“The accountable care organization is like a unicorn, a fantastic creature that is vested with mythical powers. But no one has actually seen one,” said former California HealthCare Foundation CEO, Mark Smith, MD, in 2010. Over the last four years, we’ve certainly seen a proliferation of unicorns and we’re now reaching the point where fantasy—at least in a handful of cases—is becoming a reality.

A growing number of large employers are piloting accountable care organizations (ACOs), working through their health plan; in some cases they are doing so directly with provider systems, such as the new Boeing arrangement with Providence Health and Services, Swedish Health Services, and University of Washington Medicine and Intel’s contracting efforts in Albuquerque, New Mexico and Portland, Oregon. The large employers and other health care purchasers with whom we work — eager, if not desperate, for solutions to contain the costs of health care and improve its quality — are watching these first movers carefully to see if ACOs prove to be a viable strategy for improving population health and bending the cost trend.

These leading purchasers intend to set the bar high. They cannot make the investment to pursue these ACO relationships if they are not assured that their populations will see meaningful, measurable gains in their health care and its affordability, as well as their health. That often means contractual commitments to lowering total costs of care and showing improved patient outcomes for targeted populations — like high risk, medically complex patients. Our purchaser colleagues who have been among the early adopters of ACO arrangements have begun to identify the features critical to successful ACOs; these are the elements other purchasers will look for when deciding if it’s worth proceeding.

Lessons For Purchasers Considering ACOs

Catalyst for Payment Reform (CPR) and the Pacific Business Group on Health (PBGH) have captured these elements in a recently released Accountable Care Organization Toolkit for purchasers. The Toolkit contains a questionnaire for health plans about their efforts, and model contract language for health care providers wanting to serve as ACOs. Providers, health plan leaders, and other health care experts may find these lessons useful as well.

  1. Not just any network will do. When creating or entering into an ACO, employers should be choosy about the provider network, requiring that the ACO select providers based on quality and efficiency.
  2. Insist on outcomes measures. The whole point of trying an ACO strategy is to improve quality and reduce costs. An ACO must be able to generate outcomes measures that assess its impact on quality and costs.
  3. Patient-centered, coordinated care is critical. ACOs promise to improve care quality, in large part through better care coordination. A purchaser shouldn’t work with any provider system that can’t support targeted care management; physician engagement in shared decision making with patients is a key piece as well.
  4. Transparency is key for patients, providers, and purchasers. Part of the appeal of an ACO is that it is next generation health care. As such, it should provide cost and quality information to patients, preferably at the individual provider level. The same information can help providers gauge and improve their performance. And purchasers must absolutely insist on transparency regarding the spending, savings, and the savings distribution within the ACO.
  5. Information technology helps with all of the above, and then some. Online data administration is essential to help track care management and performance.
  6. The ACO movement will largely negate its own goals if it stifles competition. ACOs should not use contract exclusivity and/or provisions that prohibit transparency, thereby reducing competition within the market. Rather, ACO development and growth should advance competition among providers in the market.
  7. Finally, and perhaps most importantly, providers should have some financial “skin in the game.” To assess cost savings, ACOs should apply a prospective patient attribution model. As a result, the ACO will have a defined patient population for cost and quality measurement and targeted care coordination strategies. Ideally, ACOs will be subject to a shared-risk payment arrangement in which any savings will be distributed based on meeting medical cost and quality goals. Furthermore, ACOs should promote payment for value not volume, such that the performance incentives are passed through to primary and specialty care providers.

One of CPR’s working hypotheses is that the payment reform models that will prove to be most successful long term at reducing costs and improving care quality will be those that put providers at financial risk while requiring transparent outcomes reporting. These arrangements allow providers to reap rewards for delivering care more efficiently, but also penalize providers for being inefficient, or for providing poor quality care; in short, both the carrot and the stick are key to changing behavior and creating lasting positive change. CPR has discussed this concept in The Payment Reform Landscape series here on Health Affairs Blog throughout this year.

A More Gradual Glide-Path Option

Yet not every provider system wanting to create an ACO will be able immediately to take on shared financial risk. For those systems, CPR and PBGH propose a “glide-path” that allows purchasers and providers to enter into an arrangement in which shared financial risk evolves over time.

Step one can include an ACO-type arrangement where providers receive fee-for-service payments with shared savings. This initial phase should be brief, and the ACO should achieve a minimum savings rate (e.g. 2 percent) before payers distribute any savings.

As step two, the purchaser and provider network system embark on a shared risk arrangement, where the provider can earn savings — or lose funding — based on performance, and where the provider stands to bear the risks of spending over budget. Of course identifying and agreeing to performance measures is a huge part of this process. Our ACO Toolkit contains recommended performance measures that are outcomes-oriented and patient-centered.

Finally, the ACO should transition to a population-based-payment model. In this phase, the ACO should receive a population-based payment using a predetermined amount based on the ACO’s expected costs. Underlying provider payment structures may include bundled payments or case-management fees in place of traditional fee-for-service payments that can be modified or reduced over time. This design incentivizes the ACO to practice “population management,” caring for specific populations of patients (e.g. those with chronic illness) while focusing on quality and efficiency.

The ACO should be eligible for shared savings only if it reduces costs (for example, by at least 3 percent). As the ACO achieves savings goals, the proportion of risk shared can increase. The ACO’s ability to manage risk successfully should determine the pace of the transition from shared savings to full risk sharing.

If ACOs are truly going to offer a new way to deliver and pay for care, it is important for purchasers to set the bar high, with expectations for improvement in care quality and cost savings. Like everything in health care, we know change won’t happen overnight, but we look forward to seeing progress, with more ACOs emerging to meet the criteria we outlined above and willing to take on shared financial risk.