Marilyn Moon, PhD
Senior Fellow
The Urban Institute

February 2002

The importance of the traditional fee-for-service portion of Medicare, which accounts for about 85 percent of Medicare spending, and the failure of Medicare+Choice to develop innovative programs for caring for chronically ill beneficiaries has led many to advocate for improvements in the traditional fee-for-service portion of Medicare to meet this important need. Such "modernization" of Medicare can take many forms, ranging from greater physician participation in managing medications to formal care management activities involving a team of providers.

A number of obstacles stand in the way of implementing these types of changes in fee for service. First, proposals for changing Medicare often have the explicit goal of achieving savings for the federal government. For example, the CMS-sponsored demonstrations now underway have as a condition of participation that these benefits cannot add to what it would otherwise cost Medicare to serve these beneficiaries (Chen et al 2000). And the recent release of a notice for yet another demonstration project in February of 2002 also includes a call for savings to the federal government (CMS 2002). In part, this reflects a strongly held view that there is substantial overuse and waste in Medicare. But underuse also occurs, and improving coordination and provision of care for those on Medicare may mean increased use in some circumstances. This is particularly true when considering the important role of outpatient prescription drugs -- a benefit not now included in Medicare -- in treating chronic illness. If coordination results in more appropriate care, at the least, it will likely result in better use of resources and perhaps lower use of some services over time. But requiring savings from the outset will likely eliminate many potentially valuable approaches.

A second major issue is overcoming the reluctance of participants in fee for service -- both physicians and patients -- to give up any of the flexibility that they now enjoy. If coordinated care approaches are voluntary, it will likely be necessary to provide incentives to make patients and physicians willing to participate. Successful models could be applied universally over time and not require direct incentives, but if the rules are restrictive, many of the stakeholders may resist such change. Further, additional groups such as hospitals, may also oppose change, if, for example, it appears that they will lose business.

Fee-for-service rules also can serve as a strong barrier to care coordination. As Wagner (2001) has pointed out, a "problem-oriented visit" approach to care as we have now is not consistent with care coordination approaches. Another important issue is the level of payments to physicians under Medicare for general office visits; low payments both in Medicare and elsewhere have resulted in many physicians spending less time with their patients and hence less likely to spend time on issues beyond immediate treatment concerns. In addition, Medicare rules sometimes create rigid requirements concerning how often certain tests can be given and whether a doctor can have an interest in a facility to which he refers patients, for example. Some of these may be needed for other reasons, but they can complicate coordination activities.

The Role of Economic Incentives

The contribution of economics to health policy is based on the importance of financial incentives and the feasibility of using such incentives more effectively. Financial incentives can be strong motivators in health care as well as in other parts of our economy. But they have not always been applied sensitively, resulting in unintended consequences and a poorly functioning market for health care. For example, the approach of many economists to the recognition that fee for service medicine implicitly encourages overuse of services was to embrace the concept of full capitation of health care coverage. On the other hand, capitation offers equally strong incentives for payers to skimp on care, resulting in underuse.

The failings of fee-for-service medicine are well known. Physicians and other providers make money in this system by providing care. Even if they are trying to be responsible to their patients, the need for tests and procedures is often a judgement call. For a patient anxious to know the cause of a problem or to treat a condition, the easiest and most profitable path for providers is to do more rather than less. Moreover, doing less likely means spending more time with the patient explaining why a test is not needed. This is not only unprofitable for the physician, it is not a position that patients may readily accept. Because Americans like technology and dislike uncertainty, patients themselves may push for more care even when they bear some of the financial risk. The problems with fee for service consequently are often cited as reasons why even modest coverage expansions can be very costly.

One of the theoretical promises of managed care was that if the incentives were properly aligned by offering a fixed payment per patient, it would be in the interest of the insurer to coordinate care to achieve good outcomes and greater efficiency. The flexibility offered in such a system could lead to innovation and more rapid change. Implicitly, support for managed care was based on a belief that such efforts could generate savings. In practice, however, the activities of managed care under Medicare have fallen far short of this goal (Gold 2001; MedPAC 2001; Thorpe and Atherly 2001; Moon 2000). First, plans participating in Medicare discovered early on that the easiest way to be profitable was to attract healthier patients to their plans. In that way, they could offer benefits with few constraints to those who signed up. Unless the system finds ways to adjust for differences in health status and health risk, plans will not face incentives to coordinate care. Further, managed care plans for Medicare and the population as a whole have found coordination of care difficult, time consuming and often not a big cost saver -- at least initially. Thus, when plans can choose from an array of activities to restrain growth in the costs of health care, coordination is not necessarily high on their list.

There are some examples of plans using innovative approaches to disease management or other coordination activities in managed care environments, but these remain limited to a few good plans (Felt-Lisk and Kleinman 2000). Loosely organized networks of providers yield cost savings through discounted fees, but not from innovations in the delivery of care. Such plans have sought other ways to hold down their costs with strict rules about referrals, penalties to physicians who order too many tests, and other rather arbitrary limits that have served to anger their patients with little evidence of achieving better care. The resulting loss of confidence in managed care in general is likely to slow if not eliminate growth in managed care enrollment under Medicare. Fee for service is likely to be the choice of a majority of beneficiaries for some time to come. Thus, there is a strong interest in finding ways to work within the traditional fee-for-service framework of Medicare to provide more incentives to coordinate and sometimes manage care being received there.

Is there a middle ground? That is, is it possible to improve upon incentives in fee for service without going the whole way to a fully capitated, managed care model? Part of the answer depends upon the goal of such an approach: is the desire to simply save costs to the federal government of providing care, or is it to improve the delivery of care with the hope that efficiencies in the care provision can be achieved and at least some savings attained over time? The latter is a more realistic goal because it recognizes that in a fee-for-service environment where patients will be able to make choices on their own, they will only participate willingly if they see some advantage for themselves.

Thus, one key take away message is that if care coordination is to occur in fee for service, incentives must be appropriately established for both providers and patients, unless the system is to be mandatory. And quality of care needs to be rewarded explicitly as well. But proper financial incentives are not enough. The interest and dedication of providers of care to improve delivery and to be willing to change their behavior will be essential as well. The key role of economics is to find a way to align the incentives so that such innovation can occur, while recognizing that accountability for public spending will always be an important element as well.

Thinking About Approaches to Care Coordination

What is care coordination? This is an important question because it can take many forms, each raising particular advantages and disadvantages. For purposes of this discussion, think of care coordination broadly; indeed, one of the challenges for moving forward in this area is to explore a range of issues and narrow them on the basis of feasibility and other criteria.

Chen et al (2000) describe care coordination programs as "those that target chronically ill persons "at risk" for adverse outcomes and expensive care and that meet their needs by filling the gaps in current health care." They conclude that if these programs do all that they are supposed to do in terms of identifying needs across a broad spectrum, addressing those needs and monitoring patients, that quality of care can be raised and need for costly services reduced. They further divide the area into two types of programs: case management and disease management. The former is needed for a small group of patients with multiple conditions, while disease management is likely to be applicable to a greater group of beneficiaries with needs in only one disease area. Disease management programs, they argue, can be more standardized as there is likely to be an agreed-upon course of treatment. Case management deals with high risk patients who likely need more customized care.

The fifteen programs underway in the CMS demonstration also illustrate how disparate these activities can be. The programs all involve risk taking with capitated payments for the extra services. And most, but not all, include care delivery and/or assessments in the home. Many of the projects focus on patients with specific disease. But efforts vary on the level of physician involvement, whether participation by patients is voluntary or if they are assigned to a program, and both the flexibility and comprehensiveness of benefits offered. The newly announced demonstration sets out to explicitly involve disease management organizations whose responsibility it will be to manage all care for those who enroll. This seems to be providing full managed care for persons in the specified disease categories. The enticement for beneficiaries would be the promise of prescription drug coverage.

Many advocates of care coordination activities caution that it makes sense to first make very modest changes and have modest goals to validate the approach (Leapfrog 2001). This is likely to be both less threatening to all parties: payers, patients and doctors. It will also, they argue, create momentum for going further in this area and taking more risks in the future. Thus, with a number of demonstrations underway, is there more that could or should be done in the short run?

Lessons from Various Medicare Programs

The evidence on why individuals have gone into Medicare+Choice indicates that, for many, it is the lure of additional benefits and not the philosophy of a system of coordinated care. In fact, patients are often unhappy to learn that they face restrictions, particularly in the choice of physicians. This suggests that any model that limits freedom of movement by the patient may need to offer a quid pro quo of some sort, particularly if broad-based participation is desired.

On the other hand, there are examples of willingness of individuals to go into alternative settings where a main function and selling point is a different type of care. The PACE program, for example, offers patients with substantial health problems a very intensive coordinated care model, and although it offers extra benefits as well, patients are likely to be attracted to the model itself. Perhaps because of this or because it is aimed at a small and specialized group of very frail beneficiaries, PACE remains a small program, serving just over 8,000 beneficiaries at present in a number of sites around the country.

A more relevant model in Medicare may be hospice. Hospice is instructive as an alternative delivery system for persons in the last six months of life. In this case, beneficiaries must forego curative care for their terminal disease in exchange for additional support services and prescription drugs for palliative purposes. Enrollment in this program has grown substantially over time, now serving about 400,000 beneficiaries each year. When it was introduced, there were concerns that it could raise the costs of care as a new benefit; consequently, the rules for participation were made rather restrictive, including the requirement for a physician to certify that the patient had a life expectancy of under six months. Today, there seems to be little evidence of abuse or misuse of the program; indeed, the major issue facing hospice is whether the barriers are too intrusive and more flexibility is needed (Moon and Boccuti 2002).

One of the important aspects of hospice is that it offers a package of services through a specific entity, but allows patients to keep their own primary care physician. This is likely to be an important factor when looking for models of care. Nonetheless, one of the barriers to expanding hospice cited by many proponents is the lack of knowledge and understanding of the program by physicians.

The Range of Potential Activities and How They Would Change Delivery of Care

Drawing from some of the literature in the area, including descriptions of the Medicare demonstrations, Figure 1 lists the types of changes that are often included in care coordination activities. Some are clearly additive: that is, they build on the current benefit package, but others seek in some cases to replace or substitute for other types of care. An emphasis on cost savings would tend to favor activities of the second type, although other activities could also improve health and functioning by older and disabled persons over time.

Among the additive pieces are a number of relatively simple activities to enhance physician’s practices. For example, improved information systems to track patients and their treatments, history and medications, and to provide physicians with suggested protocols or better drugs are examples of some possible enhancements. While some physicians may invest in these activities on their own (particularly if they have the economies of scale of a large group practice), is it in the interest of the Medicare program to help subsidize such new systems? Further, should there be additional resources made available to physicians to actively manage care for their patients? And should this continue to be on a fee-for-service basis?

One potentially very expensive add-on is the provision of prescription drugs to patients participating in a care management program. Even if limited to only those drugs associated with a particular chronic condition, this may represent an important component for achieving compliance with recommended care, but it may be a "budget breaker" for those seeking savings. Further, activities that involve visits to patients in their homes can provide help to improve the overall environment and provide education about drug use and other self-administered care, but raise the same issues as prescription drugs.

Applying protocols, on the other hand, is an approach designed to steer physicians and patients, likely substituting some types of care for others. Further, this can often mean management oversight by nurses or other non-physician professionals for routine follow-up, for example, instead of more expensive physician office visits.

Alternative Structures for Improving Care Coordination

All of these offer different possibilities for care coordination; but a key criterion in choosing among them is what will work in a fee for service or largely fee for service environment. This will affect how programs would be organized in terms of both payment approaches and where control for the coordination is centered. That is, freedom to choose your own doctor is greatest under fee for service where most participate in Medicare and there is less likelihood of physicians dropping out of Medicare as sometimes occurs under managed care. Thus, physician-based approaches may have an advantage in a voluntary system. Or at the least, they need to be an integral part of care coordination activities. From a beneficiaries’ standpoint, incentives to join a program such as a disease management effort that is outside of their physician’s control or primacy may need to include coverage of additional services such as prescription drugs.

Physician-Based Approaches. Since physicians are at the heart of the delivery of care, it makes sense to first consider what physicians need and how far they are willing to go in care coordination. This could be either a quite modest effort or one that is more expansive. For example, additional fees for physicians who take on the responsibility of assessing and coordinating care on a small scale could be done by simply offering a management code to allow for an annual visit for this purpose. And the level of reimbursement could be adjusted for patients’ conditions. Qualifying patients with chronic disease could designate a physician to serve in this role, for example, receiving a broader overview of issues such as drug monitoring, and creating a plan of care.

Further, to track patients and their needs, physicians must have access to good information, both on treatment issues and on their specific patients (Oswald 2002). What could be done in this area? Information is a resource that would be difficult for physicians themselves to set up, but could be developed and then offered to doctors for their own use. Some of this activity is already underway through innovations in office-based systems and internet information. But an explicit effort to facilitate tools for better care of specific groups might need a jumpstart on its own. Certainly a survey of what is now available and how extant systems could be modified to be useful should be undertaken. Grants and contracts could also be designed to tap key care providers in various communities to experiment with such activities as a spur to innovation. Support for such information systems at the physician level could either be built into this assessment activity or added onto all fees that physicians participating in the program are allowed to charge under Medicare.

One frustration commonly expressed by physicians is the inability to spend more time with patients and have a broader sense of their needs. This additional tool might be viewed very positively by physicians and patients as a way of turning back the assembly-line feel that has frustrated both groups. This would likely, however, emphasize more of a medical model rather than addressing the broader range of social and functional needs. To get to that stage may require a more formal team approach.

A more expansive approach might be to provide a capitated payment to such physicians on behalf of each of their qualified patients to include assessments but also a more flexible range of services including telephone monitoring and consultation, communication and education provided by nurses or other professionals under the management of the physician. The appeal of such an approach is that it is physician-centered and hence would not require patients to adjust to a new overall provider. The non-physician members of the team could undertake some activities that were previously restricted to physicians (or simply not covered under Medicare). The physician could decide how to organize such a system, offering nutritional and educational counseling for diabetics in a small group setting, for example. They could contract with home health or other agencies to provide some of the follow-up services.

Beyond economic incentives, however, is the important barrier of the lack of acceptance of this type of approach by physicians, particularly when it involves a team. While physicians may not want intrusion from outside into their practices (as is suggested in some of the models described below), it is not clear that they are willing to take this on themselves. Physician willingness to be proactive may be a very important indicator of how care will change over time.

This physician-level approach has the advantage of keeping physicians in the loop and finding ways to motivate change that may be viewed positively. But, it may not bring about the multidimensional aspects of coordination that is the hallmark for many of this type of an approach: bringing together a team of professionals with different backgrounds, addressing functional and psychosocial needs as well as medical conditions, and provision of care in a variety of settings.

Disease-Management Models. Commercial entities already exist that contract with managed care organizations to provide intensive care for individuals with particular health care problems. Often in exchange for a monthly fee, they agree to bear the risks for the costs of care in a particular area for the patients they enroll. Since they have been used more often on the younger population, they have tended to focus on treatments for patients who have only one problem area, such as asthma, for example. These entities also pose themselves as offering cost saving activities by using established protocols, and pushing for patient compliance, substituting nurses and other professionals for physician’s care. In theory these activities could reduce hospitalizations and other complications.

To be attractive in a fee-for-service environment, however, voluntary compliance would likely require that beneficiaries get additional services such as prescription drugs, help with home modifications, nutrition counseling or other valued benefits that are otherwise not included in the basic Medicare package. Although drug coverage would likely be limited to drugs for the particular medical problem, such as heart disease, for example, this could still be an expensive add-on to the costs that the disease management firm would have to absorb in its capitation payment.

Another unknown is how well such organizations would interact with patients’ physicians, both their primary care doctors and other specialists that they may see. A very strictly managed program might turn out to be essentially an HMO for beneficiaries who have only one major health problem. Efforts in this area thus could lead to the development of "specialty" HMOs.

Broader Care Management Activities. For patients with multiple conditions or who are very frail, a more intensive form of care coordination may be in order. Some of this is likely to come close to long term care activities, supporting patients at home and ensuring that their physical and social surroundings are supportive of the medical problems they face. As noted above, the PACE program offers an intensive model of this sort. PACE receives payments from both Medicare and Medicaid for the dual eligibles it serves, helping to link acute and long term care services. In theory, dual eligibles who do not wish to join a formal PACE program, or for whom PACE is not available, could still be good candidates for this type of care coordination, but a less intensive basis, likely using Medicaid home and community-based services to supplement Medicare.

But, when Medicare is the prime payer, are there care coordination activities that make sense (and are not inordinately expensive)? Coordination of care with a team managing acute and some home-based needs is likely what many envision as an ideal for helping persons with chronic care needs. This approach could allow control over care to be more patient-centered, although the organizational model, the degree of flexibility and how such a program is managed would determine the extent to which beneficiaries would have a greater say in how their care is delivered. Home care agencies could be the organizers of such systems since they already have delivery mechanisms in place. Medicare might be able to modify its prospective payment system to include some of the activities described here. But to work well, better coordination with physicians would likely be needed as well as the establishment of reasonable methods for overseeing the effectiveness of these programs.

Conclusion and Issues that Need Further Study

Many important questions remain unanswered about the coordination of care and what might work in fee for service. Some of these may be addressed by the demonstrations now underway. But other, more generic issues also need to be tackled. For example, should the focus be on the very sick where care management can have immediate effects, or should it begin earlier to prevent problems and to encourage people to embrace this type of model? If we expect consumers to play more of a role in their own care, how does that fit into the models of care coordination being discussed? Should such activities always be voluntary, for example? Should we make consumers responsible for electing such activities as is the case now with hospice? Should they have other responsibilities for participating like completing education training and demonstrating compliance as a condition of receiving subsidies for prescription drugs or other additional benefits?

Certainly the unit of payment matters in terms of the incentives that are established. But there is likely to be some controversy over the appropriate unit. For example, on a service by service basis, overseers of the program have a standard by which to judge whether care is being delivered. When services are "bundled" this creates the advantages of flexibility and encouraging coordination, but the disadvantages are that it is difficult to ascertain whether patients are getting "value" for the dollars being spent. When this issue arises for managed care, the usual answer is that competition among plans will keep them honest in providing good care. But it likely makes little sense to imagine competing operations in a local area in the same way unless individuals are willing to change doctors and make other arrangements as well.

Many analysts who advocate care coordination also stress the need to hold providers of all types accountable for providing high quality care. But as yet the payment system changes being contemplated are not sufficiently well developed to incorporate this important factor.

Resource needs exist on several levels and could be put into place in Medicare in a variety of ways (Wagner et al 1999, Oswald, Chen et al. 2000). These include (but may not be limited to):

C Support by patients to participate in and comply with recommendations. This requires time from professionals to foster and maintain patient efforts. It also means a full array of services need to be covered or largely covered. That is, the lack of prescription drug coverage serves as a major potential impediment for achieving compliance with many recommended treatements for the chronically ill.

C Sufficient reimbursement for key providers, including physicians, to take time with patients. The environment of keeping fees low has encouraged physicians to spend less time with patients; it will take considerable efforts to change the way that practice styles have evolved.

C Information resources on both recommended care plans and with detailed patient information. The goal needs to be a system that serves both patients and providers with reminders, information and feedback. These could come from private entities but some oversight of quality and appropriateness as well as support for putting this information into a practical form will likely be needed.

C Community resources and support. Since all care is local, it is important that the changes made to improve the delivery of care to the chronically ill have the support of the community and change the overall way in which care is delivered, not just for the Medicare population. Finally, information sharing among both the physician community and patients is likely to be important for acceptance of a new approach.

The areas where little experimentation is underway is at the physician level. In the likely upcoming debate over levels of payments to physicians, and with potentially billions of dollars at stake over the next ten years, it makes sense to use some of those resources to give physicians the opportunity to become more involved in care coordination. Particularly in rural areas, community-based systems could be developed for providing better information, but where the locus of control remains with the physician.

From an economics perspective, the key issue is how to get the incentives right to help foster the environment in which innovators are encouraged and supported. Just because the system is in place does not mean that "they will come." Indeed, most of the literature reviewed for this paper stressed the non-financial factors that were also important in motivating providers and consumers to seek new modes of delivering or receiving care.