In the hothouse of health care innovation that is Minnesota's Twin Cities, executives and physicians at North Memorial Health Care arrived at a crossroads a couple of years back.

A fiercely independent lot consisting of two of the last unaffiliated hospitals in the region and several independent, mostly smaller, physician groups, they had succeeded for years in the northwest metro area in the shadow of some of the largest integrated systems in the country. But the expense of developing advanced data and management infrastructure to meet the growing demand for higher quality and cost accountability threatened to swamp their resources.

"Allina, Fairview and Park Nicollet all have pursued the Pioneer ACO model, so there is a movement afoot in the broader Twin Cities market to pursue this level of clinical integration," says Craig Matticks, M.D., president of the North Collaborative Care network, Robbinsdale, Minn. "Among payers, there is an appetite for collaboration with physician networks that may not have existed in previous attempts to control costs."

Since the physicians and hospitals mostly shared geography and patients, there was a strong basis for collaboration, Matticks says. So a task force of primary care and specialty physicians representing the entire care continuum convened to explore creating a network.

In mid-2012, the nonprofit North Collaborative Care network was incorporated. It comprises 353-bed North Memorial Medical Center and 130-bed Maple Grove Hospital, as well as 13 hospital-run primary care clinics, specialty clinics and dozens of independent physician clinics. Taking care to accommodate the diverse cultures of its members, the network set about developing the care paths, care management across the continuum, and electronic health record and information technology integration required to manage population health.

Despite early physician wariness, consolidation efforts continued. "To their credit, our board quickly saw the advantage of deepening clinical integration and decided to pursue ACO status with the Medicare shared savings program in 2014," Matticks says. "To move from forming a network and sharing data to getting into bed together as an ACO speaks volumes about how far we have come in 2 1/2 years. At first, they were explicitly saying they did not want to be an ACO."

Ready or Not, Risk is Coming

But an ACO is not always the best solution. For example, Metro Health, a 208-bed osteopathic teaching hospital with 204 employed and independent physicians affiliated through a physician-hospital organization in Wyoming, Mich., has developed sophisticated disease registries and care management systems that allow it to monitor performance and contract for performance bonuses.

However, Metro Health's population is spread out across western Michigan and its payer mix is fragmented, making accepting full risk difficult, says Frank Belsito, D.O., chief medical officer. "To really enter into those contracts, you need a large population paid by one payer. There is a critical mass you need to reach and we are not quite there."

Similarly, Lehigh Valley Health Network, comprising three hospitals and 1,100 employed and affiliated physicians, headquartered in Allentown, Pa., has shared savings contracts with several private payers, but has not pursued an ACO or bundled payments with Medicare. "We decided not to do it, but we are evaluating a commercial bundled-payment approach," says Sue Lawrence, senior vice president for the care continuum.

Metro and Lehigh's decisions put them well within the mainstream. In 2013, just 15 percent of hospitals had established or were part of an ACO, with another 7 percent actively working on it, according to the American Hospital Association's Survey of Care Systems and Payment [See Few and Far Between, below].

Still, executives at Lehigh and Metro — as well as every other hospital executive interviewed for this series — acknowledge that risk contracting based on quality and cost is coming, and they are developing management structures, clinical integration strategies and risk management capabilities to get ready. "We are not yet in an ACO, but everything we are doing supports it. When and if we are part of an ACO, we will have processes in place to do very well," Belsito says. But the capability is more important than the designation, he adds. "ACO is alphabet; it's what you do inside that counts."

Big is Good, not Essential

Hospitals and health networks are wise to think twice before undertaking an ACO strategy because it usually requires a large initial investment in management, IT, network development, care protocols, quality reporting and financial reserves, says Robert James Cimasi, CEO at Health Capital Consultants in St. Louis. He considers these sophisticated systems essential. "You can't get Buck Rogers without spending the bucks."

Current rules governing the Medicare Shared Savings Program strongly favor larger organizations in several ways, Cimasi notes. He estimates up-front costs of about $5.3 million to launch an ACO serving 5,000 Medicare beneficiaries, or about $1,040 per enrollee, rising to $12 million to serve 80,000 beneficiaries, or about $150 each. Large organizations, particularly those with managed care experience, also are more likely to have in place some essential infrastructure, such as experienced contract managers, advanced EHRs and sophisticated quality management, potentially further reducing out-of-pocket startup costs.

A shorter payback period is always better from a capital return perspective, Cimasi says. He notes that an ACO with a payback period exceeding three years may not be financially justifiable since it would run past the current Medicare contract limit, and renewal terms are unknown.

The advantage larger ACOs have from lower per-beneficiary initial costs is compounded in the one-sided, gain-sharing-only Medicare model, which requires the smallest ACOs to achieve 3.9 percent savings before receiving bonuses, falling on a sliding scale to 2.0 percent for the largest players.

Using Cimasi's cost and best-case cash flow assumptions, an ACO would need about 25,000 beneficiaries to reach a one-year payback of its initial cost under the two-sided shared savings and shared losses model, or about 50,000 beneficiaries under the one-sided shared savings-only model. ACOs serving 5,000 members are not financially viable under either model using these assumptions, and plans as large as 35,000 beneficiaries have a payback period pushing 12 years.

"Unless getting a reasonable return is feasible, there may be better places to invest scarce capital," Cimasi says. He recommends developing a comprehensive business plan that examines a range of structural factors affecting potential performance before proceeding.

However, finance alone should not drive the ACO decision, Cimasi says. Other positive externalities also should be considered. These may include strategic concerns, such as repositioning the system away from a provider of hospital beds to a convener of health resources across the continuum in anticipation of future payer demands for greater efficiency. It also might include mission-related concerns, such as quantifying the organization's role in improving community health.

"There could be a move to redefine community benefit in terms of positive externalities," Cimasi says. "The whole idea of an ACO has really positive benefits for the community, and that should not be overlooked as a reason to start one if it is not going to crush you [financially]."

Given the economies of scale, degree of management sophistication required, control over more resources across the care continuum, and efficiencies of delivering care in a densely populated area, the profile of hospitals participating in ACOs is not surprising. According to the 2013 AHA survey, ACO hospitals are three times more likely to be urban, and more than twice as likely to be teaching institutions and system-affiliated than non-ACO hospitals. Participation rates are almost directly related to size, ranging from 7 percent of those with up to 99 beds to 29 percent of those with 400-plus beds [See Something in Common, below].

Still, smaller ACOs can be very successful. For example, the Heartland Regional Medical Center ACO serving 12,000 Medicare beneficiaries in northwest Missouri was one of only two two-sided MSSP ACOs launched in 2012 to generate shared savings in its first year, earning an interim payment of $2.86 million.

However, the system developed its accountable care approach over a long period and applies it over a much broader population, notes Mark Laney, M.D., president and CEO of Heartland Health and Mosaic Life Care. "For more than 10 years, we successfully applied the accountable care model to our own employees — long before becoming an ACO under the CMS shared savings program." Heartland also uses the approach to get the most out of its free clinics and charity care operations, which remain substantial in part because Missouri declined Medicaid expansion under the Affordable Care Act.

Overall, 54 of 114 MSSP ACOs that were started in 2012 generated shared savings totaling $128 million, CMS announced in January. Similarly, a November 2013 independent review of Medicare Pioneer ACOs by L&M Policy Research, Washington, D.C., found that eight of 32 achieved significant cost savings totaling $155 million, with 23 close to local market costs and one with significantly higher costs totaling $8.5 million for the first year. That works out to a savings of $20 per beneficiary per month for the entire program, with the eight plans achieving significant savings ranging from $33 to $104 per beneficiary.

The review also found that successful Pioneer ACOs varied in geographic location, size, market conditions and organizational structure. This suggests that many types of ACOs, including integrated systems, group practices and hospital-physician networks, are viable, and can succeed even in markets with low initial Medicare costs.

Making It Work

The old health care saw, "when you've seen one integrated system, you've seen one integrated system," also applies to ACOs, Cimasi notes. Their diversity reflects both the flexibility of CMS ACO regulations, which to some extent were designed to accommodate differing markets and delivery systems across the country, and commercial programs, which vary even more. "An ACO is in the eye of the beholder and the federal version is much different from the commercial."

One thing they all share, however, is a commitment to reorient health care from delivering procedures to delivering value. Value is defined as delivering evidence-based care, including coordinating care across the continuum, as well as clinical, financial and patient satisfaction outcomes for both individuals and populations.

This puts providers at risk, but it is a risk they can conceivably control, Cimasi says. And unlike HMOs, which often created incentives to undertreat by transferring financial risk indiscriminately through capitation, ACOs reward providers for improving both quality and efficiency.

Defining value in terms of concrete clinical, financial and satisfaction outcomes also puts providers in the driver's seat, says Howard Gold, executive vice president for managed care and business development at 16-hospital North Shore-LIJ Health System in New York. "The insurers add no value unless they work with a provider who adds the value of a positive experience for their members. They have to pick sides and work with some of the more innovative providers."

The converse is true for providers; they must find payers who will support their transformation to accountable care, Gold says. Savvy payer partners are critical to develop arrangements that align rewards with value, and to manage the inevitable financial hit from reducing inpatient care and procedure volume. "The problem is most insurance companies have not made the leap that the future is working closely with providers who want to work with them. A few really get it, but others are just clueless."

Because the strategic, financial and operational implications are profound, system-level leadership is essential to building an accountable care infrastructure. At University Hospitals Case Medical Center, an integrated academic system comprising 10 hospitals and 1,500 employed physicians headquartered in Cleveland, the journey began in 2010 with a commitment of $100 million, says Eric J. Bieber, M.D., chief medical officer and chair of UH Coordinated Care Organization, the system's umbrella ACO.

"It wasn't ACOs in isolation. At the same time, we were looking at driving value through new models of care and leveraging IT," Bieber says. "Reading the tea leaves, we knew value would be the driver and we needed to make things as efficient as we could."

That the system included everything from the very urban Case Medical Center to suburban hospitals to two critical access hospitals made the challenge all the more intriguing — and potentially valuable as a model, Bieber says. "How do we leverage what we do to serve a heterogeneous social demographic over a complex geography? If we can do this, it may be portable."

Lots of Prep Time

As with many systems, UH Case's project began with its self-insured plan for 24,000 employees and their families. But knowing the model must be scalable and flexible, significant time was devoted to governance and architecture, Bieber says. In addition to physicians, system finance, contracting, legal and operations managers were tapped. "The advance work was done with lots of folks at the table. That has paid significant dividends as it has become more complex."

To accommodate the different needs of different populations, UH Case developed a flexible approach, resulting in formation of three ACOs — one for pediatrics, one for commercial patients and one for Medicare. Each is governed by a physician advisory council and relevant advisers, such as county public health officials, school board and Head Start representatives, and Medicaid managed care officials for the pediatric plan.

The three ACOs developed their own care management protocols using information sources and IT solutions tailored to their respective populations. The three report up through the umbrella ACO, which helps to ensure consistency of approach and accountability at the system level. But the structure is not frozen.

"ACOs are a vehicle today to get where we are going, but we have to be able to adapt," Bieber says, adding that building a culture of continuous improvement with a structure that allows nimble response is essential for long-term success.

The need to overhaul clinical practice, physician engagement and leadership is especially important — so much so that CMS requires extensive physician representation on ACO boards. Matticks believes that putting physicians in charge of developing the North Memorial network helped to overcome initial resistance to the ACO concept. Indeed, the more the doctors examined the facts, the more compelling the case became.

"In general in the United States, we have too many fixed assets, we are overbedded and we have an overabundance of specialists, and this is also true in the Twin Cities," Matticks says. "Because our foundational work was done by physicians, not the health system, I think we laid the groundwork for the difficult conversation we will need to have to get different outcomes from what we are currently getting." Doing so will be essential to survive in a market dominated by accountable systems, he adds.

Making the Numbers Work

Although it is also a physician-driven network, the market situation is reversed for Metro Health, Belsito says. "If we had 200,000 members in a capitated plan, we could really manage [an ACO]. But when membership is all over the place under different payers and payment structure, it's hard to make the numbers work."

Instead, Metro Health focuses on programs needed to manage in the current payment structure, Belsito said. This includes an all-patient, all-payer registry that facilitates care coordination and preventive care for all patients served by the system. The system's growing IT capabilities also support quality incentive contracting through the PHO.

Metro Health also has shifted from a hospital-based care strategy to one centered on 14 community clinics located one to 30 miles from the hospital, says Michael Faas, president and CEO. With patient satisfaction rising and regional payers getting behind the concept, he sees an integrated ACO-type arrangement in the future.

With a 208-bed hospital, Metro Health lacks the resources to go it alone. But rather than a threat, Faas sees that as an opportunity to build a best-of-breed system with partners including enlightened payers and the top long-term care, home care and physician services in the area. "We can begin to think of other combinations of physicians and other care with the goal of doing the best for the families and patients we take care of," Faas says. With accountable care, "you are getting incentivized through payment for what always should have been the goal of health care. It just highlights something we have always worked toward."

And as many systems — including those already formally operating ACOs — continue transforming operations toward a comprehensive ACO model, the model itself may be a transition to even more comprehensively integrated approaches, Cimasi says. "Managed outcomes, mutually beneficial partnerships with doctors, evidence-based care, transparency and proof of value to patients — these are all things we are trying to move toward as a society. Whether ACOs are the vehicle to get there, well, the jury is still out. We may be heading toward a national risk pool and ACOs may be the bridge that shapes the care continuum that lets us get there."

Howard Larkin is a contributing writer to Hospitals & Health Networks magazine.

AHA Recommendations for ACOs

The American Hospital Association in April sent a letter to Patrick Conway, acting director of the CMS Innovation Center, recommending changes to the Pioneer ACO Model and the Medicare Shared Savings Program. "While our members are committed to the concept of accountable care — many are pursuing public and private models — we continue to have significant concerns about the design of the current Pioneer ACO and Medicare Shared Savings Program," the letter stated. "The Pioneer ACO and MSSP programs place too much risk and burden on providers with too little opportunity for reward in the form of shared savings." To increase participation in the ACO programs, the AHA recommended the following changes:

  • improve the timeliness and accuracy of performance data;
  • extend the Track 1 agreement period;
  • set a standard minimum savings rate of no more than 2 percent, regardless of the number of attributed beneficiaries;
  • create more achievable financial thresholds in the early years;
  • implement technical adjustments to the benchmark to account for policy changes outside the control of provider;
  • allow beneficiaries to "opt in" to the ACO programs;
  • allow ACOs to vary beneficiary cost sharing;
  • simplify and align quality measures, and set the required thresholds prior to the performance year.

To read the entire letter, go to