Lessons Learned from Vermont on Building Community ACOs

Tucked into the new healthcare reform law are provisions to test accountable care organizations (ACOs)—groups of providers receiving set fees to deliver coordinated quality care to a select group of patients. But before most people even knew what ACO stood for, Vermont was ahead of the curve—putting together a strategy to launch three community-based ACOs by 2012.

In 2008, the Vermont legislature instructed its Health Care Reform Commission to assess the feasibility of piloting at least one ACO model as the next phase of payment reform. Along the way, it had some help from the nearby Dartmouth Institute for Health Policy and Clinical Practice, in Lebanon, NH, says Jim Hester, PhD, the commission's director.

Working with Dartmouth's and the Brookings Institution's 50-plus member ACO Learning Network, three ACO pilots have emerged in Vermont. The ultimate goal, as stated in a new report from The Commonwealth Fund on the "Vermont ACO Pilot," is to "achieve delivery system reform based on the development of a true community health system that both improves the health of the population it serves and manages medical costs at a population level."

The project started off by communicating with the providers. "We met with all [13] hospitals in the state at one point or another," Hester says. While several of the hospitals—including those doing medical home pilot work or additional medical work—turned down the initiative, others expressed interest—and became part of finding the way to "bend the medical cost curve" at the community level.

So how to finally pull an ACO together? A working design was developed for each ACO pilot that was built on three major principles:

Local accountability. “Given the natural care patterns of patients and provider referral patterns," no "lock in" of patients to the ACO would be necessary. Instead, the ACO patient population would be determined through "historical patterns of patients" who visited providers in the ACO—using a methodology developed by Dartmouth.

Payment reform. To address misaligned incentives "between fee for service payments and the need to better support providers taking steps to improve quality at a lower overall cost," the ACO model should include payment reforms based on shared savings, the report says. However, this shared savings should reflect an ACO's "varying stages of integration and sophistication.

For instance, the simplest option would be just shared savings—where providers do not assume any of the risk for above target spending. This option is designed for "newly formed entities with little experience managing care or risk."

However, more sophisticated provider organizations—such as Vermont Managed Care (in northwestern Vermont) or the United Health Alliance (in Bennington)—could consider "shared savings plus risk." Here, providers would assume a portion of the risk for above target spending (e.g., 20%) and are eligible to keep a greater portion of the savings.

Performance measurement. Measuring more than just medical expenses will be essential to ensure that appropriate care is being delivered and that cost savings are not the result of limiting necessary care, the report says. ACOs will be reporting patient experience data, in addition to clinical process and outcome measures.

Through Vermont's earlier "Blueprint for Health" initiative for chronic care patients and the state's collaboration with the Dartmouth Population Health Research Center," Vermont has been exploring additional sets of population health based measures for its ACO pilots, Hester notes.

Any ACO pilot also will need to have threshold capabilities in five areas to get started, the report said.

• First, the ACO must be able to manage the full continuum of care settings and services for its assigned patients—both public and private—beginning with a patient centered medical home approach to primary care.
• It must be financially integrated with both commercial and public payers (Medicare would begin in 2012). All payers will need to participate, so that at least 60% to 70% of patients in a provider's practice can be eligible for inclusion in a shared savings model.
• A health information technology platform that connects providers in the ACO—and allows for proactive patient management is essential—along with a strong financial database and reporting platform for managing the global medical budget. This information starts at the practice level with a Web based clinical tracking system, which provides flow sheets for individual visits, a registry for chronic illnesses, and population based reports.

  In addition, Vermont has an all-payer claims database that all commercial payers already are "syncing into a common system" and then generating reports, Hester notes.
• Physician leadership—as well as the commitment of the local hospital CEO and leadership team—will be vital to driving changes in process, cost structure, and mission.
• And, it must have process improvement capabilities to change both clinical and administrative procedures to improve the ACO's performance to achieve financial and quality goals.

But what will the ACOs eventually look like? Probably if you're seen one pilot ACO in Vermont—you've seen one. Achieving the necessary critical mass of patients that would support statistically meaningful measures of performance may require either a consolidated performance pool involving multiple payers—or expanding the ACO to include multiple hospitals.

The former is more likely, Hester says. "You'll need a certain critical mass in order to make the shared savings pools work," Hester says. It probably would have to be in the area of at least 15,000 patients. The latter could greatly complicate implementation issues, particularly governance. "We have a lot to learn and a lot of experimentation [to do] in terms of what kind of structures it can support, Hester says.