Since 2006, 4 lines of development have converged in Vermont's health care reform program, creating a unique opportunity to test a new model for improving population health. First, more than a decade ago, Kindig (1) issued a challenge to improve the outcomes of an American health care system that spends twice as much per capita for health care services as other developed countries, while achieving third-world rates of illness and death. Recently, he renewed the challenge, calling for the development of a "pay-for-population health performance system that goes beyond medical care to include financial incentives for the equally essential nonmedical care determinants of population health" (2).

Second, in 2006 Vermont enacted legislation creating one of the nation's most ambitious health care reform programs (3). Building on foundations laid in the previous 5 years, the state attempted to achieve a sustainable reduction in the number of uninsured residents, accelerate the implementation of health information technology, and transform the prevention and treatment of chronic illness through a program called Blueprint for Health. Treatment of chronic illness accounts for more than 65% of all health care expenses in Vermont, but current practices offer major opportunities for improving performance. Blueprint for Health is based on the Chronic Care Model (4) and is a true public-private partnership supported by a broad base of stakeholders (5). Having only 600,000 residents, Vermont proved to be an ideal laboratory for testing meaningful delivery system reform as a major component of its broader health care reform effort to improve coverage and health information technology. Its small-scale, noncompeting delivery system and history of collaboration between stakeholders provided a supportive, nurturing environment for the proposed changes. Every year since the initial health reform legislation passed in 2006, Vermont has added legislation to strengthen and broaden health reform, including mandating a model called the enhanced medical home and coordinating strategies to prevent chronic illness. The legislature and its Health Care Reform Commission have led this process, but the implementation of delivery system reform has required sustained shared leadership by both the legislative and executive branches and by private-sector stakeholders.

Third, in 2007 the Institute for Healthcare Improvement began its Triple Aim project to drive large-scale system change by 1) controlling total per capita medical costs, 2) improving the population's health, and 3) improving the care experience of health care consumers (6). The institute created a learning collaborative that brought together an international collection of health care organizations implementing the Triple Aim project. The Vermont Blueprint for Health accepted the invitation to join the initial learning collaborative and continues to participate.

Finally, Vermont adopted the model of the accountable care organization (ACO) suggested by Fisher et al (7) based on their research documenting widespread, large variations in health care use without improvement in outcomes. The ACO model is built around creating a new set of financial incentives for a community provider network of physicians, local hospitals, and other caregivers for a defined population. The financial incentives are based on a pool of shared savings that is distributed when specific quality criteria are achieved.

This article describes Vermont's statewide effort, which weaves together these 4 lines of development and offers the prospect of creating a prototype for Kindig's pay-for-population health system in similarly rural areas. As part of its broader health care reform agenda, Vermont is attempting to build a statewide network of community health systems, which would provide both the infrastructure and financial incentives required to improve population health. The community health system involves multiple levels of reform to create the integration needed for effective population health incentives. The first, most basic, level is the enhanced medical home, which gives primary care practices the ability to better coordinate care with other providers and support behavior changes in their patients. The second level is the ACO, composed of the local hospital, specialists, and other key providers who work with the medical home practices. The Vermont community health model incorporates a prevention and population health incentive.

By the end of 2009, phase 1 of system reform was implemented in 3 pilot communities serving 10% of the state's population. Planning is under way for phase 2 pilot programs that combine the ACO concept with an incentive model built on the Triple Aim goals. Legislation enacted in May 2010 expands the enhanced medical home program from a pilot to a statewide initiative and commits state support to phase 2: 3 ACO pilots that use incentives based on the Triple Aim goals. Several characteristics make Vermont a unique statewide laboratory for implementing these reforms. It has a small population, a delivery system with no directly competing hospitals, a simple payer system with only 3 major commercial payers, and a long tradition of collaboration between major stakeholders. Health care reform has enjoyed long-term bipartisan support from both a Republican governor and a Democratic-majority state legislature. These qualities make it unlikely that other states will implement community health systems in exactly the same way that Vermont has, but the conceptual framework developed in Vermont can be generalized to other settings, particularly those with more rural delivery systems. This article will first present the conceptual framework of the proposed network of community health systems, focusing on the different types of integrator roles necessary for success. Then, it will describe the design of the enhanced medical home pilots and the results of the feasibility study for the ACO pilots.

The Vermont experience has revealed the necessity of integration at 3 geographic levels.

The 3 geographic levels are interdependent, interacting through the following 5 categories of functional capacity that create the required integration.

Vermont's reform plan consists of 2 phases, the first at the medical home level and the second at the community health system level. These reforms include changes in financial incentives to transform the delivery system; they have been challenging to design in a multipayer environment. Although payment reform is necessary, it is not a sufficient requirement for building a community health system. Too often, policy makers have assumed that simply changing the financial system will drive other necessary changes. Vermont's experience shows that the substantial structural changes needed require building new capabilities in all 5 functional categories.

To concentrate resources and coordinate efforts, Vermont used pilot communities. This approach had several benefits. First, because the changes were pilots and not systemwide, they were less threatening and easier to adopt. For example, it would have been impossible to implement all payment reforms statewide. Second, the competition to become a pilot community galvanized local leadership and created a more receptive climate for change. Third, scarce state resources could be focused in a more concentrated way, which prevented their premature dilution. Finally, the pilot design incorporated formative evaluations, which allowed the state to learn while implementing and recognize that these efforts are a work in progress. The corollary to the use of pilots is that scaling them statewide will require federal support through national health reform. Vermont can begin the process of building a community health system but cannot finish the task with state resources alone.

The enhanced medical home pilots involve primary care practices in 3 communities (9). These pilots are designed to strengthen the functional capacity of primary care practices to coordinate care across settings and to support behavior changes in their patients while providing the infrastructure to enable them to serve as a medical home. The objective of the pilots is to reduce the prevalence of chronic illness and its complications and to improve compliance with national prevention and treatment guidelines. The pilots have 5 components.

If the phase 1 pilots achieve results similar to those of other closed-system settings such as the Geisinger Health System (12), they will be able to meet the first of the Triple Aim goals, per capita cost savings, by reducing unnecessary hospital admissions and emergency department visits for treatment of chronic disease. To meet the other 2 Triple Aim goals of improving population health and experience of care, the local community health system must be able to share in those savings and reinvest them locally. The community activation and prevention plan created by the pilot community health teams will guide investments in each community, including priorities for key nonmedical determinants of health. However, in the absence of a second phase of reform, the financial benefits of the enhanced medical home simply flow downstream to the payers. The primary care practices have received an enhanced payment, but otherwise the community has no additional resources available to improve the health of its population.

Vermont's ACO model incorporates the Triple Aim incentives to address this issue. The model creates a shared savings incentive pool based on projected medical expenses, which is distributed on the basis of agreed-on quality measures and population health targets. As the next stage of health system reform to build a sustainable community health system, ACO pilots will be implemented. The Health Care Reform Commission has conducted a feasibility study for implementing a community-level incentive system based on ACOs (13). At the same time, the Dartmouth Institute for Health Policy and Clinical Practice and the Englelberg Center for Health Care Reform at Brookings jointly developed a national learning collaborative to implement several ACO pilots nationwide. Staff from both organizations participated in the Vermont feasibility study and contributed their research findings. After finding encouraging results from this study, legislation was passed directing the Health Care Reform Commission to collaborate with the executive branch of the state government and interested provider networks to develop a Vermont application for the ACO national learning collaborative (14).

The ACO feasibility study created a working design for the pilots, building on the medical home as the essential first step (15). The Health Care Reform Commission created a broad-based work group that identified potential obstacles to building the community level of integration. The group focused on 3 categories.

Qualified ACO pilot sites were identified, and the Vermont ACO pilots are being developed. The state regulatory agency for insurance is facilitating conversations with commercial insurers regarding a shared savings pool. Vermont's state Medicaid agency is also developing a plan to participate in the ACOs. The Blueprint for Health program is contributing to the design of the ACO model to ensure effective coordination between the medical home practices and ACOs.

Vermont has not yet created a true pay-for-population health system, but the state has found no obstacles that cannot be overcome. A substantial missing piece, federal participation, is not assured, but  national health care reform legislation explicitly authorizes and funds Medicare participation in ACO pilots. Vermont provides a statewide laboratory for assembling a bench model that will allow the state to test design issues that still need to be explored. Building a replicable, functioning pay-for-population health system should be just a matter of time.