care missions.

The second part to designing a better system is program assessment. In looking at programs, Dr. Kirch includes the specific activities that are run in each department because each department has many different programs. Programs should be examined from two perspectives—their contribution to the mission and their financial performance. Some programs make important contributions to the mission, but will never cover their costs, such as the libraries.

Those programs are, and always will be, underwritten. This process helps to identify priorities and direct investments.

The third part to designing a better system is quantifying and valuing productivity. Although faculty may resent this process as being intrusive, it also sheds light on the common knowledge that the organization tends to under-reward high producers and over-reward low producers. Assessing productivity requires an agreement on the mission metrics to measure productivity. AAMC has done work in this area to estimate, for example, how much time should be allocated for preparing a lecture or for teaching undergraduate medical students.

The fourth part of designing a better system is to rethink budgets. We have to “open the books” so that department chairs can understand the budgets of the system as a whole. In Dr. Kirch’s view, the departmental structure is not the obstacle to change. Rather, it is the lack of transparency in the core financial information and planning for the organization.

Three questions arose regarding cross-subsidization. Jordan Cohen pointed out that in his view, external funding is not adequate, although he agrees that fixing these internal operational management issues would go a long toward relieving much of the stress that currently exists in the institutions. Jeff Goldsmith suggested that if the hospital and its clinical services are the primary source of capital for many AHCs, then the hospital's primary purpose becomes earning the profit so the funds can be transferred elsewhere in the organization. To the extent this is true, the clinical enterprise may not receive the resources it needs to be a high performing organization. He also questioned the assumption that research has to lose money. AHCs have expanded their research enterprise and built buildings to support it, creating large fixed costs. James Curran asked if AHCs are advocating sufficiently for needed increases in the NIH salary cap for better payment of indirect costs. These are problems specific to AHCs that the university often does not understand. In response, Dr. Kirch noted that the goal of creating a better system is not necessarily that each mission be self-supporting, but rather, that more data are available and that there is a better understanding of the financial and other performance indicators of the AHC to improve decision-making. In his experience at Penn State and at Georgia, understanding the funds flow for each mission permitted the organizations to make management decisions that improved the operations of each function. It also permitted the identification of inefficiencies to reduce their cross-subsidization.

Larry Lewin noted that in many AHCs, the faculty is under pressure to see patients, which puts constraints on the time available for teaching. On the other hand, he asked if sufficient attention is given to making the types of changes in the educational approach that Edward Hundert described earlier in this workshop that can make teaching more creative and effective, and therefore, more rewarding to faculty. Paul Ramsey asked how, after the AHC has an understanding of its funds flow, decisions are made about the cross-subsidies to support the AHC role in its local community. In response to both questions, Dr. Kirch responded that in his view, the core problem is that funds for teaching have been obscured in the co-mingling of the revenue sources. As a result, faculty has lost sight of the fact that monies are being contributed for teaching. AHCs also need to recognize that they have the power to allocate resources to teaching or to other activities and priorities. Understanding the funds flow permits the organization to allocate its resources more rationally.

John Porter asked: a) how it is possible to understand the costs of each mission if, as Bruce Vladeck pointed out, the joint production functions makes all such allocations inherently

arbitrary and b) why administrative costs were pulled out separately rather than allocated back to their function. He also commented that understanding the funds flow and which missions are profitable does not recognize that inefficiencies can exist even within the profitable missions. In terms of the first issue, Dr. Kirch responded that assumptions can and have to be made, even if they are not perfect. Secondly, administrative costs were highlighted so the faculty would know how much money was supporting committees and other administrative functions to assess whether all were needed.

Roy Wilson pointed out that although mission-based management may be a good approach for improving the availability of information, is it sufficient for making change happen? For example, tenure can impose constraints on AHC efficiency. In response, Dr. Kirch did not view tenure as a limiting factor. AHCs have to take responsibility for defining the financial guarantees that are or are not associated with tenure and providing options for investigators who work in areas that are not fundable.

Nicole Lurie asked if the morale at Penn State has improved since all of these changes were put in place. Dr. Kirch said he could not answer this objectively, but does believe it has improved and that people feel more in control of the decisions affecting the AHC.

Gerard F. Anderson, Ph.D., Professor and Director, Center for Hospital Finance and Management, The Johns Hopkins Bloomberg School of Public Health, Johns Hopkins University

Gerard Anderson’s presentation covered three areas: understanding the distribution of funds across AHC roles in research, graduate medical education and disproportionate share; issues of accountability; and the synergy of conducting the roles as a joint product.

Medicare originally paid hospitals for the costs of providing care. Over time, as alternative payment approaches have been implemented (Section 223, prospective payment system), AHCs have consistently been the most affected by any changes intended to control the costs of care because AHCs costs have always been higher than other hospitals. When the PPS was implemented in the mid-1980s and Medicare began paying hospitals a prospectively determined rate per case, AHCs were expected to be negatively affected because of their consistently higher costs. As a result, such hospitals received direct and indirect medical education payments.⁷ Since the implementation of the PPS, the United States. has spent about $100 billion on just indirect and direct medical education. In addition, approximately $75 billion has been put into AHCs by the NIH, and another $10-20 billion has been provided by the states. Dr. Anderson asked what has been obtained for the approximately $200 billion that has been provided to AHCs? What did the money buy? Funds flow can tell you where the money went, but not what it bought. If an equivalent amount is spent in the next 15 years, what should it buy?

To answer these questions, Dr. Anderson reviewed the provision of funding to AHCs from NIH (as a measure of the size of the research activity), from Medicare for graduate medical education (as a measure of the size of the teaching function), and Medicare disproportionate share funds (as a measure of the size of the indigent care program).

NIH funding in the year 2000 to medical schools was approximately $8 billion, most of which is concentrated in about 135 institutions.⁸ One-third of it goes to the top 10 institutions, which average about $280 million each. The next 40 institutions receive about 50 percent of the money for an average about $110 million each, and the remaining institutions receive about 15 percent of the total. The criteria for allocating this money is through a peer-review process.

Money for graduate medical education (GME) is much less concentrated. Almost 1,000 institutions receive GME funds. The top 10 receive just 12 percent of the money (an average of $60 million each), the next 40 institutions receive about 24 percent of the money (an average of $30 million each) and the remainder of the group receives about 64 percent of the total money. The money is allocated to hospitals based on the number of residents trained, their costs in 1985, and their share of Medicare beneficiaries. Are these the right criteria for distributing education funds? In Dr. Anderson’s view, there is “no real accountability.” When this formula was developed, there was no requirement to do anything except to train residents.

Money for Medicare disproportionate share (DSH) is the least concentrated, going to approximately 4,000 institutions. The top 10 institutions get relatively little money, about 5 percent of the total, for an average of $20 million each. The next 40 institutions receive about 11 percent of the total, for an average of $10 million each. The remaining institutions receive about 85 percent of the total. The criteria for receiving DSH funds is based on the number of Medicare and Medicaid patients served. He asked if those are the right criteria.

Dr. Anderson then asked if AHCs have to perform in all three of these missions. He arrayed the top 50 institutions in each mission and looked for the overlap. He found that 11 institutions are in the top 50 for all three missions, 8 are in the top 50 for both GME and DSH, 15 are in the top 50 for both research and GME, and 2 are in the top 50 for the roles in both research and DSH.

Are there economies of scale when AHCs provide all three of these missions or are these separable? Dr. Anderson reviewed the characteristics of the top 50 institutions, dividing them into categories according to the number of missions (graduate medical education, NIH research funding and disproportionate share) in which they appeared in the top 50 (e.g., whether they appeared in the top 50 for conducting all three missions described, the top 50 for conducting two of the three missions or the top 50 for conducting one of the missions). Those institutions that appeared in the top 50 for all three missions tended to be larger and more expensive (based on Medicare cost per discharge). Compared with institutions not in the top 50, they tended to care for fewer Medicare patients, but more Medicaid patients. They also trained the fewest proportion of primary care residents. If that is an important education goal, it is not happening in these AHCs.

Dr. Anderson challenged the idea that much is known about funds flow within AHCs. In his view, the money that goes in AHCs is very fungible and the question for this committee to consider is whether there should be accountability for its use.

Dr. Anderson has been working with others, with support from the Department of Health and Human Services, to identify the early warning indicators that show when an AHC is in distress and changing its mission. Indicators were developed for each mission based on measures that were viewed as important and had data available or that could be collected. The indicators for the education mission are: 1) primary care residents as a percent of total residents, 2) number of unfilled chairs over one year, 3) board certification exam first-time pass rates for residency graduates, 4) faculty attrition, 5) faculty satisfaction, 6) undergraduate student ratings of educational experience, and 7) proportion of international medical graduates. The indicators for the research mission are: 1) total government research dollars, 2) total government research dollars per FTE faculty member, 3) total NIH dollars per principal investigator, 4) total private sector research dollars, 5) total private sector dollars per FTE faculty member, and 6) number of publications by faculty. The indicators for the patient care mission are: 1) nurse- to- patient ratio, 2) re-admission rates, 3) episodes of uncompensated outpatient care, and 4) unfilled positions in the intensive care unit.

In closing Dr. Anderson reiterated his view that the core issue for AHCs is one of accountability. What should be expected from AHCs in return for public support of their roles.

Nicole Lurie asked if most AHCs have the same kind of funds flow shown by Dr. Kirch. Dr. Anderson indicated that is unknown, but believes that funds flows vary across AHCs.

Robert Dickler agreed on the need to understand where these public dollars go and the need for accountability. However, he raised a number of concerns with the analysis presented. It is hospital-centric, attributes NIH funds that go to medical schools to the hospitals, mingles direct and indirect medical education funds, and uses absolute dollars without accounting for institutional size. He believes that additional analysis and refinement of these figures are needed. Furthermore, he also noted that methods of accountability are built into the current system. For example, residency programs are not funded if they are not accredited. Institutions receive money based on how many residents they have and their real costs (at a point in time). NIH money is provided based on grant applications and peer review. Disproportionate share money is distributed based on an underlying methodology. Prior attempts in setting policy objectives have been difficult. For example, efforts at predicting workforce supply needs and mix have not been successful. In his view, the question is how to support accountability for certain criteria while retaining flexibility.

Dr. Anderson agreed that the analysis is primarily based on hospitals because those are the data available. Data are not available on the AHC as a whole. The concentration on hospitals was because most of the monies described flow to the hospital, not the AHC broadly. He continues to believe that there is less accountability for graduate medical education compared with the other missions. The standards set by the Accreditation Council for Graduate Medical Education (ACGME) do not talk about the balance between specialists and generalists or outpatient training, for example.

Linda Aiken commented that Dr. Anderson's data showed that AHCs do not appear to be very different relative to nurse staffing, but that case mix varies substantially. She asked if this suggests that AHCs are under-investing in patient care, or perhaps the cross-subsidies are drawing resources out of patient care to fund other missions. Dr. Anderson suggested that there may not be enough nurses in AHCs, however, it is also possible that residents are substituting for nurses.

However, research he conducted in the 1980s suggested there was stronger evidence that residents were substituting for the attending physicians.

Brian Biles asked about the Medicare formula for indirect medical education. The goal for implementing it was to keep teaching hospitals financially solvent as the system moved to prospective payment. Is a more sophisticated formula required today? Dr. Anderson replied that more sophisticated formulas can indeed be developed, but the objective needs to be known so the formula can be designed.

Al Dobson commented that the formulas in place today were designed to meet a policy goal of ensuring access for Medicare and Medicaid beneficiaries in AHCs that have high costs and providing support for interns and residents. There may be more questions today, but those were the policy goals in 1983. Dr. Anderson agreed that the funds are flowing exactly according to the formulas developed. The question, however, is whether those are the right objectives today.

Jeff Goldsmith reiterated his view that there will be tremendous pressure on AHCs to expand the number of trainees in the next 10-15 years. Much of the discussion has focused on using cost as a basis for deciding how to subsidize various activities. The result has been wide variations in the amount of support provided to specific institutions. Dr. Goldsmith asked if this is a viable basis for policy going forward. Dr. Anderson responded that expanding training programs will require an assessment of what we want them to accomplish. NIH faces these questions annually when it has to decide how much spend on cancer or AIDS. Although the future cannot be predicted, we can say where we think the problems will be and allocate money accordingly. We know that the needs of the 21^st century are going to be more oriented toward chronic care and will be less hospital focused. Even if projections are imprecise, the alternative is to let each AHC make its own decisions with the current set of economic incentives, which will continue to have an acute care, inpatient focus.

Larry Lewin challenged the statement that DSH and NIH funding mechanisms have more accountability built into them than GME funding does. He noted that DSH funding is imprecise, yet has a significant impact on the size of the safety net, where it is located and whether it is oriented to outpatient or inpatient services. Dr. Anderson replied that in his view, graduate medical education was the least accountable, but it may be true that accountability for DSH dollars could also be clarified.