The Adirondack Medical Home Demonstration: A Case Study

Transcription

1 The Adirondack Medical Home Demonstration: A Case Study A N I S S U E B R I E F

2 OFFICERS J. Barclay Collins II Chairman James R. Tallon, Jr. President William M. Evarts, Jr. Patricia S. Levinson Frederick W. Telling, PhD Vice Chairmen Sheila M. Abrams Treasurer David A. Gould Sally J. Rogers Senior Vice Presidents Sheila M. Abrams Deborah E. Halper Vice Presidents Stephanie L. Davis Corporate Secretary DIRECTORS Richard A. Berman Jo Ivey Boufford, MD Rev. John E. Carrington Philip Chapman J. Barclay Collins II Richard Cotton Richard K. DeScherer William M. Evarts, Jr. Paul Francis Michael R. Golding, MD Josh N. Kuriloff Patricia S. Levinson Howard P. Milstein Susana R. Morales, MD Robert C. Osborne Peter J. Powers Katherine Osborn Roberts Mary H. Schachne John C. Simons Howard Smith Michael A. Stocker, MD, MPH Most Rev. Joseph M. Sullivan James R. Tallon, Jr. Frederick W. Telling, PhD Mary Beth C. Tully Howard Smith Chairman Emeritus United Hospital Fund The United Hospital Fund is a health services research and philanthropic organization whose mission is to shape positive change in health care for the people of New York. We advance policies and support programs that promote high-quality, patient-centered health care services that are accessible to all. We undertake research and policy analysis to improve the financing and delivery of care in hospitals, health centers, nursing homes, and other care settings. We raise funds and give grants to examine emerging issues and stimulate innovative programs. And we work collaboratively with civic, professional, and volunteer leaders to identify and realize opportunities for change. HONORARY DIRECTORS Donald M. Elliman Douglas T. Yates Honorary Chairmen Herbert C. Bernard John K. Castle Timothy C. Forbes Barbara P. Gimbel Rosalie B. Greenberg Allan Weissglass

3 The Adirondack Medical Home Demonstration: A Case Study Gregory Burke C ONSULTANT Sean Cavanaugh D IRECTOR OF H EALTH C ARE F INANCE U NITED H OSPITAL F UND UNITED HOSPITAL FUND

4 Copyright 2011 by United Hospital Fund ISBN Free electronic copies of this report are available at the United Hospital Fund s website,

5 Contents INTRODUCTION 1 SETTING AND MAJOR PARTICIPANTS 2 The Community: Geography, Population, and Demographics 2 The Health Care Providers 3 The Payers 4 The State 5 OVERVIEW OF THE DEMONSTRATION 5 DEVELOPMENT OF THE DEMONSTRATION 6 Phase I ( ): Reaching Consensus on the Problem 6 Phase II ( ): Designing the Project 8 Phase III ( ): Selling and Funding the Project 10 State Health Department 10 Discussions with Payers 11 Health Information Technology 11 Summer-Fall, 2009: Funding and Approval 12 Phase IV (2010 and Beyond): Implementation 13 Overall Organization and Oversight 13 Role and Function of the Pods 14 Guidelines for Participation 15 Outstanding Payer-Provider Issues 15 Looking Forward: A Focus on the Deliverables 16 WHAT ELEMENTS WERE CRITICAL TO THE DESIGN OF THE DEMONSTRATION? 16 A Clearly Defined Population with a Real Problem 16 A Promising Intervention 17 A Community-Oriented Provider System 17 Engaged and Collaborative Payers 17 State Leadership 18 Investment Capital and Targeted Grantmaking 19 Personal Leadership 19 Strategic Use of Expert Consultants 20 CONCLUSION 20 REFERENCES 22 iii

6 Introduction With the passage of the Affordable Care Act (ACA), the next phase of health reform is now under way. The focus is shifting from laying a new foundation for national policy to actually building it at the local and regional level. That is where changes in medical practice must occur, if the nation s health system is to improve quality, safety, and health status, and reduce avoidable utilization and cost. Local and regional initiatives are now being designed and implemented across the country. Partnerships of providers and payers are mounting demonstration projects focused on improving the performance of local health care delivery systems particularly the way they provide and organize care for the chronically ill, who are the health system s highest-cost users. Historically, such initiatives have had limited success because of their small size. Most providers serve patients who are covered by many different plans, and the proportion of practices covered by any one payer is generally relatively small. This lack of scale limits any one payer s ability to effect change in provider behavior. To be effective, such experiments require a critical mass of both providers and payers serving the region. Payers need enough of the providers in a given region to participate in order to effect a measurable change in their enrollees health status, utilization of services, and costs of care in that region; and providers need enough of the payers with whom they work to agree on a common set of desired changes and a consistent set of incentives to drive behavior change (Cavanaugh and Burke 2010). The evidence indicates that most successful models occur within integrated delivery systems like Geisinger, Mayo, Kaiser, and Group Health, in which physicians are employed or organized into groups and payment systems are more unified, often using capitated payment systems that include incentives for improved care and outcomes (Shih et al., 2008; Crosson 2009). Achieving meaningful change is substantially more difficult in mainstream delivery systems, which have many unaffiliated, communitybased private practices operating in a feefor-service payment environment, serving patients covered by many different payers. One promising integrated effort in New York is the Adirondack Medical Home Demonstration (AMHD). In the northeast corner of the state, providers, payers, and the state government have been working together for four years to improve the way primary care is delivered in the region by implementing the principles of the patient-centered medical home in roughly 40 primary care practices. This pilot project is testing the effect of implementing advanced patient-centered medical homes in an entire region that is rural and underserved. It involves essentially all primary care providers in the region, who have committed to achieving National Committee for Quality Assurance (NCQA) certification as patient-centered medical homes and other specified goals; and essentially all major payers in the region, who have agreed on a specific set of services and deliverables for which they would be willing to pay and a common payment rate for those services. The project is focused on improving access, quality, and continuity of care for all patients, while reducing avoidable utilization and cost. The purpose of this case study is to Case Study: The Adirondack Medical Home Demonstration 1

7 describe the Adirondack initiative, and to draw from it some lessons that may be useful to other regions in New York State and the entire nation. In the following sections, we will describe the setting, the issues, the key players, and the events leading to the AMHD; identify some key principles involved in the development of the AMHD; and highlight the public policy and other issues raised by this project, and their potential applicability elsewhere. This case study is based on research conducted during summer of 2010, including interviews conducted with many of the leaders involved (who were extraordinarily generous with their time), and the review of key documents related to the evolution of the Adirondack Medical Home Demonstration. Any mistakes or misstatements are our own. Setting and Major Participants The Community: Geography, Population, and Demographics The Adirondack Medical Home Demonstration spans four of the Adirondack Mountain region s counties in their entirety (Clinton, Essex, Franklin, and Hamilton), and parts of a fifth (Warren). The overall region (Figure 1) is an area the size of Connecticut. Dominated by the Adirondack State Park (nearly 6 million acres), it includes one-fifth of New York State s land area but less than one percent of its total population. It is highly rural, with a population density of 28 people per square mile, compared to 180 for upstate New York and 412 for the state overall. Figure 1: Regional Map of the AMHD St. Lawrence Franklin Clinton Jefferson Essex Orleans Niagara Monroe Wayne Genesee Erie Ontario Wyoming Yates Schuyler Chautauqua Cattaraugus Allegany Stueben Livingston Seneca Chemung Lewis Oswego Oneida Onondaga Madison Cayuga Tompkins Cortland Chenango Tioga Broome Herkimer Otsego Hamilton Warren Fulton Saratoga Montgomery Schenectady Albany Schoharie Greene Delaware Washington Rensselaer Columbia Sullivan Ulster Dutchess Orange Rockland Putnam Westchester Bronx New York Nassau Queens Richmond Kings Suffolk 2 United Hospital Fund

8 The region s population is substantially older, on average, than the state population. A recent study by the LA Group showed that the Adirondack North Country has proportionately more people over the age of 65 than anywhere in the nation except southwest Florida the result of young people leaving the region because of limited economic opportunity (LA Group, 2009). The region has a depressed economy, divided into three sectors: harvesting natural resources, tourism, and service jobs. The first two sectors center on small enterprises offering seasonal employment, resulting in high levels of near-poverty and lack of health insurance coverage. The largest employers in the region (accounting in total for over 30 percent of total employment in the five counties) are municipal, county, and state governments. Over half of the land in the region is owned or controlled by the state for wilderness preservation. Residents of the region are disproportionately poor and sick. Compared to their counterparts in upstate New York, Adirondack Service Area residents are more likely to lack wage income (27.4 percent versus 24.6 percent), live in poverty (14.4 percent versus 12.0 percent), live with a disability (19.5 percent versus 15.3 percent), and lack insurance coverage (17.9 percent versus 12.9 percent). On the positive side, Adirondack communities have a history of collaborative efforts in which towns and counties have worked together to analyze shared challenges and then organize, develop, and share services. 1 This community approach to problem-solving has been supported by the New York State Association of Counties and the Adirondack Regional Planning Association. The Health Care Providers The health care system in the Adirondack region, like that in many rural areas, is comparatively lean. There are five hospitals in the region, with a total of 545 beds: Adirondack Medical Center (AMC), a 98-bed full-service hospital in Saranac Lake, which operates a 2-bed critical access hospital in nearby Lake Placid; Alice Hyde Hospital, a 76-bed full-service hospital in Malone; Champlain Valley Physicians Hospital (CVPH), a 341-bed full-service hospital in Plattsburgh; Elizabethtown Community Hospital, a 15-bed critical access hospital in Elizabethtown; and Moses-Ludington Hospital, a 15-bed critical access hospital in Ticonderoga. The Adirondack region has historically had difficulty attracting and retaining physicians, a problem faced by many rural communities; much of the region has been designated a federal Health Professional Shortage Area. The SUNY Center for Healthcare Workforce Studies reports that the four core counties of the AMHD initiative had the lowest supply of physicians in the state, as shown in Figure 2: 116 primary care physicians, a ratio of only 65 primary care physicians per 100,000 people, compared to a statewide average of 95 per 100,000; and 216 specialists, a ratio of only 122 non-primary care physicians per 100,000 people, compared to a statewide average of 220 per 100, There are several examples of collaboration among communities and health care providers in the Adirondack region. Most relevantly, the Adirondack Rural Health Network ( prepared regional needs assessments, a unified approach to developing hospital community service plans, and a collaborative approach to acquiring state, federal, and foundation grants. Other collaborative efforts include the Upper Hudson Primary Care Consortium and the Regional Health Information Exchange (HIXNY). Case Study: The Adirondack Medical Home Demonstration 3

9 Figure 2: Doctor/Population Ratios in Adirondack Region, 2005 MDs per 100K Population Clinton, Essex, Franklin, and Hamilton Counties (combined) New York State Average 0 All MDs Primary Care MDs Specialist/Non-Primary Care MDs Note: The four-county region described in the figure differs slightly from the AMHD region as it does not include any of the communities in Warren County. Source: Authors analysis of data from Center for Health Workforce Studies, December There are 123 primary care physicians participating in the AMHD. One-quarter (32) practice in 17 solo and small practices in the four-county Adirondack region. One-third (39) are employed by the Hudson Headwaters Health Network (HHHN), a large and wellestablished 13-site federally qualified health center that is the main provider of primary care in the southern tier of the region (Warren County, Essex County, and the surrounding communities). The remaining primary care providers (52) are employed by local hospitals. Finally, the region has three distinct and largely non-competitive geographic areas within which the providers are already organized, each with its own natural leadership: Champlain Valley Physicians Hospital in the north, Adirondack Medical Center in the west, and the Hudson Headwaters Health Network in the south. The Payers The insured population in the Adirondack region is covered by a mix of private and governmental payers, with roughly half covered by each. The seven private insurance carriers participating in the demonstration Excellus, The Empire Plan (which is administered by United Health Care and covers state employees and retirees), Fidelis, Empire Blue Cross, Blue Shield of Northeastern New York, the Mohawk Valley Plan (MVP), and Capital District Physicians Health Plan (CDPHP) collectively represent nearly 90 percent of all commercial covered lives in the region. Together with Medicaid, these seven plans represent approximately two-thirds of patients being cared for in the primary care practices in the demonstration (Figure 3). Another 18 percent are covered by Medicare, and the rest are covered by other insurers or are uninsured. 4 United Hospital Fund

10 Figure 3: Payer Mix in Practices Participating in AMHD (Initial Estimates, 2007 Sample) 18% Participating Plans Medicare Other and Uninsured 20% 62% The State The State of New York s role in the demonstration went well beyond that of a payer on behalf of Medicaid and state employees. The State has had a longstanding interest and involvement in the health system in the Adirondack region, and it has been a critical player in the development of the AMHD. Overview of the Demonstration The Adirondack Medical Home Demonstration is a multiprovider, multipayer collaboration involving essentially all primary care practices serving about 200,000 residents in a five-county region of northeast New York State. The impetus for this initiative was a crisis in the region s ability to attract and retain primary care physicians. From 2006 to 2007, the region lost nearly two dozen primary care physicians, significantly eroding its already small base of primary care physicians. The providers and the community viewed this crisis as an opportunity to effect real change, and identified the patient-centered medical home model as their best chance of improving primary care for physicians and patients alike. Over a four-year planning period, physicians, payers, state officials, and community leaders worked to achieve broad consensus on the structure and funding of this demonstration. The aim of this fiveyear demonstration project is twofold: to strengthen the Adirondack region s ability to attract and retain primary care physicians by improving their quality of life and increasing their income, and to transform the delivery of primary care in the region by increasing quality and continuity of care, reducing avoidable utilization and costs, and improving the health of the community. To achieve these ambitious goals, the participating physicians have committed to meeting criteria necessary for their practices to become NCQA-certified patient-centered medical homes, and to defining and ultimately achieving a variety of access, quality, and utilization goals over the next five years. The patient-centered medical home (PCMH) is a model intended to improve access to primary care while enhancing care coordination. It incorporates advanced health information technology, and revolves around Case Study: The Adirondack Medical Home Demonstration 5

11 provision of evidence-based preventive and chronic care coordination services delivered by a team of providers led by a physician with whom the patient has an ongoing relationship. The PCMH can have particular value for patients with chronic diseases who require ongoing surveillance and care for those conditions and training and support in selfmanagement. Such patients often experience more referrals to specialists and hospital admissions, and thus need more assistance from care managers with medication management and transitions between care settings. Patients with chronic diseases also generate more costs through higher numbers of emergency department visits and inpatient admissions many of which may be preventable. Early evidence supports the effectiveness of the PCMH in reducing emergency department visits and preventable hospital admissions and readmissions, within the context of integrated delivery systems (Paulus, Davis, and Steele, 2008; Grumbach and Grundy, 2010). The participating payers agreed to reimburse participating providers $7 per member per month (or $84 per year) over and above the normal fee-for-service payments for office visits to cover the costs of the expanded services of a medical home. These payments will enable the practices to put in place the elements of a PCMH, with the expectation that this model of care will result in fewer avoidable emergency department visits and hospital admissions in the future. The demonstration formally began on January 1, Over a five-year period, the participating physician practices are being held to a series of performance benchmarks to ensure that the new payments are financing improvements in patient care and generating cost savings that exceed their investment. Development of the Demonstration The AMHD is the result of more than four years of planning, in four broad phases. These phases are outlined in the accompanying box and detailed in the following section. Phase I ( ): Reaching Consensus on the Problem Over the years, the Adirondack region has had difficulty recruiting and retaining physicians. This problem reached crisis proportions in 2006 and 2007 when the region lost two dozen primary care physicians nearly 10 percent of the region s total supply of primary care physicians as a result of retirements and relocations. As one of the hospital executives noted, We recognized that if we didn t do something dramatic, we were at risk of losing our primary care base. These losses generated serious concern among the physicians, communities, employers, and insurance companies serving the region. All worried about the region s ability to maintain access to quality health care, and about the impact on the region s already weakened economy. In 2007, with the assistance of the New York State Association of Counties, local health care leaders convened a summit to define the magnitude of the problem, identify its root causes, and build a constituency for responding to it. The Adirondack Health Summit gathered business leaders, schools, service organizations, environmental advocates, and public officials from all levels of government along with members of the media to alert 6 United Hospital Fund

12 AMHD DEVELOPMENTAL PHASES Phase I ( ): Reaching Consensus on the Problem Reach consensus on the magnitude of the problem and its root causes Quantify its impact on access to care, management of chronic disease, avoidable emergency department visits and hospitalizations, and cost Identify possible solutions Phase II (2007-8): Designing the Project Physicians and groups evaluate the Patient-Centered Medical Home model Consensus among providers to pursue regionwide PCMH project, and agreement on goals Baseline assessments Existing capabilities and needs relative to NCQA criteria HIT infrastructure required vs. current status Resources required in individual practices and as shared services Resources required and return on investment Phase III (2008-9): Selling and Funding the Project Discussions and negotiations State discussions: antitrust protection, Medicaid participation, role as convener Discussions with payers on costs and payment rates for the AMHD Regional application for HEAL-10 funding Funding and approval New York State legislature passes AMHD legislation (antitrust, Medicaid funds, DOH/OHIP role) HEAL-10 application approved, funded Medical Society of the State of New York funding for pod development, and consulting support to practices Payers agree to scope of program, deliverables, and payment rate Phase IV (2010 and Beyond): Implementation Organizing the project, the Pods, and the ADK Governance Council Guidelines for Participation signed by all participating providers and payers Pods build capacity to support practices in care/case management, HIT support, etc. Engage consultants to help practices implement HIT/EMR changes and achieve PCMH certification January 1, 2010: Project goes live (includes payment of PCMH rate to participating providers) Year 1 (2010): Targets and milestones By January 2010, each practice to complete detailed readiness assessment and develop work plan By July 2010, each practice to activate electronic prescriptions By February 2011, each practice prepared to apply for NCQA certification as a medical home During 2010, establish two project databases (clinical and claims) and establish baseline data Years 2 5: Project operational Providers operate NCQA-certified medical homes Providers submit regular reports and data tracking operations, process, and outcomes measures Measures focus on access, continuity, clinical care (six conditions), and satisfaction Utilization and costs tracked using clinical and claims databases Case Study: The Adirondack Medical Home Demonstration 7

13 the entire community to the need for action. Projects of this scale need a forcing function, said one of the participants. This was a community in pain. Providers, employers, and the community all felt keenly the impact of losing primary care physicians. There was a broad consensus that something had to be done. Summit participants identified two root causes for the loss of primary care physicians: Low reimbursement. According to one participant in the study, private insurance payment rates to primary care providers in the region were substantially lower than in other areas in the state and the nation. This affects both the privately practicing physicians (whose average income was reported to be roughly two-thirds of the Medical Group Management Association national average) and the region s federally qualified health center (more than half of whose patients are covered by private insurance). Eroding quality of professional work life. Primary physicians were working long hours without appropriate support staff, unable to spend adequate time managing and responding to the needs of their most complex and chronically ill patients. Many primary care physicians worked in solo practices or small groups, with little support staff, limited opportunities for professional interaction, and inadequate coverage arrangements for after-hours care. An additional challenge noted was the inability of most practices to invest in new health information technologies. In 2007, fewer than 40 percent of the primary care physicians in the region had installed an electronic medical record (EMR) system; and of those who had done so, many had unsuccessful or incomplete installations. Lacking training funds, hardly any practices with EMRs developed patient registries, implemented electronic reminders, or turned on automated features. The summit produced a broad-based consensus regarding the importance of developing a regional solution to this regional problem. There was a commitment among providers to work together with payers and the state to craft and implement a response to the problem and its causes. The summit also identified a specific solution: the emerging model of the medical home, which had the potential to enhance the way primary care practices function improving quality and continuity of care, and reducing avoidable utilization and costs in return for increased reimbursement to support expanded primary care services consistent with the chronic care model (Bodenheimer, Wagner, and Grumbach, 2002a&b). Phase II ( ): Designing the Project Over the next two years, at the request of provider leaders, the New York State Department of Health (DOH) convened a planning group that included representatives from the area s primary care practices, the hospitals, the HHHN, the seven private payers, the Medical Society of the State of New York (MSSNY), the New York State Department of Civil Service, and the Office of Health Insurance Programs (Medicaid) within the DOH. In addition, physicians were convened locally to discuss the project in sub-regional groups (which became known as the pods ) that followed the region s natural geographic alignments: Pod 1 (Tri-Lakes), organized around AMC s existing physician-hospital organization; Pod 2 (Lake George), organized around HHHN s health center infrastructure; and Pod 3: (Plattsburgh), organized around CVPH s medical services organization. The use of structures that the primary care physicians knew and trusted was critical to developing a physician constituency for the demonstration. It was very important to have 8 United Hospital Fund

14 Figure 4: Patient-Centered Medical Home Standards (Screen Shot from NCQA Web Site) Source: National Committee for Quality Assurance. Available at (accessed January 19, 2011). physician leadership in the pods, said one of the current pod managers. Doctors listen to doctors. The goal of the planning group was to develop an action plan to implement the medical home model in primary care practices across the region, and to achieve certification by the NCQA (Figure 4). In mid-2007, the planning group engaged a consultant from EastPoint Health, a group with national experience in developing medical homes, including the development of a network of 11 new medical homes in New Orleans following Hurricane Katrina. With this assistance and leadership, the group helped participating physicians better understand the PCMH model, its requirements, and its potential benefits. The consultants helped the planning group and pod leaders assess the baseline capacities of the practices involved and identify what they would need to do to achieve NCQA certification as a PCMH. This included specifying which services were beyond the capacity of small practices to support and would instead need to be developed and shared among practices. With help from the Mass ehealth Collaborative (a group with extensive experience in EMR implementation in Massachusetts), they clarified the health information technology (HIT) infrastructure needed to support Case Study: The Adirondack Medical Home Demonstration 9

15 the medical home. Finally, the planning group and its consultants estimated the type of resources and investments required to implement the model, as well as the improvements in quality and safety, reductions in avoidable utilization, and cost savings that could all result from the initiative. Ultimately, the groups agreed upon four key objectives for the medical home initiative: 1. Improve quality and outcomes. Improve quality, coordination, outcomes, and safety of medical care for patients in the demonstration through the development of advanced medical homes within participating primary care practices. 2. Lower overall health care costs. Reduce unnecessary utilization of health care services and lower health care costs to generate a return on investment to ensure the program is sustainable. 3. Improve access. Significantly improve access to care by improving the retention and recruitment of primary care clinicians and thereby ensure a strong, sustainable primary care network for the region. 4. Create a new clinically integrated model. Create a successful model for public-private and payer-provider collaborations to improve health care value and transform primary care a model that can then be replicated in other parts of the state. The PCMH model for delivering better care, coupled with the possibility of being paid more for delivering better care, had immediate appeal for the physicians. As one physician noted, PCMH is the only really good frame-work out there for primary care in the U.S. Patients really love it. The participants became increasingly enthusiastic about the possibility of a regionwide demonstration project. As one physician put it, Why are physicians involved? First, because it s good for the patients, and second, because it means that they re being paid for what their care is worth. However, the group also noted that the PCMH would disrupt existing systems and ways of doing things in the practices and would involve new costs. They also noted the need for financial support from a critical mass of payers. If you do a medical home said one of the physicians, you have to do it for an entire practice, not just for an individual payer. You can t have a medical home for some of your patients. It s all or nothing, at the practice level. From a practice standpoint, we have to be able to say, Medical Home is our model of care. We don t care who your payer is. Ownership of the demonstration by the providers was crucial. The Adirondack project was very much a provider-driven effort said one of the payers. The magic here came from the providers, not the payers. They came to us with a proposal. Another payer put it, This demonstration is not about payment reform, but about implementing medical homes in primary care practices, and how to pay for it. Phase III ( ): Selling and Funding the Project The leadership of the AMHD next turned to gaining support for the proposal on three fronts: discussions with the state, discussions with payers, and securing funding for the necessary HIT improvements. State Health Department In early 2008, the leaders of the AMHD met with senior officials in the DOH and the Governor s office, and gained their support for the initiative in three areas: 1. Antitrust protection. During the initial planning process, the payers had been uncomfortable meeting with their competitors and with providers 10 United Hospital Fund

16 to discuss products (what they would be willing to buy as a PCMH) and pricing (what they would be willing to pay for it). Many payers felt antitrust immunity would be required. 2. Medicaid budget. Although leaders from Medicaid had been involved in the development of the project, state budget approval would be required for Medicaid to participate in the demonstration. 3. State role as a convener. Finally, the project needed an organization to convene the stakeholders to work through the design and implementation of the demonstration. The state seemed like the logical choice for such a role. Discussions with Payers With the state Department of Health serving as convener, the project leaders then met with the payers to discuss the project s deliverables and the possibility of augmented payment methods that would be made available to the participating practices. Although generally supportive of the AMHD, the payers were concerned about paying separately and in addition to their normal fee schedules for the expanded services of a PCMH without some analysis of the specific additional services and costs involved, and some estimate of the savings they could expect. As one of the payers put it, Medical homes show promise. We re cautiously optimistic, but we need more evidence. In July 2009, the providers engaged a financial consultant (RSM McGladrey) with extensive experience in reimbursement analysis to estimate the costs of implementing the PCMH in participating practices, and to develop a methodology for augmented payments to practices for the expanded services of the PCMH. While there was some literature on the costs of existing programs, the variation in design and payments among these national models made it difficult to simply use the data from other demonstrations. The consultant reviewed the available literature, identified the discrete elements in an NCQA-certified medical home, and estimated the total incremental cost of those services per practice and per enrolled patient. This analysis was presented in June 2009 at a meeting of stakeholders, who agreed with most of the underlying assumptions and results. Many of the payers agreed in concept to providing additional reimbursement for medical homes to cover the costs of implementing and operating the medical home model, and to provide a substantial increase in the income of the participating primary care physicians. However, there was less agreement regarding the actual amount to be paid, the populations for whom the payments would apply, and mechanisms for payment. In return, the participating primary care physicians agreed, in concept, to pursue and achieve NCQA medical home certification in their practices by a set date, and to meet additional performance benchmarks over the following years. Health Information Technology Achieving recognition as a PCMH requires physicians to implement and use electronic medical records in caring for their patients. At the time the AMHD was being developed, most of the region s primary care practices had not implemented this technology. Coincidentally, in mid-2009, the Department of Health issued a solicitation for grant funds (HEAL-NY Phase 10, also known as just HEAL-10 ) with a specific focus on projects using health information technology to develop and operate medical homes. The AMHD project leadership led a regionwide process to develop an application for HEAL- 10 funding. While that application included a bigger group of payers and providers (including hospitals and specialists not participating in the AMHD and the local regional health Case Study: The Adirondack Medical Home Demonstration 11

17 information organization, HIXNY) and a broader set of deliverables than the AMHD called for, its main focus was on technologies and capacities that would support the medical home project: EMR acquisition or remediation for all PCPs involved in the AMHD; Regional data exchanges enabling PCPs to exchange data with hospitals and other providers; Electronic prescribing capabilities; Patient management and quality improvement infrastructure; Internet portals to enable non-traditional communication between PCPs and their patients after hours and on weekends; and Creation of two regional databases one to aggregate clinical data and the other to aggregate the payers regional administrative claims data. Both of these databases were considered essential for the management and evaluation of the AMHD. Summer-Fall, 2009: Funding and Approval In the late summer, the state legislature passed the 2009 budget bill, which included a DOH-sponsored section that provided the state s endorsement of the demonstration, including the antitrust immunity the payers had requested, and also provided funds for Medicaid to participate in the AMHD. The budget also articulated the role of the Department of Health as convener, chair, and monitor of the AMHD s planning and implementation, which was necessary to ensure the antitrust protections. In August of 2009, the Adirondack proposal was awarded a $7 million HEAL grant to support investments in health information technology that were key to the AMHD. HEAL-10 was particularly important, noted one of the pod managers. It s the funding to make sure that the physicians have and can effectively implement EMRs in their practices; that their EMRs connect to the region s RHIO [regional health information organization], and from there to the specialists and hospitals; and that the two data warehouses necessary for QI [quality improvement] and evaluation are in place. A month later, the AMHD was awarded a $2.7 million grant from the MSSNY to further support the regional HIT plan, including technical assistance provided by EastPoint Health and the Mass ehealth Collaborative to the participating primary care practices, and the creation of the three regional pods, intended to provide shared services to participating practices and assist in providing care coordination. Although the payers agreed in concept to pay the new fees for each enrollee receiving care from a participating practice, implementing this proved challenging for payers whose IT and payment systems were not capable of generating monthly capitation payments. As one payer put it, Claims payment systems are built to do one thing: pay claims for visits and procedures. Everyone pretty much agrees that the payments should be per-member, per-month. However, the mechanics involved in doing that are substantial. This issue was resolved by giving payers the option of paying either a $7 per member-month fee, or its equivalent via an add-on payment attached to a specific, existing CPT (current procedural terminology) billing code. In October 2009, the leaders of the AMHD sponsored a second Adirondack Health Summit to review the project to date, and to gain final agreement by all of the payers to participate in the AMHD. Throughout the planning process, payers had varied in their responsiveness and interest in the demonstration. Following some targeted advocacy from local and state officials, as well as regional employers and purchasers, by late October 2009 all of the payers had agreed to participate. After further negotiation, all of the 12 United Hospital Fund

18 payers agreed to pay the participating primary care physicians $84 per member per year (i.e., $7 per member per month) for each of their plans eligible enrollees who received primary care at their practices. This medical home payment was to be over and above their current fee-for-service payments for visits and procedures. The financial commitment of the payers was significant. The new payments will generate an estimated $30 $35 million over the five-year project. Of particular note, in this demonstration the payers committed to support the costs of developing the PCMH model in the practices. According to several participating payers, in most other demonstrations around New York and the country, payers have not made additional payments until after the practices achieved PCMH accreditation. As one payer put it, The provision of the up-front payments was, for the payers, a leap of faith. But we re committed to this for the five-year cycle. We expect to break even by Year 3, and to achieve a positive ROI with savings in Years 4 and 5. Another departure from the usual approach was the breadth of physician inclusion. The AMHD includes essentially all primary care practices in the region. In most cases, payers work with fewer practices, those with more advanced HIT systems and further along the road toward becoming a medical home. Usually, experiments like this use only gold standard providers, one payer noted. In return for these payments, the providers committed to achieving a series of milestones, including gaining NCQA certification for their practices by the second year of the demonstration, and to reducing avoidable utilization of emergency departments and inpatient services so that, by the third year, they generate savings that offset or exceeded that year s medical home payments. The providers also agreed to generate and report performance data related to their practice operations. For practices to receive the medical home money, they must prove that they are gathering and reporting data on their performance and that they meet quality standards. Practices must be able to demonstrate that they are adhering to evidence-based guidelines in six areas. Phase IV (2010 and Beyond): Implementation In the fall of 2009, the AMHD took final steps of preparation for its January 2010 implementation, beginning with the formal organizational structure of the project. Overall Organization and Oversight The AMHD formally established a 17- member oversight body (the ADK Governance Council), with representation from primary care physicians, hospitals, payers, the Department of Civil Service, the Department of Health, and the Department of Insurance. The ADK Governance Council is responsible for overall project evaluation, overseeing quality improvement activities, evaluating the performance of the participating primary care providers against agreedupon milestones and clinical and administrative measures, and collecting and analyzing key data, including the creation and management of the two consolidated AMHDwide databases (clinical data and claims data). The Governance Council uses the clinical database to support provider-specific quality assessments, identify implementation issues, and identify training and educational needs among participating practices. The claims database is used to track and evaluate patterns of utilization and cost for the overall study population and provide timely performance data to practices for clinical quality improvement. Case Study: The Adirondack Medical Home Demonstration 13

19 Role and Function of the Pods The three pods provide local governance and direction for physicians in each of the three sub-regions. The central focus of the pods is to support the physicians and practice managers in making the changes they need to make, helping them to succeed in implementing their EMR systems and in changing their operations to achieve medical home standards. They also provide a number of essential shared services (e.g., health education, care coordination, and pharmacy consultations) that are important for effective care management but difficult to mount economically in small practices. To cover the expenses of these shared services, the physicians have agreed to allocate the $7 per member-month in PCMH funding as follows: $3.50 goes to the primary care practice, as increased compensation; $3.00 goes to the local pod, to support shared services; and $0.50 goes to support the overall project governance, management, data, and evaluation functions at the ADK Governance level. The different roles of the participating practices, pods, and the Governance Council, along with the proposed flow of funds, are shown in Table 1. Table 1: Roles and Flow of Funds within the AMHD Proposed Body Roles Flow of Funds ADK Governance Council - Overall governance, evaluation, and measurement $0.50 PMPM - Provider/payer participation - Health information technology - Project databases (EMRs; clinical and claims databases) - Overall quality improvement Pod - Local, physician-led governance of pods $3.00 PMPM - Assistance with PCMH certification and EMR implementation - Organize, staff, and provide shared services: Care coordination, complex case management Care transition management Pharmacy and social service consultations Practice - Implement EMR, electronic prescribing, registries $3.50 PMPM - Achievement of PCMH standards, certification - Improved access, team-based care, referral tracking - Reengineered visits, evidence-based medicine - Improved chronic disease management PMPM: Per member per month. Total: $7.00 PMPM 14 United Hospital Fund

20 Guidelines for Participation The AMHD Governance Council has issued Guidelines for Participation in the AMHD, an agreement that clearly spells out the requirements for primary care physicians and payers participating in the demonstration. For a primary care practice to participate in the AMHD and receive the augmented payments, it must: be located in the specified geographic region; commit to monitoring by the Governance Council, including the provision of practice data for evaluation purposes; and agree to meet performance standards established by the National Committee for Quality Assurance (NCQA) for Level 2 or Level 3 recognition as a patient-centered medical home within 18 months. These agreements have been signed by all participating primary care physicians, payers, and hospital partners. Outstanding Payer-Provider Issues As of fall 2009, two issues remained to be resolved in order for the payers to participate in the demonstration: the methodology for attributing the payers members to specific practices, and the question of whether to include patients whose employers were self-insured (and purchasing administrative services only from the health plans) or under experience-rated contracts. Attribution methodology. The AMHD leadership worked with the payers to resolve the question of how to attribute enrollees and patients to a specific primary care physician. This was an important step, since it was the basis for making the monthly payment and for holding the right practice responsible for the care, utilization, and costs of each enrollee. Attribution proved difficult to resolve because many enrollees either did not have a specified primary care provider or used more than one. During the summer of 2010, the payers and providers agreed to a uniform attribution methodology, which resulted in an initial allocation of enrollees to specific physicians. Table 2 summarizes the outcome for each payer within each pod. Table 2: Participating Enrollees, by Payer and by Pod Participating Northern Adirondack Lake George Tri-Lakes Total Patients Payers Patients Patients Patients in Pilot A 20,994 1,887 4,798 27,679 B 5,500 2,946 2,751 11,197 C 2,200 3, ,650 D 2,562 2, ,525 E 2,328 2, ,524 F 548 3, ,512 G 26 3, ,962 H 11,901 5,035 1,648 18,584 Medicare 12,722 13,382 4,872 30,976 Total Enrollees 58,781 39,666 15, ,609 Percentage of Total in Pilot 52% 35% 13% 100% Case Study: The Adirondack Medical Home Demonstration 15

21 ASO enrollees. Resolution of the status and inclusion of members covered under Administrative Services Only (ASO) arrangements was a complex task. The plans argued that they could not include these patients in the demonstration without the consent of the employers. One payer observed, To convince the self-funded and experience-rated purchasers to participate in this demonstration, we need to be able to show them documented ROI. The data are not strong enough yet. In addition, gaining agreement to participate in this demonstration means convincing each of them that it is in their interest to expend new dollars for health benefits. That s a lot of work, and we do have day jobs. In the end, different plans handled this issue differently. Some sought and received approval of the purchasers to include their members in the demonstration, some simply absorbed the incremental costs of the demonstration, and others chose to exclude the ASO members in the demonstration. Looking Forward: A Focus on the Deliverables By late fall of 2009, the multiprovider, multipayer AMHD had established a clear set of milestones that participating practices had to reach in order to remain part of the demonstration and be eligible to receive the augmented medical home payments: Assessment and plan (December 2009). An assessment of the status of each practice had to be completed and submitted to the pods and ADK Governance Council by December 31, This assessment formed the basis for the practice-specific work plans developed and followed by the pod staff and the consultant teams working on the implementation of AMHD and HEAL to address any open issues. That milestone was met by all participating primary care practices. Electronic prescriptions (July 2010). Each practice had to demonstrate that it was capable of electronically submitting prescriptions to pharmacies by July 1, That milestone was met by all participating primary care practices. NCQA application (February 2011). Each practice had to apply for NCQA certification as a Level 2 or Level 3 medical home by February 28, Working with the participating practices to ensure that they can achieve this milestone is currently a top priority of the AMHD, and the main focus for pod and consultant staff. Performance reporting (ongoing). The physicians have agreed to a series of reporting requirements that include a common set of quality metrics, measures of patient and provider satisfaction, and quarterly measures of utilization and cost. What Elements Were Critical to the Design of the Demonstration? In our interviews, we were able to identify eight factors that contributed to the successful development of the Adirondack demonstration. A Clearly Defined Population with a Real Problem The Adirondack community recognizes itself as a community, and it has a history of successful collaboration to meet urgent needs. The region is not an artificial construct imposed by policymakers. The providers were clearly identified as members of and contributors to the communities, closely aligned with and responsive to their needs. Physicians, hospitals, and health centers all had a local constituency and strong local support (a number of towns were already 16 United Hospital Fund

22 providing office space free of charge to primary care physicians). The erosion in the region s primary care base the impetus for this initiative was a crisis felt by the entire community. A Promising Intervention The medical home model has received a great deal of attention within health policy circles in recent years. 2 It was developed by respected professional organizations, it appeals to both providers and patients, and it has a growing track record of success in improving care processes and outcomes. The ability to rely on established medical home standards from NCQA, a respected external accrediting body, also allowed participants to focus on implementing an existing model rather than designing one from scratch. A Community-Oriented Provider System Because of the Adirondack region s geography and the distances between providers and systems, its health care sector is less competitive than many, which enabled this multiprovider collaborative to take root. As one of the hospital executives noted, Up here, when the nearest other hospital is an hour or more away, there s less competition. There s a lack of the negative energy from competition. It s a good environment; it lets you focus on your community. The providers also understood that the process needed to be data-driven, and they were willing to take part in a process that was transparent to other providers, payers, and the state. They have participated in an open decisionmaking process, prepared to accept accountability for their own behavior and that of their colleagues, and actively pursued quality and performance improvement in their practices and across the region. As one of the physicians noted, The demonstration is about establishing a common set of standards, about providers holding each other accountable for the care they provide, its quality and costs reminding each other, we ought to be better than that. The existence of organized groups and a large number of employed physicians was an important factor in reaching consensus on a course of action. The use of existing organizational structures was also important; these were organizations with legitimacy and a track record, to which the primary care physicians felt some historical relationship and alignment, and in which they felt they had a meaningful voice. These sub-regional vehicles the pods were used effectively in the planning process, and they will house key elements of the clinical and care management infrastructure needed for implementation. Engaged and Collaborative Payers The complexity of organizing a multipayer collaborative to support a project of this type is covered in more detail in a companion paper, which details how the payers in the AMHD were able to work together to craft common payment solutions to common problems despite the many impediments they faced (Cavanaugh and Burke 2010). Although the payers were challenged by some elements of the AMHD initiative (methods related to attribution, ASOs, and payment, in particular), they exhibited substantial flexibility and were, in most cases, able to provide solutions to those problems. 2 See bibliography of PCMH Foundational Articles at the AHRQ website: home/1483/foundational_articles (accessed January 19, 2011). Case Study: The Adirondack Medical Home Demonstration 17

23 The state s provision of antitrust protection for payers ( not absolute immunity, noted a Health Department leader, but case-specific, around this particular project ) was a key to enabling them to participate in discussions about products, costs, and pricing. The Centers for Medicare and Medicaid Services has agreed to allow Medicare to participate in the AMHD starting in The Adirondacks was one of only eight sites in the United States selected by CMS in its Advanced Primary Care Practice Demonstration (CMS Office of Public Affairs, 2010). Although participants believed that Medicare participation was essential to the success of the AMHD, they also anticipate some additional challenges, such as new reporting requirements and a possibly a separate evaluation. State Leadership The leadership of the state Department of Health and other agencies has been critical to the implementation of the AMHD. Throughout the planning and implementation process, New York State has played many different roles (see accompanying box). THE ROLES OF THE STATE IN THE AMHD The State of New York has a longstanding interest and involvement in the health system in the Adirondack region, and it has been a key player in the development of the AMHD. Its leadership and support, in particular that of the Health Department and Medicaid, and the Civil Service, have been critical to the success of this project thus far. That interest and involvement is multifaceted, based on the state s different roles: As Payer: Over 20 percent of the region s population is covered by Medicaid or Medicaid Managed Care. As Employer/Purchaser: New York State is the region s largest employer, and the region s largest purchaser of private health insurance (via United s Empire Plan). The leadership of the New York State Department of Civil Service and United Health Care was instrumental in the development of the AMHD. As Regulator: Dealing with antitrust issues, crafting a safe space for the AMHD participants to work together on regionally based plans. As Public Health Agency: As lead agency for the state s public health and prevention agenda, focused on responding to the needs of the underserved and rural communities, reducing the burden of chronic disease, and supporting the role of federally qualified health centers. As Lead Agency for Health Policy and Innovation: Taking an activist role as proponent for health system performance improvement, the DOH is heavily involved in areas related to quality improvement. Payment system and delivery system innovation, redesign, and reform (e.g., experimentation with pay-for-performance demonstrations, quality improvement initiatives, using Medicaid payment incentives to stimulate ambulatory care delivery system reform, and stimulating adoption of PCMH). Expanding the use of HIT, funding key health information technologies, and infrastructure. 18 United Hospital Fund

24 Investment Capital and Targeted Grantmaking Although the payers are making a substantial financial investment in this demonstration, a variety of grant funds have filled important gaps throughout the process (Table 3). These grants have supported planning, HIT adoption and implementation, and the hiring of consultants to help practices implement the medical home model. As one pod manager noted, The HEAL and MSSNY grants and the up-front payments in Year 1 were critical. They supported the IT investments and some of the major costs of transformation in the practices, things that were needed to effect change. Personal Leadership The process would not have achieved what it has thus far without the ongoing commitment and strong personal leadership of key individuals from the provider community, physician and administrative leaders of the Hudson Headwaters Health Network, Champlain Valley Physicians Hospital, and the Adirondack Medical Center; from the Department of Health and the Department of Insurance; and from the purchasers and payers, including the Department of Civil Service. Table 3: Investments to Date in the AMHD Funding Source Amount Function Activities HRSA: Rural Health $80,000 Concept development Problem identification and proposed solution Network and early organizing Langeloth Foundation $16,000 HRSA: Rural Health $580,000 Strategy formulation Developing shared strategy for providers, Outreach (3 years) payers, and policy makers New York State Rural $35,000 Health Network New York State $7 million HIT infrastructure Adoption/remediation of EMRs for primary HEAL-10 (2 years) care practices and certain specialists; regionwide connectivity through HIXNY Medical Society of $2.7 million Implementation HIXNY interfaces; HIT design; medical home the State of New York (2 years) consulting, pod design, consulting assistance to primary care practices HIT infrastructure (pod design and startup) Participating Payers $40 $50 million Medical home Supporting costs of developing, implementing (est.) operational funding medical homes in participating primary care (5 years)* practices *Does not include potential contributions of Medicare, approved to join the AMHD in January Case Study: The Adirondack Medical Home Demonstration 19

25 Strategic Use of Expert Consultants The use of credible, expert outside consultants EastPoint Health, RSM McGladrey, and the Mass ehealth Collaborative was important in clarifying issues and generating consensus around several technical points: spelling out the presenting problem and its causes, identifying baseline capabilities in the participating practices, and mapping out actions required for the practices to achieve recognition as medical homes; assessing practices HIT capacities and needs, and developing plans to implement EMRs and other HIT capacities as part of the demonstration; and establishing and negotiating the cost basis and rates to be paid for the medical home capacities in the participating practices. Conclusion Developing and implementing the medical home demonstration, with the myriad of interests involved and the complexity of the undertaking, has been a remarkable achievement for the Adirondack community. Clearly, one goal increasing physician income was advanced with the implementation of the new monthly payments. However, whether the project can meet its other goals improving access, quality, continuity, and effectiveness of care, reducing preventable admissions and costs, improving the primary care physicians quality of work life, and improving the patient experience will be the real test of this demonstration. As one of the hospital executives noted, The real challenge in this demonstration is to establish a higher standard of practice not just for a single community, but for an entire region. Fundamentally, this initiative s success rests on the ability of the region s primary care physicians to embrace and make fundamental changes, including acquiring and implementing electronic medical records, redesigning their practices, successfully sharing support services across practices, and hiring and training new staff. In addition, the success of the demonstration also requires the creation of two new data warehouses: a clinical database that will be used for quality and performance improvement, and a claims database, that will be used to track utilization and costs. This will be a serious operational and financial challenge. As one of the payers noted, To have regionwide and practice-specific metrics on quality requires data aggregation and analysis, and the ability to report it on a practice and physician basis. That is complicated, and it is expensive. Currently, there s no incentive, no way to aggregate data across payers in a single data warehouse. Each payer looks at its own utilization and cost. All of these activities are supported by three different funding sources, each of which has its own specific focus, organizational structure, accountability for reporting, milestones, and deliverables. Even if the practices succeed in implementing all of these changes, it is unclear whether the medical home model alone can achieve the desired goals in long-term quality, utilization, and cost saving. The evidence on medical homes implemented in community practices, as opposed to within integrated delivery systems, is thin (Shih et al., 2008; Crosson 2009). As one physician noted, The medical home is necessary, but it will not by itself be sufficient. Just passing the NCQA test won t automatically generate savings. You need to do a number of additional things if you re going to make a difference in utilization and costs: create and manage registries, develop systems for care coordination, improve transitions management, provide pharmacy consultations and improve 20 United Hospital Fund

26 medication management, and provide effective health education and self-management training. The next step in the Adirondacks could be the evolution of the AMHD into some more integrated form of delivery, possibly an accountable care organization (ACO). ACOs were created in the Affordable Care Act to describe organizations that could assume clinical and financial risk for the care of a group of Medicare beneficiaries. If the AMHD participants are successful in implementing the medical home model in the region s primary care practices, they will have put in place some of the essential elements of an ACO by linking a defined population with an organized set of providers focused on improving quality and willing to accept financial risk for better outcomes and cost savings. The AMHD is clearly of and about the Adirondacks, but it also offers lessons that may apply elsewhere in New York State or beyond. If a defined and recognized community has a pressing need, and if the providers who serve that community can work together, and the payers agree to participate, it is possible to put together an initiative with the potential to transform the delivery of health care at the front end, in primary care physicians offices. But defining a recognizable and actionable community is not always easy, particularly in dense urban and suburban regions, where there may be neither a sense of shared identity and connectedness, nor a history of collaboration. Achieving an alignment of providers, payers, and community like that seen in the AMHD the foundation for achieving broad-based change in the system is not a given. It is something that may occur more easily in rural areas, often in one-hospital towns, and in some areas where there is a provider group that works well together and is dominant. However, in some towns with several hospitals, and certainly in dense, multi-hospital urban areas, the market is often split among competing providers. This competition results in a lack of alignment between a given provider group and its community that could make replication of the multipayer, multiprovider collaboration of the AMHD experiment more challenging. The final lesson of the AMHD case study is the need to reconsider the case for regional health planning, for creating opportunities to gather providers and payers in a given region and developing a sense of shared responsibility for improving access and quality, for consistently employing available evidence-based models for care when managing the care of the chronically ill and populations with particular needs in short, having local health systems operate more like integrated delivery systems, in the interest of the public and the communities they serve. Case Study: The Adirondack Medical Home Demonstration 21

27 References Bodenheimer T, EH Wagner, and K Grumbach. 2002a. Improving Primary Care for Patients with Chronic Illness: the Chronic Care Model. JAMA 288: Bodenheimer T, EH Wagner, and K Grumbach. 2002b. Improving Primary Care for Patients with Chronic Illness: the Chronic Care Model, Part 2. JAMA 288: Cavanaugh S and G Burke. November A Multipayer Approach to Health Care Reform. New York: United Hospital Fund. The Center for Health Workforce Studies. December Annual New York Physician Workforce Profile, 2007 Abridged Edition. Rensselaer, N.Y.: The Center for Health Workforce Studies, University at Albany, State University of New York, School of Public Health. Available at (accessed January 19, 2011). CMS Office of Public Affairs. November CMS Introduces New Center for Medicare and Medicaid Innovation, Initiatives to Better Coordinate Health Care. Press release, November 16, Baltimore, MD: Centers for Medicare and Medicaid Services. Crosson FJ. September st-Century Health Care The Case for Integrated Delivery Systems. N Engl J Med 361: Grumbach K and P Grundy. November Outcomes of Implementing Patient-Centered Medical Home Interventions: A Review of the Evidence from Prospective Evaluation Studies in the United States. Washington, D.C.: Patient-Centered Primary Care Collaborative. The LA Group, P.C. May Adirondack Park Regional Assessment Project Executive Summary. Available at Documents/File/16.pdf (accessed January 19, 2011). Paulus RA, K Davis, and GD Steele Continuous Innovation in Health Care: Implications of the Geisinger Experience. Health Affairs 27(5): Shih A, K Davis, SC Schoenbaum, A Gauthier, R Nuzum, and D McCarthy. August Organizing the U.S. Health Care Delivery System for High Performance. New York: The Commonwealth Fund Commission on a High Performance Health System. 22 United Hospital Fund

28 Additional copies of The Adirondack Medical Home Demonstration: A Case Study may be downloaded, at no charge, from the United Hospital Fund website,

29 Shaping New York s Health Care: Information, Philanthropy, Policy. Empire State Building 350 Fifth Avenue, 23rd Floor New York, NY (212) ISBN