Barbara S. Cooper
Institute for Medicare Practice
Mount Sinai School of Medicine

For

Medicare Coordinated Care Conference
March 21, 2002

INTRODUCTION

More than one quarter of community-dwelling Medicare beneficiaries, or over 17 million individuals, suffer from 3 or more chronic conditions and about 10 percent suffer from at least 5.  Some 9 million beneficiaries (not all of whom are in the first category) are unable to care for themselves in at least one significant way, such as feeding, bathing, toileting. [1]

These frail and vulnerable beneficiaries are not well served by our health care system -- they do not receive the care they need or much assistance obtaining the care they do receive; their treatment often fails to conform to evidence-based guidelines; appointments are frequently rushed with little follow-up; and insufficient time is devoted to assessing functional impairment, providing instruction on patient self-management, addressing emotional distress, or providing help navigating across multiple settings and multiple payers.  It is no wonder that care-management is on many people’s lips.

Ask any room full of health policy analysts whether to add a case-or care management benefit to the Medicare program and it is likely that every one of them will answer in the affirmative.  However, ask those same analysts what they mean by care management, and you are likely to get as many answers as there are analysts. Case-management, care management, disease state management, disease-specific case management, service coordination, patient care management, personal agent –  multiple terms with multiple meanings.  Say care-management to some, they immediately envision the possibility of limiting all that “unfettered” access to unnecessary services in fee-for-service.  Say it to others, they envision assisting beneficiaries to navigate through the maze of programs and services.  Before adding this new benefit, we must define what we mean.  What are its goals – gatekeeper or beneficiary advocate?  What services does it provide?  Who is eligible?  There are a host of questions that must be addressed and interactions among them considered in the design of such a benefit.  This paper raises some of these issues as a starting point for consideration and discussion.

PRINCIPAL GOAL

The first, and perhaps most critical, question that policymakers must answer is: What is the principal purpose of the care-management benefit -- to save Medicare money or maximize beneficiary quality of life?  Is it to serve as an advocate for the beneficiary, assuring that the beneficiary receives the services he or she needs; or to serve as a gatekeeper for the Medicare program, assuring that the beneficiary does not receive any unnecessary services and that the program saves money?  Both?  Unfortunately, there is no silver bullet here and it is unlikely that you can get both a patient advocate and program gatekeeper at the same time.   Can a gatekeeper reduce unnecessary care, improve care, and reduce hospitalizations and nursing home admissions?  Possibly.  But, if the care-manager is serving as a beneficiary advocate, a lot more services may be provided than the beneficiary would otherwise receive and these services may wipe out any of the savings that better management has achieved. 

Case-management received attention as early as the 1920s, as a means of addressing the multiple needs of people with chronic health problems.  Gradually, as concerns over rising costs of health care grew, the case manager began to be seen as an organizational and system agent asked to limit costs and inefficiencies.  In a Workshop on Making Case-Management Work in Government Programs,  A.E.”Ted” Benjamin, Ph.D, pointed out that,  “We began by defining case managers as advocates for clients.  And case managers have evolved into agents charged with becoming the bureaucratic ‘watchdog’.“[2]    According to participants in the Workshop, cost containment is case-management’s best friend and mortal enemy.  “The prospect of cost savings is one of the main reasons that case-management has drawn and continues to draw significant interest from a wide range of public and private policymakers, systems, and institutions.  But, improving access to services through case management or using case management as a mechanism for advocating for more adequate services will likely cost more, not less.” [3]

A study by researchers at RAND found that fee-for-service elderly Medicare beneficiaries experienced substantial underuse likely to result in negative outcomes.  Using Medicare claims, researchers examined utilization between 1994 and 1996 for 47 indicators of necessary care and found that for almost one-half of the indicators, less than two-thirds of beneficiaries received needed care. [4] RAND did not quantify the cost of providing all those missing services, but it clearly was not trivial.  If a care-manager is serving as the patient advocate, the patient may receive those needed services, but those services may very well cost more than any savings generated.

If the principal goal is to save money, or maintain budget neutrality, even that goal may be illusive as well, unless strong controls are imposed on use.  The research on cost-savings from care-management is equivocal at best[5]

Medicaid programs have had considerable experience with one model of care-management, primary care case management (PCCM), structured around basic care planning and gatekeeping.  Enrollees receive their services, or are required to obtain a referral, from their primary care provider.  But states have been moving away from use of PCCM programs in favor of capitated HMO alternatives “as the preferred strategy to not only improve access and accountability and reduce costs, but also achieve budget predictability.”[6]

The Balanced Budget Act of 1997 required The Centers for Medicare and Medicaid Services (CMS), formerly known as HCFA, to design a demonstration project for the Medicare fee-for-service population based on an evaluation of best practices in the private sector for methods of coordinated care.  The evaluation, performed by researchers at Mathematica, used six criteria for inclusion of a case-management program, but the first was that “programs must show quantitative evidence of reductions in hospital admissions, or hospital costs or total medical costs…”  [7] The search for programs was exhaustive – a Federal register solicitation, publication in several professional journals, hundreds of letters and e-mails, and extensive literature reviews.  Only 69 programs of 157 possible programs were eligible and the most common reasons for ineligibility by far was lack of impacts on hospital cost or use outcomes and lack of impacts on any outcome measures.  Mathematica was unable to ascertain eligibility of 236 programs.  Almost 2/3 of the eligible programs were disease-specific.  Even after completing their study of the eligible plans, Mathematica was “unable to address program cost-effectiveness.  Few programs had any actual data on their operating costs, and what data existed were of poor quality.” [8] Research on cost-effectiveness had to await the demonstration.

In July 2000, CMS solicited proposals for Medicare Coordinated Care Demonstrations.  CMS required that Medicare expenditures not exceed what would have been spent in the absence of the demonstration. Proposals were to  provide clear and convincing evidence that the care coordination services were “likely to achieve reductions in the use of medical services and likely to improve the quality of care for these individuals.” [9]  These demonstrations are currently underway. 

Advocate or gatekeeper?   In addressing that question, perhaps we should ask why cost-containment should be necessary for an intervention that has the potential to improve quality of care and quality of life.  Surely no such demand is made of medical or pharmacological interventions, many of which do not accomplish nearly as much as care-management.   If a pill improved quality of life as much as care-management can, would this even be an issue?  As Alan Weil eloquently stated in his discussion of care integration, “Rather than claiming value-neutral “cost-effectiveness,” we should claim that such a system is worth it …. It is the right thing to do, and it is worth new investments and even some financial risk to make it happen.”  [10]

BREADTH OF BENEFIT

Care management services can range anywhere on a continuum from simply reminding people to take their medicines to figuring out what care they need, helping them to get it, and making sure it is working out.   It can focus on one condition or on the person and all of his/her conditions.  It can focus just on Medicare services or it can address all of a person’s health care and social service needs as well.  Although it is not necessary to use the same model, or the same services within a model, for all eligible beneficiaries, a decision must be made as to the potential breadth of services to be included.  This decision will have different effects depending on other decisions on goals, payment methods, beneficiary eligibility, etc.  

The most common distinction is made between care-management and disease-management models.   Mathematica summarized the difference between the two types of models as “generalists”  -- dealing with a wide range of patients and problems -- and “specialists” –expert in one type of patient and their problems.

Disease-management programs are usually run by commercial entities that assume some share of the cost risk or contract for specific savings.  They target persons whose primary health problem is a specific disease, such as diabetes, although some other conditions may be addressed as well.  The patients generally do not have a high prevalence of difficult geriatric syndromes, such as incontinence, falling, cognitive impairments, or social problems such as inadequate family support, housing, or transportation.  The care coordination interventions tend to be highly structured, emphasize the use of standard protocols and clinical guidelines, and provide disease-specific education in self-management, lifestyle, diet, and medication compliance.

Care or case-management programs typically serve a select group of frail, disabled patients who suffer from severe illness, often multiple chronic health problems, and a high risk of recurrent adverse medical events.  These programs rely more on case manager judgment and individualized approaches than standardized protocols, and can involve community resources and social support services in planning care.

Mathematica found that some form of needs assessment, care planning and implementation, and reassessment is conducted in each model, but the intensity of those services varies enormously. The following indicates the broad categories of services or combinations of services that might be included.  

Potential services:

• Assessment of patient needs, barriers, and goals

• Development of plan of care

• Patient education about conditions(s) and self-care

• Medications reminder calls

• Nurse-advice hotline

• Advice on providers of care – Medicare-only, or all medical, or medical and social services

• Arranging for services – Medicare-only, or all medical, or medical and social

• Coordination with providers and programs (Medicare, Medicaid, Meals on Wheels, etc.)

• Monitoring and reassessment
  

CARE MANAGEMENT PROVIDERS

Who should be eligible to provide this new benefit?  Any primary care physician?  A nurse?  A multi-disciplinary-team?   Should there be a new category of certified Medicare provider with specific requirements to be met?  Although a number of disease management entities exist, there is not currently an extensive infrastructure of independent comprehensive care managers or care-management.    

Many advocates of care-management suggest that care-management works best if it is provided by an interdisciplinary team.   The best known example of multidisciplinary teams is found in PACE, capitated plans serving the frail elderly.  The hallmark of PACE is the interdisciplinary team, in which all persons who interact with the patients, even the van driver, meet daily to discuss every patient.   Regardless of whether or not the PACE approach represents the pinnacle of good care, the PACE kind of interdisciplinary team is extremely rare and it is not practical to require such a model for a fee-for-service benefit.   In fact, interdisciplinary teams in general are also rare so that it may not be practical to require any kind of team to deliver a care-management benefit.  The payment structure can encourage the development of teams, however, paying for team meetings, for example.

Should the care-manager, or leader, be required to be a primary care physician? On balance, physicians have not shown a great deal of interest in performing these functions, but in previous Medicare case-management and Social/HMO demonstrations, CMS found that failure to gain physician participation of some kind hindered the success of the demonstrations.[11]  Physicians may not need to be the care-managers, as long as they have some formal connection to the process.

In all case management programs in the Mathematica Best Practices study, case-managers were nurses with at least a bachelor’s degree in nursing.    In some cases, the nurses were part of the same practice or clinic staff as the primary care physician and formed permanent teams, physically located together and sharing patients; in some cases the nurses operated independently.  

Should Medicare require that the care-manager be a nurse, or permit either a physician or nurse to assume those responsibilities?   Can she be an independent nurse, working with a primary care physician, but billing separately or must the nurse be part of a physician practice and/or under the physician’s supervision.  If there is a certification requirement, who is certified – the nurse, the physician, or both?

BENEFICIARY ELIGIBILITY

Who should be eligible to receive this benefit?  The sickest beneficiaries?  The costliest?  Persons most in need of assistance?   Or, should care-management services be treated simply like any medical services, up to the provider to determine who and when?

In a disease-management model, eligibility is simple: it is a specific disease, sometimes at a specific severity level.  Eligibility for more comprehensive care-management benefits is more difficult to define and all of the above categories of beneficiaries could benefit in some way.  If eligibility is based on diagnosis, on what basis are the diagnoses chosen?  Should it be focused on high cost beneficiaries only or should a specified level of functional impairment trigger eligibility?  Eligibility criteria could also include specific demographic characteristics, such as poverty, living alone, or lack of an informal caregiver, where the beneficiary is in greater need of assistance.  Perhaps the dual eligible beneficiary – eligible for both Medicare and Medicaid services – represents the best candidate for care management.  Duals are poor, nearly 4 times more likely to suffer from 3 or more ADLs, less likely to have a usual source of primary care, and much more likely to suffer from chronic conditions such as diabetes, stroke, Alzheimer’s. [12]

Some legislative proposals offer a care-management benefit to beneficiaries with 2 or more deficiencies in Activities of Daily Living (ADLs) -- a large number of beneficiaries -- but information on ADLs is not routinely collected for the fee-for-service population.  The CMS Solicitation for the Medicare Coordinated Care Demonstration listed a number of specific conditions, such as congestive heart failure, diabetes, psychotic disorders, Alzheimer’s and stated that it would “give preference to proposals aimed at beneficiaries with one or more chronic conditions that represent high costs to the Medicare program.” [13]

PAYMENT

There are a variety of payment approaches that can be used and the one selected depends -- it depends on the goal, the type of program, the breadth of services provided, the connection with the primary care physician, and other factors.  [14]

Care-management can be billed service-by-service and a fee provided for each specific care management service, such as a team meeting, a conference call with family-members, development of a care plan, etc. Or, payment could be in the form of a bundled fee, similar to the all-inclusive rate paid for the prenatal care and delivery of a baby.  New codes would need to be developed and payment amounts calculated.

The CMS Demonstration, which requires voluntary monthly enrollment, specifies the use of a monthly administrative fee for each enrollee.  That fee, which can be case-mix adjusted, includes all the care management services and any extra benefits provided, but it must stay within the budget neutrality requirements.  In other words, the fee plus all the Medicare spending on behalf of the enrolled beneficiaries cannot exceed what spending for those beneficiaries would have been without care-management.   The Demonstration also permits consideration of alternative payment models in out-years, such as a financial incentives program paying a percentage of net Medicare savings -- a bonus -- in addition to a monthly fee.

Two different methods have been used by Medicaid plans to pay PCCMs.  Under one, PCCMs are paid a fixed fee of about $3 to $6 per member per month to manage primary care of individual beneficiaries; under the other, primary care physicians are rewarded with potential bonuses from inpatient cost savings. [15]

The Mathematica study reported that many hospital-based systems and academic medical centers indicate that they would not be interested in participating in a demonstration that, if successful, would result in a substantial loss in revenue from medical care.   In other words, if care-management reduces hospitalization, hospitals are not interested in being care managers without some offsetting revenue.

Disease management vendors, usually contracting with managed care plans, often participate in some form of risk-sharing.   In order to get started, disease management firms originally provided the entire infrastructure and were totally at risk for savings, but the amount of risk assumed now is usually less.  [16]  The medical director of HealthNet, a large managed care firm in Kansas City, observed that “risk sharing is a powerful tool to separate out disease management companies that really have confidence in their program.”  [17]  It should be noted that although risk-sharing or bonuses for savings are theoretically possible, the greater the share of risk assumed, the more control a disease-management or care-management entity will demand, control that may not be desirable or politically feasible.  The further along the continuum you move toward full-capitation, the more quality monitoring and controls are needed to guard against underuse.  There is also a potential for a comprehensive care-manager to try and shift responsibility from Medicare to other programs, promoting personal care services covered by Medicaid, for example, when Medicare skilled home care services would be more appropriate.  This, too, must somehow be addressed in program requirements and monitoring. 

Further, reaching agreement on the amount of savings (or costs) achieved, or even on the method for determining that, is not a trivial issue.

LOCUS OF CONTROL

Who is in charge of determining the care the beneficiary receives -- the care manager and/or physician or the beneficiary/family, and to what extent?  Is the care manager making all the decisions or is the care-manager acting as an adviser to the beneficiary or beneficiary’s family?  Is the program beneficiary-centered, in which decision-making draws upon an individual’s needs, values, and expectations, or is it provider-centered?   If the goal is to save money, or if it is disease-management only, the case-manager will need to have the major role; but otherwise, the beneficiary and family can be much more involved . “ I am not a case and I don’t want to be managed” is a refrain recounted by some persons involved in care management.  The degree of beneficiary and/or beneficiary’s family’s involvement is on a continuum from passive acceptance of the care-manager’s “suggestions” or edicts to total selection and payment of providers.  In many respects, it can be a function of the beneficiary’s desires and abilities.

In the 1970s, many younger persons with disabilities began to challenge the typical model of provider-centered and directed care, advocating for “consumer-directed” care.  Most of the movement and the literature about consumer-directed care centers on the personal care attendant benefit in Medicaid.  A number of Medicaid and state-funded programs have shifted the decision making, control, and autonomy for personal attendants from home care professionals and agencies to the consumer. Although consumer-direction for  personal attendants is obviously not relevant for Medicare, some of its tenets or principles giving greater control to the beneficiary, are.

Medicare beneficiaries include younger persons who have experienced incapacitating illnesses or injuries and older persons suffering from conditions such as strokes, hip fractures, congestive heart failure.  Most of the consumers in the Medicaid consumer-directed programs have been younger persons with disabilities but recent evidence suggests that a large minority of older persons are interested too. [18]  Like others, they prefer to have a say in what is done, when, by whom, and how.

The degree to which consumer preferences are incorporated in a care-management model is a critical question.  “To the extent that care subordinates or suppresses autonomy, its benefits come at a dubiously high cost of human individuality and freedom.” [19]

BENEFICIARY PARTICIPATION

What if an eligible beneficiary does not want a care-manager?  Is it totally voluntary?  Or, what if a care-manager does not want a particular beneficiary, or does not want to retain a beneficiary who fails to comply with protocols?  It is hard to imagine forcing, or wanting to force, fee-for-service Medicare beneficiaries to participate in care-management.  The CMS demonstration solicitation required voluntary enrollment.  It is also hard to imagine forcibly disenrolling a beneficiary. 

How beneficiary participation is handled depends in part on whether savings are a principal goal, on how the case-manager is paid, and whether financial risk and incentives are built into payment.  If paid on a service-by-service basis, care-management can be offered strictly as a service, or simply incorporated in some form into the practices of primary care physicians.  If payment is on an enrollee per month basis, then some form of voluntary enrollment may be necessary and incentives to join, such as free transportation or other additional benefits, may need to be offered. 

In the early 1990s, HCFA conducted a Medicare Case Management Demonstration in which 70 to 80 percent of all eligible patients declined to participate.  The Mathematica study of Best Practices did not directly address the issue of patient refusal, but it did refer to a disease management program that advocated a “passive enrollment” approach, in which patients are informed that they are automatically enrolled unless they specifically refuse.  Passive enrollment achieved a 98 percent participation rate; active enrollment achieved only a 25 percent participation rate.   In many respects, passive enrollment is just a softer form of forced enrollment, particularly in the case of Medicare beneficiaries.

A beneficiary’s willingness to participate will likely depend upon how closely aligned the case-management function is to the primary care physician,  the extent of “gatekeeping”, any extra benefits provided, and, of course, how well it is done.  It is likely that the stricter the requirements, such as preauthorization, the greater the extra benefits that will have to be offered to induce participation.

PROMOTING QUALITY OF CARE 

A separate section on quality of care is designated here because it is too critical not to be accorded one, but it remains for others to address. 

CONCLUSION

Adding a care-management benefit to Medicare sounds to most of us like a good idea.  But until we answer some critical first-order questions like what we mean by care-management, who it is intended to help, and how it is designed and paid, it will remain just that – a good idea.  How we answer these and other questions can change a good idea into a great one.  But how we answer these questions can also change a good idea into a great flop.