Bruce Leff, MD

	Using literature review and interviews, we have identified 8 attributes of programs, such as accountable care organizations, readmission initiatives, special needs plans, care transition programs, and patient-centered medical homes, that successfully treat high-need, high-cost patients. These 8 attributes—illustrated here with specific examples&mdash;are specific ways to target these types of individuals, promote leadership at various levels, emphasize interaction with the care coordinator, use data strategically to refine the program, update the program periodically, allow physicians to spend more time with patients, and promote interaction among clinicians and high-need, high cost patients and their families. 

	Many of the attempts to design programs for high-need, high-cost patients have been unable to achieve even 1 of the triple aims. In this article, we examine the published literature, government reports, and reports by managed care organizations, hospitals, and physician groups to identify the commonalities of successful programs. The 8 attributes of successful programs we identified are: appropriate targeting of high-risk patients, inclusive leadership, appropriate sizing of program, adapting to local conditions, providing helpful feedback to clinicians, promoting interaction with patients and family members, fostering successful care transitions, and reducing the workload of clinicians. This article will allow policy makers to: 

		Compare their programs to the attributes of successful programs. 

		Assess if modifications to their programs are needed. 

		Determine if they should partner with certain existing programs. 

		Design the next generation of programs.

The United States has tried, and continues to experiment with, many different programs designed to provide better care for people with high needs and high costs. These programs include accountable care organizations, readmission initiatives, special needs plans, care transition programs, patient-centered medical homes, and many others. These programs have been evolving over the past 20 years with many of the initiatives based loosely on the Chronic Care Model developed by Ed Wagner and his colleagues.

 We wanted to identify specific attributes of successful programs for this population.

In order to identify the attributes that successful programs have in common, we began by reviewing the peer-reviewed literature evaluating the various programs and subsequently identified over 1000 articles that were written about programs designed to treat adults with high costs and/or high needs in the United States. To be included in the review, the study had to involve 2 or more chronic conditions or a chronic condition and a disability.

However, we quickly realized that many programs being used to guide policy are not in the peer-reviewed literature. These include programs sponsored by the Medicare and Medicaid, Veterans Administration, private insurers or health plans, and doctors or hospitals. We used the information from both the peer-reviewed literature and the gray literature to: 1) identify specific programs that focus on high-need, high-cost individuals that were successful on at least 1 of the 3 triple aims (spending, satisfaction, clinical outcomes), and 2) to identify commonalities of success in these programs.

We became especially interested in why certain programs had varying levels of success in different settings and populations. We created a semi-structured interview guide and interviewed 45 programs that the peer-reviewed literature and/or gray literature identified as successful on at least 1 of the triple aims.

Based on the literature review and the interviews, we identified 8 attributes that many successful programs shared. An example of each specific program that typifies each attribute is provided in the 

Many successful programs use 2 steps to identify which individuals to include in the program and who should receive special attention. Many unsuccessful programs enroll a high percentage of people who are unlikely to benefit from the intervention; for example, the enrollment of people with only a single chronic disease.

 Second, some programs provide the same services to all people enrolled in the program; however, the more successful programs typically stratify program enrollees and provide additional services to certain high-need, high-cost individuals.

Nevertheless, simply being high-need/high-cost is not always sufficient.

 Many patients have conditions that are not amenable to change. For example, some cancer patients may require expensive drugs and the drug use is appropriate for their condition. An intervention will not necessarily lower the spending for that person. Other patients are unwilling to comply with treatment protocols. Being able to identify those who are most likely to benefit from an intervention is an understudied area and may be responsible for the lack of success in many programs. The key is to identify those patients most amenable to change.

Most successful programs have leadership at varying levels of the organization, not simply at the top. It is important that the leadership team span a wide range of clinical and operational disciplines, including physicians, nurses, social workers, pharmacy, and operations. Finding successful care managers and facilitating their relationship with other clinical practitioners and non-clinical professionals, such as community health workers, is a critical role for leadership. The care managers must have the skills to be assertive when necessary, to understand the practice culture in the setting, and to maintain good relationships with physicians, other clinicians, and administrators caring for the patient.

 With this in mind, some programs have developed their own training programs for care managers.

Some programs involve teams of nurses and social workers working remotely, primarily using the telephone or e-mail. Although some of these programs are successful, most are not, probably because they have not developed close relationships with the clinicians treating the patient, the patient and the family members.

 Although the initial expense is greater, successful programs are more likely to have care coordinators located in the physician’s office. Physical proximity is only a first step, however; it may be necessary for all clinicians, including specialists, to personally know each other. Developing an ongoing working relationship is important, and often this requires frequent face-to-face interactions about a specific patient, especially with specialty physicians.

 It was often more difficult to develop these personal relationships in large fee-for-service organizations, like academic medical centers or large hospitals, where there were a multitude of people treating the patient

; it was much easier in smaller multidisciplinary physician groups, where a care coordinator was embedded in the office, as well as in larger organizations like Kaiser Permanente, where teams of clinicians routinely worked together as a group.

Both electronic medical records (EMRs) and claims data have been utilized successfully in the treatment of high-need, high-cost individuals. The advantage of relying on EMRs is that the data is available in real time, which is often when the information is most actionable. However, EMRs are often not interoperable and, therefore, it is difficult to assemble a complete picture of the patient if the patient is receiving care at multiple sites. Claims data is often more complete, and the Medicare program has begun to provide claims data to programs on a monthly basis.

Providing timely data to the care team is vital.

 However, it is unclear how timely the data needs to be or exactly which data elements are the most critical. Many clinicians do not respond to cost information about individual patients or know how to respond to information about readmission rates. In addition, finding relevant data for high-need, high-cost patients is problematic because determining the appropriate care for them is challenging—most quality metrics focus on a single disease, not the ability to coordinate care for complex patients.

 Data customization for specific types of patients also presents a challenge. For example, people near the end of their life require a different data set from those with psychological or social issues, and a different data set from those undergoing a transition from one setting to another.

Successful programs seem to be more facile at interacting with patients and their families.

 In most successful programs, the care coordinator is located in the doctor’s office and has a wide range of clinical and social service responsibilities.

 They often: 1) see the patient when he or she arrives to see the doctor and ask about their priorities for care, 2) make periodic home visits, 3) occasionally interact with the patient's family, and 4) to some extent, interact with the specialists caring for the patient. Interaction with patients is especially critical during and following the hospital stay. A patient’s medications can change following physician and hospital visits, so it is often necessary for the care coordinator to explain the change in medications directly to the patient.

Also important are ongoing interactions with the social service programs since many high-need, high-cost high patients have disabilities that make it difficult for them to undertake routine activities of daily living. Both nurses and social workers seem to be successful care coordinators. In general, more in-person interaction between the patient and the care coordinator results in better outcomes and lower overall spending.

 Remote monitoring programs are less likely to achieve the triple aim.

 A paper synthesizing the evaluations of Medicare programs that treat the high-need, high-cost population found that 1 contact per month with the care coordinator is the minimum necessary for a successful program.

Most successful programs evolve over time.

 The challenge here is periodically updating the program to reflect new information about what is working, and responding adequately to the changing mix of high-need, high-cost individuals participating in the program. It appears to be important for care managers and clinicians to have enough flexibility to initiate program innovations at certain internals, but there does not seem to be a standard timeline for updating the program. It is clear that many programs treating high-need, high-cost individuals require a long period of sustained funding before the program becomes cost-saving.

There are a multitude of programs with different financial incentives for physicians. Many of them provide rather modest financial incentives for the physicians and may represent a relatively small portion of the physicians’ total patients. As a result, changing the financial incentives is frequently an unsuccessful strategy. Instead of dollars, an appeal to the physician’s work-life balance is often more effective.

 Care coordinators are able to simplify the work of clinicians by facilitating the interaction with other physicians treating the patient, meeting with family members, explaining changes to treatment protocols with patients, interacting with social service agencies to make sure the patient has necessary services, and many other roles that the physicians frequently perform. This allows the physician to spend more time with their patients and to perhaps leave the office earlier in the day.

Governments, health plans, and clinicians will soon be developing the next generation of programs designed to treat high-need, high-cost individuals. The reasons are clear: they are the most costly individuals who often have the worst clinical outcomes. The problem is, most of the interventions that have been attempted have not yet been able to achieve 1 or more of the triple aims, and many of those that have been reported in the peer-reviewed literature as being successful in this were not operational a few years later. The current evidence suggests that programs that share certain attributes are more likely to be successful and sustainable. These attributes are shown in the Table. These attributes are meant as guidelines for governments, health plans, and clinicians to consider as they develop the next generation of programs. Whereas all are important, perhaps the most important is identifying the patients most likely to benefit from the program and then targeting the program around these individuals.

 Johns Hopkins Bloomberg School of Public Health (GFA, JB, SB, CS, JW), Baltimore, MD; Johns Hopkins School of Medicine (CB, BL), Baltimore, MD; University of Wisconsin — Madison (ED), Madison, WI.

 Dr Leff is member of the board for American Academy Home Care Medicine and a member of the advisory board for Landmark Health. Dr Boyd received royalties as author of book chapter on multimorbidity from UpToDate. The remaining authors report no relationship or financial interest with any entity that would pose a conflict of interest with the subject matter of this article.

 Concept and design (GFA, CS, BL, JW); acquisition of data (GFA, JB, ED, BL, CS); analysis and interpretation of data (GFA, CB, JB, SB, BL, CS, JW); drafting of the manuscript (GFA, JB, SB, ED, CS); critical revision of the manuscript for important intellectual content (GFA, CB, SB, ED, BL, JW); statistical analysis (GFA); provision of patients or study materials (GFA); obtaining funding (GFA); administrative, technical, or logistic support (GFA, JB, ED); and supervision (GFA).

Gerard Anderson, PhD, Department of Health Policy and Management, Bloomberg School of Public Health, Johns Hopkins University, 624 North Broadway, Baltimore, MD 21205. E-mail: Ganderson@jhu.edu. 

1. Wagner EH, Austin BT, Von Korff M. Organizing care for patients with chronic illness. 

2. Bott DM, Kapp MC, Johnson LB, Magno LM. Disease management for chronically ill beneficiaries in traditional Medicare. 

3. Boult C, Green AF, Boult LB, Pacala JT, Snyder C, Leff B. Successful models of comprehensive care for older adults with chronic conditions: evidence for the Institute of Medicine’s “retooling for an aging America” report. 

4. Brown RS, Peikes D, Peterson G, Schore J, Razafindrakoto CM. Six features of Medicare coordinated care demonstration programs that cut hospital admissions of high-risk patients. 

5. Hong CS, Siegel AL, Ferris TG. Caring for high-need, high-cost patients: what makes for a successful care management program? 

6. Wolff JL, Rand-Giovannetti E, Palmer S, et al. Caregiving and chronic care: the guided care program for families and friends. 

7. Boult C, Karm L, Groves C. Improving chronic care: the “guided care” model. 

8. McCall N, Cromwell J. Results of the Medicare Health Support disease-management pilot program. 

9. Boyd CM, Leff B, Wolff JL, et al. Informing clinical practice guideline development and implementation: prevalence of coexisting conditions among adults with coronary heart disease. 

10. Wolff JL, Giovannetti ER, Boyd CM, et al. Effects of guided care on family caregivers. 

11. Naylor MD, Brooten DA, Campbell RL, Maislin G, McCauley KM, Schwartz JS. Transitional care of older adults hospitalized with heart failure: a randomized, controlled trial. 

12. Boult C, Reider L, Leff B, et al. The effect of guided care teams on the use of health services: results from a cluster-randomized controlled trial. 

Articles

Many programs attempting to effectively treat high-need, high-cost individuals have not been able to lower spending, improve outcomes, or increase
satisfaction. This paper suggests 8 attributes that many successful programs share.

Attributes Common to Programs That Successfully Treat High-Need, High-Cost Individuals

About a third of Medicare beneficiaries are covered by Medicare Advantage (MA) plans or accountable care organizations (ACOs). As a result of assuming financial risk for Medicare services and/or being eligible for shared savings, these organizations have an incentive to adopt models of delivering care that contribute to better care, improved health outcomes, and lower cost. This paper identifies innovative care models across the care continuum for high-cost Medicare beneficiaries that MA plans and ACOs could adopt to improve care while potentially achieving savings. It suggests policy changes that would accelerate testing and spread of promising care delivery model innovations.

Targeted review of the literature to identify care delivery models focused on high-cost or high-risk Medicare beneficiaries.â€¨

This paper presents select delivery models for high-risk Medicare beneficiaries across the care continuum that show promise of yielding better care at lower cost that could be considered for adoption by MA plans and ACOs. Common to these models are elements of the Wagner Chronic Care Model, including practice redesign to incorporate a team approach to care, the inclusion of nonmedical personnel, efforts to promote patient engagement, supporting provider education on innovations,and information systems allowing feedback of information to providers.The goal of these models is to slow the progression to long-term care, reduce health risks, and minimize adverse health impacts, all while achieving savings.These models attempt to maintain the ability of high-risk individuals to live in the home or a community-based setting, thereby avoiding costly institutional care. Identifying and implementing promising care delivery models will become increasingly important in launching successful population health initiatives.

MA plans and ACOs stand to benefit financially from adopting care delivery models for high-risk Medicare beneficia- ries that reduce hospitalization. Spreading these models to other organizations will require provider payment policy changes. Integration of acute and long-term care would further spur adoption of effective strategies for reducing or delaying entry into long-term institutional care.

This paper presents 9 delivery models for high-risk Medicare beneficiaries in different settings that Medicare Advantage (MA) plans and accountable care orga- nizations (ACOs) could adopt to achieve better care at lower costs.â€¨ Common to these models is a team approach to care, including nonmedical personnel, and providing care in home- or community-based settings to avoid costly institutional care.

Our review of the literature included intervention models such as Advanced Primary Care, Home-Based Primary Care, CAPABLE, MIND at Home, PACE, Hospi- tal at Home, Hospital Elder Life Program, Transitional Care, and INTERACT. Policy changes to permit MA plans, ACOs, and service providers to share Medicaid long- term care savings would further accelerate adoption.

About a third of Medicare beneficiaries are now covered by Medicare Advantage (MA) plans or accountablecareorganizations (ACOs).

 These organizations have an incentive to adopt innovations in care delivery that yield better care, improve patient outcomes, and lower costs.MA plans are at financial risk for the total cost of Medicare services, and Medicare ACOs are either at full or partial risk, or eligible for shared savings. Delivering care in a way that reduces the costly use of inpatient hospital care and emergency departments (EDs) helps these organizations to realize savings and gain market share.

Factors that place Medicare beneficiaries at risk for higher acute, post acute, and long-term care utilization and expen- ditures include age, number and type of chronic conditions, functional impairment, income level, and social support system (eg, living alone). Those Medicare beneficiaries with multiple chronic conditions, functional impairment, and low income, and/or who are living alone or in institutions, account for most Medicare expenditures and high rates of ED visits, hospitalizations, readmissions, and nursing home placements.

 These high-risk beneficiaries also tend to spend the most—5% of Medicare’s beneficiaries account for more than 40% of the costs.

Despite the potential for savings, adoption of innovative care delivery models focused on the highest cost, highest risk patients is limited—in part, this reflects the need for large- scale testing of innovations in care delivery, as evidence is limited on what works and why. The Center for Medicare and Medicaid Innovation (CMMI) is starting to fill this gap with $10 billion in funding to experiment with innovative payment models and improved care delivery systems, as part of the Affordable Care Act.

CMMI has launched a number of initiatives, with the primary focus on ACOs, bundled payments for care improvement, and primary care transformation.

These initiatives have helped spur the development of ACOs, health systems, and advanced primary care practices. The first evaluation results for Pioneer ACOs indicate some modest success in improving care and lowering costs.

Better targeting of high-risk beneficiaries and the use of models that manage care across the continuum could yield greater savings. For the high-risk individuals who transition frequently across settings of care, to date less emphasis has been placed on specific tools or care models, which could enhance quality of care and patient experience, as well as to reduce cost (

The largest obstacle to the diffusion of innovative delivery models that integrate care is a lack of aligned financial incentives and the presence of a fragmented payment system across different providers. MA plans often continue to pay providers on a fee-for-service basis, giving providers little incentive to provide lower-cost, higher-quality care. The closer integration of the insurance and provider functions, however, is starting to align provider incentives. Integrated delivery systems with MA plans or managed care at-risk provider contracts have the incentive to identify methods of providing better care at lower costs. This has spurred peer-to-peer learning and the creation of a Medicare Innovations Collaborative to provide technical assistance to promote the simultaneous adoption of multiple complex-care models.

Collaborations will continue to grow as Medicare and other payers begin paying for total care for a population over time or over an episode, with accountability for quality and patient health outcomes.

Identifying interventions that prevent long-term nursing home admissions could yield significant savings, as non-alignment of incentives across providers, families, and social service agencies has been a barrier to the diffusion of models of care that reduce these admissions. Medicaid is responsible for many long-term care costs, adding more complexity in designing incentive strategies to reduce long-term care admissions. 

This paper examines how integrated delivery systems, health systems, and MA plans bearing financial risk could benefit from adopting specific care delivery models across the care continuum that show promise of better results for high-risk Medicare beneficiaries. It also explores policy changes to create a global payment system, which is more amenable to implementing new care delivery models. 

Innovations for High-Risk Medicare Populations 

The budgetary pressure to bend the cost curve for Medicare beneficiaries is likely to persist or accelerate as those born after World War II become eligible for benefits. Since the majority of spending is concentrated among a minority of beneficiaries, the importance of appropriately targeting innovation initiatives on high-cost, high-risk subpopulations will increase. The diversity of beneficiaries by health status, functioning, living arrangements, and income suggests that rather than focus on a uniform program with uniform benefits, Medicare might look to adapt benefits and care delivery to the specific needs of Medicare beneficiary sub-groups based on their health risks. As the population ages, with unprecedented growth in the oldest cohort, considering the risks of long-term care as well as acute care grows in importance.

describes characteristics of potential high-risk beneficiaries based on the authors’ estimates from the Health and Retirement Study for 2010.

Beneficiaries eligible for both Medicare and Medicaid are 4 times more likely to have 6 or more chronic conditions than beneficiaries with incomes 200% or more of the federal poverty level not covered by Medicaid, and they are 3 times more likely to have 2 or more restrictions in activities of daily living. Dual eligibles are more likely to live alone and more likely to be disabled than higher-income Medicare beneficiaries with incomes 200% or more of the federal poverty level. Those Medicare beneficiaries with incomes below 200% of the poverty level who are not covered by Medicaid are also more at risk than higher-income beneficiaries, with a greater likelihood of having multiple chronic conditions and functional impairments, as well as to be living alone. 

To illustrate how strategies that target high-risk groups that are designed to achieve savings could work, we offer promising care delivery innovations as examples here. Interventions showing modest success have generally tried to improve coordination among the patient, family members, providers, and even social service agencies. All of the select innovations incorporate the major elements of the Wagner Chronic Care Model, including practice redesign, patient engagement, provider support and education, and information systems designed to furnish feedback of information to providers.

The Wagner Chronic Care Model is applicable to this high-risk Medicare population because most members have chronic conditions in addition to functional limitations. The model stresses the need for patients to be actively involved in their care, as well as practice redesign, ongoing provider education, and meaningful use of information systems. Successful models that are able to demonstrate savings will likely have some aspects of all 4 elements. 

										All patients benefit from better care from providers, but high-risk patients have the added challenge of moving across multiple care settings, from primary care to possibly the hospital and nursing home.

For the high-risk elderly, the care continuum starts with normal functioning in the home, with interventions needed in the primary care setting to maintain function. As the course of disease and aging continues, the elderly may need more support to stave off transitioning to nursing homes, in addition to coordinating care across multiple providers. For all elderly, reducing the number and severity of hospitalizations is crucial, and some interventions target this particular care setting. Lastly, high-risk elderly in nursing homes have their own set of needs to minimize transitions in and out of acute care settings.

The illustrative care delivery model innovations examined here were selected because: 1) they have some evidence of improved care, better patient care experiences, and outcomes with lower cost; 2) they incorporate elements of the Wagner model; and 3) they encompass the full continuum of care (eg, primary care, home or community care, inpatient hospital care, transitional care from inpatient to post acute care, and long-term care). 

 provides a brief description of each model, its target population, and promising results from at least 1 evaluative study. The models selected point to the policy changes that would be required to create incentives for adoption. The 

) describes the models in more detail and highlights particular barriers to wider implementation. 

Barriers to Spreading Innovations for High-Risk Elderly

summarizes major obstacles to the diffusion of innovative models of care for high-risk older adults. These include the fact that savings associated with the delivery of more efficient care may accrue to Medicaid, families, or providers/parties other than those incurring the cost of the innovation, thereby diluting the return to the Medicare program and adopting organizations. To realize a return, eligible beneficiaries or organizations providing services may need to meet specific requirements to ensure likelihood of yielding the anticipated savings or outcomes. This might entail screening costs or administrative burdens on Medicare and participants, and may increase the risk of fraud or system gaming. Finally, upfront assistance with capital costs or the costs of training qualified personnel, or reinsurance against high-cost cases, may be required.

Barriers to adoption are particularly onerous in the long-term care setting. Incentives are not aligned across acute and long-term care sectors. Hospitals that reduce institutional care—for example, through reduction of delirium in at-risk hospitalization patients&mdash;incur added staffing costs, yet do not benefit from Medicaid long-term care savings that may result. Nursing homes do not benefit from initiatives that reduce hospitalization and, in fact, typically lose Medicare revenues for post acute care when residents are not hospitalized for acute illnesses. Further, the added costs of nursing home staffing to prevent or manage conditions typically requiring hospitalization are not compensated. 

Payment approaches that stimulate integration of care across acute and long-term care services through sharing savings need further policy development. This is a particularly important issue for the dual eligible population. While not all at-risk Medicare beneficiaries are covered by Medicaid, lower-income Medicare beneficiaries with functional limitations are likely to eventually exhaust their financial resources and qualify for Medicaid. Interventions that reduce the need for long-term institutional care would not only benefit beneficiaries and their families, but ultimately lower Medicaid costs. 

A key determinant of success will be designing payment models that show promise of achieving better out- comes and lower acute and long-term care costs. This could be the evolution of ACOs and MA plans to take financial risk for long-term care costs as well as Medicare services, or more modest steps such as Medicaid and/or Medicare payment of care management fees for services in the home and shared long-term care savings.

These examples of innovative delivery models have important implications for how Medicare and Medicaid payment models would need to change to capitalize on these promising innovations on a large scale. In all models, MA plans, integrated delivery systems, and ACOs are the most likely to benefit from shared savings initiatives due to aligned financial incentives that would accrue with reduced use of expensive Medicare services. However, even these organizations do not benefit from reduced Medicaid long-term care costs.

summarizes the target populations and the services not currently covered by Medicare that are crucial to the delivery model innovations. In most cases, the noncovered services that are key to the delivery models are those provided by social workers or nurse case managers. Table 4 also highlights the source of total health and long- term care savings, which for high-risk beneficiaries mostly accrue from reduced hospitalizations and readmissions, as well as from reduced nursing home admissions. Additionally, Table 4 distills the payment changes needed to further align incentives and encourage adoption and spread of the delivery models, for fee-for-service providers, MA plans and provider organizations at financial risk for cost of care or sharing in savings, and Medicaid managed care plans. 

The most obvious source of savings for MA plans and ACOs come from better management of high-risk beneficiaries in their homes or in nursing homes to avoid ED use and hospitalization. Models such as Hospital at Home and INTERACT should be especially attractive to MA plans and ACOs, as they effectively reduce costly hospitalization of Medicare patients. However, mounting these interventions in patients’ homes or in nursing homes requires upfront costs—in particular, the cost of providing services in patients’ homes, or training and supervising nurse aides in nursing homes, are not now typically incurred by MA plans and ACOs. Testing and refining effective models of care on a smaller scale and closely monitoring their impact may be needed to overcome this obstacle to implementation. 

Innovative care delivery models for Medicare beneficiaries with physical and cognitive functional limitations show substantial promise of improving quality of life for these beneficiaries and their caregivers, as well as for reducing costs—especially costly long-term nursing home placements. MA plans and ACOs looking for care delivery innovations that could reduce cost of hospitalization could especially benefit from implementing the Advanced Primary Care, Independence at Home, Home-based Primary Care, Hospital at Home, Transitional Care, and INTERACT models of care delivery&mdash;all of which have potential to reduce hospitalization or readmissions. Other models such as CAPABLE, MIND at Home, PACE, show promise of helping at-risk Medicare beneficiaries continue to live independently at home by reducing or delaying nursing home placement; this has the potential to yield Medicaid savings. Medicaid managed care plans or those financially responsible for dual eligibles could well benefit from their adoption.

This paper illustrates how Medicare Advantage plans and accountable care organizations could benefit from adopting innovative care delivery models, and suggests policy changes to accelerate spread.

Innovative Care Models for High-Cost Medicare Beneficiaries: Delivery System and Payment Reform to Accelerate Adoption

Improving quality of care for people with multiple chronic conditions (MCCs) requires performance measures reflecting the heterogeneity and scope of their care. Since most existing measures are disease specific, performance measures must be refined and new measures must be developed to address the complexity of care for those with MCCs.

 To describe development of the Performance Measurement for People with Multiple Chronic Conditions (PM-MCC) conceptual model.

 Framework development and a national stakeholder panel.

 We used reviews of existing conceptual frameworks of performance measurement, review of the literature on MCCs, input from experts in the multistakeholder Steering Committee, and public comment.

 The resulting model centers on the patient and family goals and preferences for care in the context of multiple care sites and providers, the type of care they are receiving, and the national priority domains for healthcare quality measurement.

: This model organizes measures into a comprehensive framework and identifies areas where measures are lacking. In this context, performance measures can be prioritized and implemented at different levels, in the context of patients’ overall healthcare needs.

Improving the quality of care for people with multiple chronic conditions (MCCs) requires performance measures reflecting the heterogeneity of, and scope of care required by, these people.

The Performance Measurement for People with Multiple Chronic Conditions conceptualmodel is designed to facilitate development and refinement of performance measures to address the complexity and dynamic nature of care for those with MCCs.

The model centers on patient and family goals and preferences for care in the context of multiple care sites and providers, the type of care they are receiving, and the national priority domains for healthcare quality measurement.

One-fourth of Americans have 2 or more chronic conditions, yet this population accounts for more than half of overall healthcare expenditures.

 Having multiple chronic conditions (MCCs) can negatively affect quality of life,

Despite the high prevalence of MCCs and corresponding negative consequences, care for people with MCCs is often fragmented, incomplete, inefficient, and ineffective,

 resulting in potentially avoidable inpatient admissions,

 adverse  consequences of therapeutic interactions,

 and postoperative complications. The Department of Health and Human Services has identified “fostering healthcare and public health system changes to improve the health of individuals with multiple chronic conditions” as one of 4 goals in an overall strategic framework to improve the health of individuals with MCCs.

 Various  agencies, including the Centers for Medicare & Medicaid Innovation Center, Centers for Disease Control and Prevention, and     Administration on Aging, are funding projects centered around healthcare quality improvement for individuals with MCCs.  However, existing healthcare quality measures used in many of these projects are inadequate for measuring quality improvement for this population.

Existing quality performance measures are often limited by a disease-specific focus,

 do not account for patient and family  preferences and goals,

and often focus on a single setting such as hospitals. Performance measurement has been largely guided by a single disease heuristic that does not address challenges common to MCCs such as disease interactions and treatment interactions, and little evidence exists to support development or adaptation of performance measures for people with MCCs.

 To improve quality and efficiency of care for people with MCCs, performance measures need to address the  heterogeneity and scope of care, individuals’ priorities and care preferences, the high risk of uncoordinated care across settings, and the high risk for adverse health outcomes in the presence of comorbidities. Currently, few measures meet these criteria, and those that exist require further development.

There is a need to develop new performance measures and refine existing measures to address the complexity of care experienced by patients with MCCs and their families over time. Such development efforts would be facilitated by an appropriate conceptual model. Existing conceptual models of quality of care address some of these issues but do not address the broader perspective of people with MCCs.

 To help advance development of new performance measures and implementation of existing measures in this area, the Department of Health and Human Services requested that the National Quality Forum convene a multistakeholder steering committee to develop a measurement framework for individuals with MCCs. As part of this effort we developed the Performance Measurement for People with Multiple Chronic Conditions (PM-MCC) conceptual model based on reviews of the literature and existing conceptual models of performance measurement,

 input from experts on the National Quality Forum Steering Committee, and input through an open public comment period.

This conceptual model centers on a patient with multiple conditions, represented by overlapping circles (see center circle of the 

). Included are traditional diseases, but also conditions such as symptoms, disability, substance abuse, and hearing impairment that fall outside the traditional disease model.Any given condition may affect the patient to a greater or lesser magnitude at any one time, and may or may not be a dominant condition (ie, a condition so complex or serious that it eclipses the management of other conditions

). The patient and the family or friends who care for him/her have goals and preferences   for care of these diseases and conditions. Performance measurement should center on these goals and preferences.

The first circle surrounding the individual at the center of the model represents shifting sites and providers that support and care for the individual’s healthcare needs. These sites could include the ambulatory, hospital, postacute, nursing home, community, home (including formal home-based primary and skilled nursing care as well as informal care), and pharmacy settings. This list is not exhaustive; rather, it is intended to be illustrative of the possible sites of care for people with MCCs. At any given site, multiple types of providers may be providing care over time.

Moving outward from the sites and providers circle, there is a circle representing the types of care an individual may receive at any given site of care. The types of care patients receive (eg, screening, prevention, diagnosis, treatment, management, secondary prevention, community services, management of an acute exacerbation, rehabilitation, palliation, end-of-life care) are not necessarily linear or mutually exclusive. For example, a patient with congestive heart failure may be seen in the hospital for an acute exacerbation while receiving ongoing care for diabetes and depression at the same time. 

The outermost circle represents the domains of measurement that apply across sites and types of care. These domains are not mutually exclusive, and a given measure could fall into multiple domains; however, measures can be categorized under the 6 priority areas of the National Quality Strategy for improving healthcare: (1) health and well-being, (2) effective prevention and  treatment, (3) person- and family-centered care, (4) patient safety, (5) effective communication and care coordination, and (6) affordable care (Figure). These domains intentionally align with the  National Quality Strategy to promote harmonization across public- and private-sector programs supporting this population and to provide a way to track progress n filling of critical measure gaps.  In each measurement domain there are non—disease-specific measures such as pain screening (which would apply to all patients within a certain opulation regardless of their condition)  and disease-specific measures such as management of antidepressant medications (which would apply to patients with certain conditions). 

These circles, described above and shown in the Figure, re intended to be seen as a set of interlocking wheels that  rotate over time. At any given point in time, an individual may receive 1 or  more types of care in 1 or more settings from multiple providers. At the same time, the domains of performance measurement most important for that patient, site, or type of care vary. Some measures may be specific to a site and type of care, while others (eg, care coordination) apply across sites and types of care.

Each measurement domain can be measured over a period of time, which may be labeled as an episode. However, the PM-MCC model builds on the original National Quality Forum—endorsed patient-focused episodes of care framework by recognizing that distinct beginnings and ends to episodes of care are difficult to define and may need to be defined more arbitrarily in people with MCCs. Additionally, the elements of care highlighted on this model are not meant to be of equal importance. The amplification or minimization of any item in the model will vary with the individual and across time. The PM-MCC model places measures within the larger context of patient’s multiple conditions, sites of care, and types of care. We propose that in this context measurement  can be applied  and prioritized at the level of the individual, practice, site of care, health plan, or population.

Finally, the PM-MCC model falls within a social and environment context, as well as a public and private health policy context. These contextual factors surrounding the patients, their healthcare needs, and the delivery of healthcare will affect the way performance is measured. By using the PMMCC conceptual model, measurement developers, researchers, and policy makers can refine and implement performance measurement sets to effectively evaluate and improve care for individuals with MCCs.

To demonstrate how the PM-MCC model can be applied, as well as its utility, we present the case of a 60-year-old woman with known congestive heart failure (CHF), diabetes, hypertension, and depression as just 1 example of a complex patient.

 Over the course of 6 months, this patient received care from an internist in primary care, a cardiologist in specialty care, a spousal caregiver at home, and a pharmacist in the pharmacy. The types of care provided included screening, prevention, treatment, and management. Over the episode of care, the patient had some weight loss and fatigue. There was no acute exacerbation of her conditions or change in the ites or types of care provided, although it is not clear that her depression was adequately treated initially given her fatigue and weight loss. Examples of the complex medical decision making her providers faced include how low to push her blood pressure and the choice and dose of antidepressants in light of a previous fall and her ongoing risk of falls. It is not rare for patients to be more complex than the patient described here—cognitive impairment due to traumatic brain injury or dementia, or a history of substance abuse, are just 2 examples. Cognitive impairment would increase the involvement of her spouse in her healthcare decision making and management, and would make care coordination more challenging.

The ideal measurement framework would assess the goals and priorities of this patient, and tailor the performance measures to the priorities of the patient. For example, this patient, her spousal caregiver, and her physician may be more concerned about depression control and antidepressant medication management than blood pressure or diabetes control. This information could theoretically drive which denominators of performance measures should be applied to her (eg, the denominator of an antidepressant medication adherence measure but not the denominator of a glycated hemoglobin control measure) or the relative weight given to performance measures for this patient (with higher weight given to achieving antidepressant medication adherence than to achieving glycated hemoglobin control). This method of selecting or weighting performance measures based on individual patient goals and priorities is not currently feasible in most settings. However, personal health records and health information technology may make this type of patient-centered measurement more of a reality.

To demonstrate how this model can be used with current performance measurement methods, we will examine this case at the health plan level. We can start by looking at the subpopulation of patients similar to the case example: adults with MCCs receiving care from more than 1 outpatient provider with no acute medical events over the course of a stable episode. Healthcare performance related to treatment, management, and secondary prevention in both the primary and specialty care settings can be measured across all 6 measure domains that map to the national priorities for improving healthcare (see Table). Many of these measures would apply to all patients in the population (ie, regardless of conditions), such as patient and family experience of care, medication review, and assessment of treatment interactions. Some measures are specific to the diseases of the patient (ie, diabetes, hypertension, CHF, and depression). It is important to note that although guidelines exist for the treatment of each of these conditions individually, it is not clear that specific disease guidelines are appropriate when a patient has more than 1 condition; thus, disease-specific performance measures may not all be appropriate.

 For example, applying the Optimal Diabetes Care composite measure (a measure of processes and outcome associated with ideal diabetes care) to people with MCCs (eg, the patient described above) may not be clinically appropriate and may misdirect the focus of care from other more pressing conditions for the patient, such as depression or heart failure.

 Although disease-specific performance measures are still important for patients in this adult MCC population, they should be implemented with an acute awareness of how the treatment of one disease can affect the treatment of any or all other diseases or conditions. 

At the same time this population is receiving ongoing treatment and management, screening and prevention of primary and secondary conditions are also important. In the adult MCC population, some screening measures may be more important for some subgroups than for others. For example, urinary incontinence screening is recommended for adults 65 years and older, but not for younger adults. Alternatively, cervical cancer screening is recommended for women aged 21 to  64 years.

 In addition to performance measures for providing appropriate screening, measures should be developed to reward plans or physicians for avoiding inappropriate screenings that could potentially cause harm to the patient.

Finally, it is important to consider the other sites and providers where patients in this population may be receiving care. As new delivery systems and Accountable Care Organizations (ACOs) develop, measures outside the current scope of health plans should be considered, such as patient out-ofpocket costs or availability of transportation services in the community. 

Now let us examine the case described above when the patient has an exacerbation of her CHF, is hospitalized, and is diagnosed concurrently with community-acquired pneumonia. Hospital care is added to the patient’s measurement framework,and the patient begins receiving care for the acute exacerbation of CHF and pneumonia. From the level of the health plan measuring performance, this patient may now fall into a different subpopulation for measurement: adults with MCCs hospitalized for an acute episode with subsequent follow-up care. Similar to the application of the framework to a stable episode, performance measures can be non—disease specific, with a focus on the prevention of delirium and infection during the hospitalization and subsequent readmissions; or they can be disease specific, with a focus on the use of antibiotics and continued assessment of oxygen saturation. 

The measures outlined above and in the Table are not intended to be exhaustive. Rather, they illustrate the potential performance measures that fall into the measure domains that map to the 6 National Quality Strategy priority areas. On the other hand, implementing this entire list of measures could create a significant burden of measurement on patients, providers, and health plans. What the PM-MCC model offers is a framework for organizing existing measures and for prioritizing those measures based on the characteristics of the population being examined, and eventually, individual patient preference. This model can also be used to highlight measure gap areas where we currently lack rigorous tools for the purposes of quality measurement (see italicized text in the Table). As can be seen from the Table, particular areas where we lack rigorous quality measurement are non—disease-specific measures of effective treatment and prevention and measures of health and well-being. These measure gap areas may cut across condition, site, and type of care. For example, treatment burden and goal attainment may provide an important addition to the measurement of quality of healthcare and may be higher patient priorities than disease-specific measures.

Healthcare quality improvement is moving toward care delivery models that aim to improve care for individuals with MCCs. The Patient Protection and Affordable Care Act provides multiple opportunities for the development of these new models (eg, ACOs, Center for Medicare & Medicaid Innovation, Independence at Home Demonstration, Patient-Centered Outcomes Research Institute

); however, we lack the appropriate tools and approaches to measure the performance of these new healthcare delivery models for individuals with MCCs, who are often the most costly at-risk population. We present a measurement model that will move performance measurement on the path toward measurement that cross-cuts across conditions, sites of care, and types of care, and focuses on the goals and preferences of the patient and family. Policy makers, researchers, and health plans can use this model to select and prioritize existing performance measures for subpopulations of people with MCCs in the context of the complex healthcare experience, as well as to identify areas where measures need to be developed or improved.

Building on the National Quality Forum patient-focused episodes of care measurement framework, which examines an episode of care for a single condition,

the PM-MCC model shows the heterogeneity and scope of care required by people with MCCs. By illustrating this complexity, the risk of uncoordinated care across settings and the resulting adverse health outcomes becomes more apparent. Although the PMMCC model does not provide a solution for shared accountability across silos of care, it is a model for the development of future measures that can be applied across multiple conditions (eg, goal attainment),

 measures that can be applied across settings (eg, use of a coordinated electronic health record [EHR]

), and measures that can be applied across types of care (eg, assessment for drug-disease interactions

). To date, many performance measures have been designed to be setting specific so as to assign accountability for performance on a measure to a single setting or provider. These siloed measures are partially a reflection of a siloed healthcare system where each care setting or provider is only responsible for 1 piece of a patient’s continuum of care. However, new models of care delivery such as ACOs and integrated delivery systems provide an opportunity to measure performance and assign accountability across the continuum of care. The PM-MCC model is well suited to guide ACO performance measure prioritization and development because it focuses on patients in the context of the multiple settings and providers where they are receiving multiple types of care, often simultaneously. There are several important issues to consider when applying this model. Although the center of this model is a patient with individual goals and preferences for care, current data sources are not designed to track this type of information. Therefore, tailoring quality measures to the individual’s goals and preferences for care is not currently feasible on a  widespread basis, and many measures that can apply to a population of people with MCCs will not reflect the care priorities of the individual patient. In the example described in “Applying The Conceptual Model,” although the patient and her provider decided depression was a higher priority of care than diabetes at the moment, there is not currently a methodology for excluding patients from measurement of diabetes outcomes ased on preferences for care. Similarly, the measures within a domain may represent different priorities to different patient populations. For example, informed decision making may be of the highest priority to a patient with multiple treatment ptions, whereas a patient with limited means may prioritize access to care. New systems are needed to capture patientreported data that are integrated with EHRs to facilitate the use of measures based on  patient preferences and priorities for care. In the meantime, consumers, policy makers, researchers, health plans, and practices need to prioritize which measures are most important for subgroups of individuals, recognizingthat priorities may differ by age, dominant condition, or other characteristics.

There are significant methodologic and logistical challenges to developing and implementing new performance measures for people with MCCs. Current data sources are not optimal to support performance measurement for people with MCCs. Physicians and health plans frequently cite measurement burden as a significant drain on resources. While the spread of EHRs may hold particular promise for performance measurement by providing the structure for documenting key elements and allowing for easier extraction of data, misclassification of performance data can happen.

 Furthermore, EHRs are not designed to collect outcomes often of the greatest  importance to patients and families, such as symptom burden, daily function, and quality of life.

Performance measurement must be based on rigorous evidence that a process of care is the best possible treatment because it leads to improved outcomes.

However, clinical trials often exclude medically complex patients, older people, people with MCCs, or people undergoing multiple medical interventions.

 When treating patients with MCCs, providers must use clinical judgment in deciding whether to intervene without evidence indicating what treatment is most likely to cause benefit rather than harm. They should use informed decision making that involves both patients and their families. Development of new performance measures must be guided by clinical research into new treatments and guidelines that address people with MCCs.

Despite these challenges, the PM-MCC model is a useful tool for organizing and prioritizing existing performance measures on multiple levels (population, health plan, or practice) and identifying areas for future development of measures that examine quality of care across conditions, providers, and sites of care over time. Given the increasing prevalence of individuals with MCCs in the United States, the rising cost of providing care for this population, and the current investment in quality improvement initiatives, it is imperative that we know how to measure healthcare performance beyond the traditional process-based, disease-specific framework. We present the PM-MCC model as a step toward developing andimplementing healthcare performance measures that work in this complex world.

 From Johns Hopkins School of Medicine (ERG, SD, BL, CMB), Baltimore, MD; Johns Hopkins School of Public Health (SD, BL, CW, CMB), Baltimore, MD; National Quality Forum (TBV, AP), Washington, DC; University of Michigan Health Systems (CSB), Ann Arbor, MI; Interim Healthcare (BAM), Sunrise, FL.

 Dr Giovannetti reports employment with the National Committee for Quality Assurance and has received honoraria for guest lectures at Johns Hopkins. Drs Leff and Boyd report that they have received grants from the National Quality Forum. The other authors (SD, KA, TBV, ATP, CSB) report no relationship or financial interest with any entity that would pose a conflict of interest with the subject matter of this article.

 Concept and design (ERG, BL, KA, TBV, ATP,  CSB, CMB); acquisition of data (ERG, BL, CMB); Concept and design (ERG, SD, BL, CW, CMB, KA, TBV, ATP, CSB, BAM); drafting of the manuscript (ERG, SD, BL, CW, CMB); critical revision of the manuscript for important intellectual content (SD, BL, CW, CMB, KA, TBV, ATP, CSB, BAM); obtaining funding (CBM, TBV, KA); and supervision (CMB).

This work was funded by a subcontract from the National Quality Forum through a contract from Department of Health and Human Services (Contract HHSM-500-2009-0010C). Dr Giovannetti’s work was supported during her postdoctoral fellowship under a National Institute on Aging (NIA) T32 Training Grant (T32 AG00021 052507). Dr Boyd is a Robert Wood Johnson Foundation Physician Faculty Scholar and is supported by the Paul Beeson Career Development Award Program (NIA K23 AG032910, American Federation for Aging Research, The John A. Hartford Foundation,The Atlantic Philanthropies, The Starr Foundation, and an anonymous donor). Dr Leff is supported by The John A. Hartford Foundation and The Atlantic Philanthropies, and is an American Political Science Association Health and Aging Policy Fellow. This article is based on material that was prepared under, and paid for by, the Centers for Medicare & Medicaid Services under Contract HHSM-500-2009-00010C.

 Erin R. Giovannetti, PhD, Research Scientist, National Committee for Quality Assurance, 1100 13th St NW, Ste 1000, Washington, DC 20005. E-mail: giovannetti@ncqa.org.1. US Department of Health and Human Services. Multiple Chronic Conditions: A Strategic Framework. Optimum Health and Quality of Life for Individuals With Multiple Chronic Conditions. http://www.hhs.gov/ash/initiatives/mcc/mcc_framework.pdf. Published December 2010. Accessed January 2012.

2. Fortin M, Bravo G, Hudon C, et al. Relationship between multimorbidity and health-related quality of life of patients in primary care. Qual Life Res. 2006;15(1):83-91.

3. Naessens JM, Stroebel RJ, Finnie DM, et al. Effect of multiple chronic conditions among working-age adults. Am J Manag Care. 2011;17(2): 118-122.

4. Wolff JL, Boult C, Boyd C, Anderson G. Newly reported chronic conditions and onset of functional dependency. J Am Geriatr Soc. 2005; 53(5):851-855.

5. Marengoni A, von Strauss E, Rizzuto D, Winblad B, Fratiglioni L. The impact of chronic multimorbidity and disability on functional decline and survival in elderly persons: a community-based, longitudinal study. J Intern Med. 2009;265(2):288-295.

6. Parekh AK, Barton MB. The challenge of multiple comorbidity for the US health care system. JAMA. 2010;303(13):1303-1304.

7. Boyd CM, Darer J, Boult C, Fried LP, Boult L, Wu AW. Clinical practice guidelines and quality of care for older patients with multiple comorbid diseases: implications for pay for performance. JAMA. 2005;294(6): 716-724.

8. Librero J, Peiró S, Ordiñana R. Chronic comorbidity and outcomes of hospital care: length of stay, mortality, and readmission at 30 and 365 days. J Clin Epidemiol. 1999;52(3):171-179.

9. Wolff JL, Starfield B, Anderson G. Prevalence, expenditures, and complications of multiple chronic conditions in the elderly. Arch Intern Med. 2002;162(20):2269-2276.

10. Hohl CM, Dankoff J, Colacone A, Afilalo M. Polypharmacy, adverse drug-related events, and potential adverse drug interactions in elderly patients presenting to an emergency department. Ann Emerg Med. 2001;38(6):666-671.

11. Tinetti ME, Bogardus ST Jr, Agostini JV. Potential pitfalls of diseasespecific guidelines for patients with multiple conditions. N Engl J Med. 2004;351(27):2870-2874.

12. Bayliss EA, Edwards AE, Steiner JF, Main DS. Processes of care desired by elderly patients with multimorbidities. Fam Pract. 2008;25(4): 287-293.

13. Fried TR, McGraw S, Agostini JV, Tinetti ME. Views of older personswith multiple morbidities on competing outcomes and clinical decision-making. J Am Geriatr Soc. 2008;56(10):1839-1844.

14. Fortin M, Dionne J, Pinho G, Gignac J, Almirall J, Lapointe L. Randomized controlled trials: do they have external validity for patients with multiple comorbidities? Ann Fam Med. 2006;4(2):104-108.

The Performance Measurement for People with Multiple Chronic Conditions conceptual model can facilitate development and refinement of quality measures for a medically complex population.

Performance Measurement for People With Multiple Chronic Conditions: Conceptual Model

: Guided Care (GC) is a model of proactive, evidence-based comprehensive healthcare provided by physician-nurse teams for people with several chronic health conditions. Our objective was to evaluate the preliminary effects of GC on health service utilization and costs.

: Cluster-randomized controlled trial of GC involving 14 primary care teams (49 physicians) and 904 of their chronically ill patients age 65 years or older.

: Using insurance claims, we compared the health services used by patients who received GC with the health services used by patients who received usual care during the first 8 months of the study.

: After adjustment for baseline characteristics, GC patients experienced, on average, 24% fewer hospital days (95% confidence interval [CI]: 49% fewer, 13% more), 37% fewer skilled nursing facility days (95% CI: 65% fewer, 5% more), 15% fewer emergency department visits (95% CI: 38% fewer, 18% more), and 29% fewer home healthcare episodes (95% CI: 53% fewer, 8% more), as well as 9% more specialist visits (95% CI: 8% fewer, 29% more). Based on current Medicare payment rates and GC costs, these differences in utilization represent an annual net savings of $75,000 (95% CI: −$244,000, $150,900) per nurse, or $1364 per patient. 

: Initial introduction of GC into primary care practices may be associated with less use of expensive health services and a net savings in healthcare costs among older patients with several chronic health conditions. Final results from the remaining 2 years of this ongoing study will be published in 2011.

Guided Care is a model of proactive, comprehensive healthcare by physician-nurse teams for people with several chronic health conditions. If its promising early effects persist and reach statistical significance during the remainder of an ongoing cluster-randomized controlled trial and are replicated during upcoming demonstration programs, Guided Care may:

Attract public and private health insurance coverage in the future.

Become adopted widely by US primary care practices.

Healthcare for older persons with chronic conditions is very expensive.

 Growing recognition of these costs has led to the development and promulgation of numerous new disease and care management strategies designed to provide high-quality, costeffective chronic care. One such approach is Guided Care (GC).

 Guided Care is provided by a practice-based team that includes a registered nurse, 2 to 5 physicians, and the other members of the office staff. This team provides 8 clinical services to a panel of 50 to 60 of the practice’s highest-risk older patients. For each patient, the Guided Care nurse (GCN) (1) performs a comprehensive assessment at home, (2) creates an evidence-based care guide, (3) monitors and coaches the patient monthly, (4) coordinates the efforts of all of the patient’s providers of healthcare, (5) smooths the patient’s transitions between sites of care, (6) promotes patient self-management, (7) educates and supports family caregivers, and (8) facilitates access to appropriate community resources.

In a pilot study during 2003-2004, GC appeared to improve the quality of chronic care

 In the initial months of the present cluster-randomized controlled trial (cRCT), patients and physicians rated GC more highly than usual care (UC).

 The aim of the present analysis was to evaluate the effects of GC on the use and costs of health services during the first 8 months of this cRCT. We report these preliminary results now to help inform business and policy decisions related to improving chronic care that must be made before the publication of the study’s final results in 2011.

In 2006, we launched a cRCT of GC in 8 community-based primary care practices located in urban and suburban neighborhoods in the Baltimore, MD—Washington, DC, metropolitan area. Three practices were operated by Kaiser Permanente, a group-model managed care organization; 4 were operated by Johns Hopkins Community Physicians, a statewide network of community-based practices; and 1 was operated by Medstar Physician Partners, a multisite group practice. The study population comprised older persons insured by traditional fee-for-service Medicare (34%) or managed care plans (45% Kaiser Permanente, 21% Tricare/USFHP, a federal health insurance program for retired military personnel and their dependents). Additional details about the study, which was approved by 3 relevant institutional review boards, have been published previously.

We screened the insurance claims of all patients of the 8 practices to identify those who were age 65 years or older and at high risk of using health services heavily during the following year, as estimated by the claims-based Hierarchical Condition Category (HCC) predictive model.

 “High risk” was equated with HCC scores of 1.2 or higher, which identifies persons in the highest quartile of risk. Eligible high-risk patients who provided informed consent completed in-home baseline interviews.

Within the 8 practices, we identified 14 “pods,” each of which consisted of 2 to 5 primary care physicians and their consenting high-risk patients. The study’s statistician, blinded to the identities of the pods, randomly allocated each pod (physicians and their patients) to either GC (7 pods) or UC (7 pods).

Registered nurses who had completed a course in Guided Care Nursing joined their assigned pods in May 2006. During the following 6 to 8 months, each GCN established a caseload of 50 to 60 GC patients. The date on which each patient’s care guide was developed was his/her start date for receiving GC.

Patients in the UC group continued to receive the same care they had received before enrolling in the study. We randomly assigned their start dates for receiving UC to reflect the GC patients’ distribution of start dates.

Information about baseline characteristics was obtained from the prerandomization in-home interviews. Data about health service utilization were obtained from paid insurance claims for services used during the period from 1 year before the person’s start date through June 30, 2007. To maximize the capture of most health services, we waited until late November 2007 to begin the process of querying, cleaning, and merging the 3 insurers’ claims databases. We calculated participants’ baseline HCC scores from claims for services used during the 12 months before their start dates. We grouped these health services into 9 categories: hospital, skilled nursing facility, primary care physician, specialty physician, home healthcare, emergency department, durable medical equipment, tests (radiologic and laboratory), and nonmedication treatments (eg, physical therapy, outpatient procedures).

 we imputed values for missing baseline interview questions. We compared the GC and UC patients’ use of health services during the intervals between their start dates and June 30, 2007. For each category of healthcare service, we constructed a marginal regression model to estimate the effect of GC (vs UC) on the mean units of service used per person per year (

). For each category, we modeled the logarithm of the mean as a linear function of treatment group, age, race, sex, education, finances, HCC score, self-rated health, activities of daily living, instrumental activities of daily living, and practice site, plus an offset term for exposure period. Regression parameters were estimated using generalized estimating equations with a working independence covariance structure. The estimated variance-covariance matrix of all the regression estimators was obtained using the sandwich variance technique. The adjusted treatment effect for each outcome was interpreted as the ratio of the mean units of service used per GC recipient (vs UC recipient) over a common exposure period.

To facilitate computation of the difference in costs associated with GC, we first quantified the average units of service used annually by the recipients of UC and multiplied these units by 55 (the average number of patients in a GCN’s caseload). Then we computed the analogous average units of service used by 55 recipients of GC by multiplying the UC units of service by the treatment effects estimated by the marginal regression models. Finally, we calculated the difference between the 2 groups’ aggregate healthcare costs by multiplying the differences between the groups’ annualized use of services by Medicare’s average payment per unit of service.

 Excluded from these estimates are the costs of durable medical equipment, emergency department visits, tests, and treatments, which are extremely heterogeneous and for which average Medicare payment rates are not available.

Confidence intervals for the annualized use of services and cost differences were derived by using a 2-stage bootstrap procedure. In the first stage, individuals were resampled within practices. In the second stage, treatment effect estimates (on a log scale) were sampled from a multivariate normal distribution in which the mean was equal to the estimated treatment effect (on a log scale) and the variance-covariance matrix was equal to that obtained from the generalized estimating equation procedure. A total of 100,000 bootstrap samples were generated, and confidence intervals were formed using the percentile method.

 13,534 older patients were screened for eligibility, 1763 were offered the option of participating, and 904 were randomized by pod to receive either GC (n = 485) or UC (n = 419). We excluded from these analyses utilization by participants whose insurance claims were unavailable (1.4%) and those who rescinded consent (1.8%), died before their start dates (2.2%), or did not have start dates before June 30, 2007 (2.2%). We analyzed data on the remaining 835 participants (n = 433 in the GC group, n = 402 in the UC group).

At baseline, the demographic features, comorbidity, and HCC risk scores of the 2 groups were similar, but the GC participants were less likely to experience economic challenge (10.4% vs 15.5% lacked “enough money at the end of the month”; 

The early months of a cluster-randomized trial suggest that Guided Care is associated with a trend toward a net reduction in total healthcare expenditures.

Bruce Leff, MD

Articles

Attributes Common to Programs That Successfully Treat High-Need, High-Cost Individuals

Innovative Care Models for High-Cost Medicare Beneficiaries: Delivery System and Payment Reform to Accelerate Adoption

Performance Measurement for People With Multiple Chronic Conditions: Conceptual Model

Guided Care and the Cost of Complex Healthcare: A Preliminary Report