3
Keynote Address
John M. Eisenberg
This workshop addresses one of the fundamental issues of medical care in the future. As health care systems change, as we improve our ability to avert death from acute diseases, and as the population ages, the care of individuals with chronic disease becomes more central to the mission of health care worldwide.
As requested, today I will address "changes in the system of health care organization and finance as they relate to management of persons with chronic disease." I will explore what our best leading indicators predict the effects of these changes to be. I will discuss some broad trends in medical care today, comment on the payment of physicians, and reflect on the overall system of care. I will also discuss managed care, the training of physicians, both the tension and the collaboration between primary care physicians and rheumatologists, and what these forces mean for the rheumatology work force. I will conclude with comments on issues related to the intersection between health policy and clinical practice.
Changing Health Care
In preparing this talk, I looked into the crystal ball that sits on my desk. Reflected there I saw trends in the future for which we should be planning today.
First, it is clear that hospitals will downsize and close or, as an intermediate step between downsizing and closing, merge. Whichever scenario occurs, a significant decrease in the number of hospital beds will take place.
This decrease in hospital beds will continue the trend of moving the focus of medical care, especially for chronic diseases, away from the inpatient setting.
Second, managed care organizations will compete on price. What is not so clear yet is whether they will be able to compete on quality as well. This is a fundamental issue for health care delivery and certainly for the future of chronic disease care.
Third, technical services, which have long been sources of revenue, are becoming sources of cost. This changes the way we look at services such as endoscopy, cardiac catheterization, and surgery. We can no longer use them to comfortably cross-subsidize specialist care for chronic conditions.
Fourth, physicians' rewards and relationships will continue to change. We are moving away from traditional fee-for-service payment and higher payment for more technical procedures. Our professional relationships no longer will be determined solely by our personal preferences, but will be determined by new payment systems and re-engineered systems of care.
Fifth, gatekeepers will manage patient care, primary care will become primary, specialists will specialize, and consultants will consult.
Finally, physicians will redefine their relationships to organizations and third-party payers. More physicians will be employees; fewer will be self-employed.
Hospitals
The decreasing use of hospitals is an important element in the changing face of medical care. Medicare-reimbursed hospital stays are already shorter and admissions fewer. The national average hospital use for Medicare is about 2,800 hospital days per 1,000 people annually. The California health maintenance organizations (HMO) rate is about half that, and for California integrated systems of care, the rate is only 960 days per 1,000 people per year. Soon, we will likely experience as few as 800–900 hospital days per 1,000 Medicare beneficiaries per year. That is about one-third of the days of hospitalization currently used.
Commercially reimbursed hospital stays are shortening as well, and admissions are fewer. As a result, the national average is about 500 hospital days per 1,000 people annually; for California HMOs, 250; and for California integrated systems of care, 200. Many will remember Kerr White's description, a generation ago, of the health care system being like a pyramid. For every 1,000 people there were 1,000 days of hospitalization per year. No more. We will soon be experiencing 200 days per year per 1,000, at least for relatively young populations.
Managed Care
As different types of managed care plans emerge, it becomes increasingly difficult to consider managed care as a monolithic form of health care delivery. They are as different from each other as they are from traditional fee-for-service medicine. The heterogeneity is remarkable!
- Staff model HMOs hire their doctors directly. There are few of these plans—Group Health of Puget Sound and the Harvard Community Health Plan, for example.
- Group model HMOs such as Kaiser Permanente contract exclusively with a single physician group.
- Network model HMOs differ from staff and group in that they contract with several physician groups, usually on a non exclusive basis.
- Independent practice associations (IPAs)—in an IPA model, doctors generally practice in community-based offices. They may have individual practices or share a practice with a group of physicians. They contract with the IPA on a non exclusive basis, sometimes, but not always, taking capitation.
- Preferred provider organizations (PPOs) are organizations that contract with individual doctors for fee discounts to list the doctor as a "preferred provider" for their enrollees. This is not really a new model. In fact, in the mid-1980s, Blue Shield of Pennsylvania contracted with almost every doctor in Pennsylvania, set fee limits, had a vigorous utilization program, and could drop doctors from the plan. PPOs are basically a way of finding a group of doctors willing to accept a discount and putting them on a list of physicians eligible for a negotiated or previously established fee-for-service payment.
- Point-of-service (POS) plans are still called "managed care" plans, but they are really open-ended plans with a list of participating doctors. If subscribers go to the listed doctors, they can get covered services for either no copayment or a lower copayment than from doctors who are not plan participants. Sometimes in these plans, patients have an even lower copayment if they are referred by their gatekeeper than if they seek service from consultants on their own. Thus, there may be three levels of fees in some point-of-service plans—referral in-plan, nonreferral in-plan, and out of plan.
What can we conclude from this brief survey? Managed care is not any one of these particular models. When we talk about managed care we should avoid getting trapped in the notion that there is a single model of managed care synonymous with Kaiser Permanente. In fact, a substantial amount of managed care is paid fee for service, albeit at a discount.
Enrollment in group and staff model HMOs has been relatively fiat for some time. In fact, in some areas they have lost market share and numbers of patients. PPO and POS plans, however, have been growing rapidly. They do
not alter the nature of medical practice as much, change relationships among physicians, or change the organization of care. The more restrictive plans may alter referral patterns or, at least, redirect referral patterns. However, in most instances all they do is force physicians to sign with a plan and accept a lower fee. They also may not save as much money as group and staff model HMOs do.
Much of the change in health care systems stems from buyers, and now sellers of care having organized in response to the need to negotiate fees. As a result, more physicians are moving toward larger multispecialty group practices, such as physician networks, to gain negotiating power. If the buyer of services has market power and the seller does not, there is an inequity. The sellers of service, in this case doctors, have now begun to consolidate in order to gain market power that matches that of the purchasers.
How much of this can occur before doctors are in restraint of trade? If consolidation reaches the point of one organization (or a very few) negotiating on behalf of all doctors, the market advantage swings from buyers to sellers, from health care plans to doctors. In that case, medical practice will border on monopoly or oligopoly and will be subject to intensive antitrust scrutiny.
Many believe that there will be both vertical and horizontal integration of medical practice. Vertical integration is coordination, sometimes but not always common ownership, of all the various levels of care from hospitals to specialty care, primary care, nursing home care, and home care. Horizontal integration means that multiple hospitals and multiple groups of physicians come together at the same level of care—be it hospital, specialist, or primary care—to consolidate their activities in a particular geographic area. We are moving from a model of physicians in solo practice to one of physicians enmeshed in different modes of consolidation as the market propels us toward both vertical and horizontal integration.
Physician Payment
Physician payment will change dramatically. There are three major questions about payment: What are we going to be paid for; who is going to be paid; and what controls will there be on the way we are paid? Traditionally, in the United States, we have been paid on a fee-for-service basis according to a market price structure—whatever the doctor could get. These market-driven fees were soon locked into place by a customary, prevailing, and reasonable payment system used by Medicare and many third-party payers. The market had stopped influencing physicians' fee levels; third-party payer fee schedules, based on historical data, were now determining fees.
With the 1989 Omnibus Budget Reconciliation Act, we moved to the resource-based fee-for-service system for Medicare, built on the premise that
in a free market, what one is paid is determined by the cost of producing a product as well as the demand for that product. Doctors are now paid by Medicare based on the resources consumed in taking care of patients.
A major payment change is the increasing popularity of payment not for actual services rendered but for taking responsibility for the care of a patient. This could be capitation payment—taking responsibility for subscriber's care for a year—or it could be global payment—taking responsibility for a disease or disorder (e.g., diabetes) or a procedure (e.g., bone marrow transplantation). Hospital payment may or may not be included. Another form of payment that we may see in the future is bundled payment for an episode of illness. However, these forms of payment are not really a new concept. Obstetricians have long accepted fixed payment for prenatal care and delivery. In any case, the physician is clearly being paid for accepting responsibility for care rather than for providing a particular service on a particular day.
Traditionally, doctors have been paid individually for their services. However, increasingly we are practicing in teams and in organizations. Of course, this is a an important cultural change for our profession. We will need to develop payment mechanisms that recognize the contributions of individuals practicing in teams or organizations.
What controls will govern our payment systems? Capitation, of course, is the ultimate control. It tells physicians, ''Here is what you get; live within this budget.'' Short of capitation, four other payment controls are popular.
First, under the gatekeeper system, a medical arbiter (usually a primary care physician) decides what care will be given, sometimes with incentives for that gatekeeper. The gatekeeper controls the flow of services that the system provides.
Second, copayment puts some responsibility on patients. If they have to pay more, they will be more concerned about the cost of care and its appropriateness.
Third, utilization control represents the classic American approach. Let everybody do what they want, but slap them on the hand if they do too much of it. This approach invokes monitoring medical practice and auditing doctors for aberrant practice patterns, then penalizing them for inappropriate utilization.
Fourth is the use of supply constraints. That is, limit the number of physicians, limit the number of hospitals, and do it either through certificates of need (which don't work very well), regional planning (which didn't work very well), or the market, which does seem to be working—although causing major perturbations and dislocations as it decreases our capacity to provide medical care.
Specialist-Generalist Interaction
Who will provide chronic care services in the future? Will we have a
work force of health professionals capable of providing needed services?
About a decade ago, the American Board of Internal Medicine (ABIM) established a Task Force on the Future Internist to propose the kind of training internists will need in the future. Although there might be more agreement on this list today than 10 years ago, not much has changed in medical education in response to the changing nature of practice—the emerging dominance of chronic disease, the need for more attention to the healthy patient, and the changing organization of health care.
About the same time, that the ABIM task force was studying changes in physician education, I wrote an article in which I addressed the need to train primary care physicians, especially general internists, for their role as gatekeepers. This training should prepare physicians to
- Evaluate risks, benefits, and costs of potential treatments;
- deal with a wide variety of clinical problems and settings;
- maintain strong primary care skills;
- develop skills in working with consultants; and
- teach primary care physicians and specialists how to collaborate.
Each training area needs more attention, but in my opinion, none is greater than the need to train specialists to collaborate with primary care physicians. Medical residents still ask subspecialists, "Why don't you send patients back to me when I refer them to you? Why don't you write me a note? Why don't you learn to be a consultant instead of stealing all my patients away from me?" The subspecialty fellows respond, "How come you never send me any consults? You are managing all these patients with rheumatoid arthritis and lupus, and you never send me a patient?"
This debate between specialists and generalists can end up as a war over patients. It is not only the patients who will suffer but also future specialists who now so rarely learn how to be consultants and how to collaborate with primary care physicians.
What does this mean for rheumatology and other specialties? What will rheumatologists do in the future? What will their practices look like? Unfortunately, rheumatologists have not clearly defined for themselves how they are likely to practice in the future. They may consult on referred patients with severe or unusual diseases of the musculoskeletal system and immunologic disorders. They may be primary care doctors for patients with rheumatic diseases. They may be primary care doctors and also provide rheumatologic care for selected patients.
To what extent do rheumatologists function as primary care doctors now? There are no adequate, unbiased data on the quality of care they already provide. The best data available, although somewhat dated, suggest that rheumatologists spend about half of their time as principal physicians. Whether
they are truly serving as primary care physicians is not clear. Consultation comprises only about one-sixth of the activities of rheumatologists. Thus, even though it is a consulting specialty, consulting is a relatively small percentage of rheumatology practice.
Can rheumatologists be effective general internists? I believe that a two-year rheumatology fellowship should not nullify prior board certification in internal medicine. As well-trained clinicians, rheumatologists should maintain and update their skills in general medicine (e.g., rheumatologists should be familiar and comfortable with performing pap tests themselves rather then sending patients to a gynecologist). Equally important, they should enjoy providing primary care in addition to specialty care.
William Winkenwerder has written that the functions of primary care and specialty physicians clearly differ. Primary care doctors navigate, negotiate, evaluate, educate, and make decisions, he writes. Consultants consult, providing diagnostic and rehabilitation expertise, which is underemphasized in most of our training programs. Consultants also periodically review patients and educate both patients and their primary care physicians.
I believe it is time that we consider an alternative to such clearly demarcated boundaries between primary care and specialty medicine. Perhaps some doctors will be generalists with an area of expertise. For example, is there is a role for a generalist who has completed six months or a year of rheumatology training? Is there a role for the fully trained subspecialist who maintains his skills as a generalist?
The American Board of Internal Medicine has taken the position that specialists not only can, but should, maintain their skills as generalists. The upcoming recertification exam encourages physicians to become recertified in internal medicine as well as in their subspecialty.
Do specialists provide better care? There is evidence that receiving care from a rheumatologist seems to slow disease progression and to be associated with higher functioning levels for rheumatoid arthritis (RA) patients. However, there are also data showing that specialists do not provide better care and that, when they do, it is because of systems of care, especially the involvement of non physician health professionals.
Even if we could agree on the proper role of rheumatologists, work force issues would likely remain fuzzy. Data of the Graduate Medical Education National Advisory Committee from the early 1980s demonstrated that the supply of rheumatologists totaled about 3,000, whereas we needed only 1,900 (i.e., we had about 1.5 times the number of rheumatologists needed).
Jonathan Weiner39 also concluded that an oversupply exists, suggesting that we have 0.9 rheumatologist per 100,000 in the United States but need between 0.4 and 0.7. Another study also suggested that we have about twice as many rheumatologists as we need. In Europe, the need for rheumatologists is perceived to range from 0.3 to 0.5 per 100,000, and that is actually close to the European supply today. Meenan and colleagues suggested that if the methodology of the Committee on Graduate Medical Education is used, the supply of rheumatologists is actually less than the need.
I think there are some problems with using specialty as a proxy for capacity. Specifically, there is overlapping capability of providers. Providers may be capable but unwilling to perform some necessary services. We assume that a rheumatologist trained 30 years ago provides comparable services to a rheumatologist trained today.
Another reason for the uncertainty about the number of rheumatologists needed is that estimates of rheumatologic populations vary substantially. These work force studies also make assumptions about referral patterns that are likely to change in the future. As we think about the need for rheumatologists to care for chronic disease patients, we also must consider whether there will be alternative providers for any of these chronic diseases.
Policy Issues
How will the contributions of this conference affect the care of individuals with chronic rheumatologic disease? I believe there are three levels of policy that may be influenced by a conference such as this.
The first level is that of public policy. Is there a set of recommendations that can be made at the public policy level, for example, approval of a new service by the Food and Drug Administration (FDA) or new coverage by Health Care Financing Agency (HCFA)? The second level is clinical systems policy. In emerging systems of care, policies are being instituted not by individual physicians, but by managed care organizations, hospitals, and health care systems. Often these policies are aimed at saving costs, but often they are also aimed at improving clinical outcomes or patient satisfaction. The third level is clinical policy, for example, the institution of guidelines, audits, or criteria. These may come from professional societies or expert groups.
New research can inform policy. The Agency for Health Care Policy and Research is a beleaguered agency, but it is the only one that has accepted
primary responsibility for linking health services research to policy. This responsibility for linking research to policy, as well as for linking policy to politics, is essential to improving systems of care. As medicine and rheumatology enter a new practice era, we need to address these key issues. With careful thought about innovative payment schemes, appropriate numbers of trainees, and responsiveness to the need for outcomes assessment, the future of rheumatology and—more fundamentally—the care of patients with chronic disease will be brighter.