Explaining Sluggish Savings under Accountable Care

By Valerie A. Lewis, Ph.D., Elliott S. Fisher, M.D., M.P.H., and Carrie H. Colla, Ph.D.
The New England Journal of Medicine, November 9, 2017

Despite aggressive targets set by Medicare for the spread of value-based payment arrangements and widespread agreement on the importance of delivery-system reform, progress toward lower spending growth and a transformed delivery system has been slow. Accountable care organizations (ACOs) are a prime example: nearly 1000 organizations operate as ACOs, but they have generated limited savings.

Myriad context-specific factors affect the performance of individual ACOs, but few characteristics have been shown to systematically explain ACO performance. We identified four cross-cutting explanations for the failure of ACOs to achieve savings right away (or at all) — two economic and two organizational. Economic explanations emphasize the breadth and depth of financial incentives in ACO programs, whereas organizational explanations emphasize organizational processes and complexity.

Economists worry about weak incentives. In the Medicare Shared Savings Program, nearly all participants opted to receive a bonus if they generate savings but bear no financial responsibility for losses. Transitioning providers to sharing in downside risk may result in greater behavior change and savings, although the literature on this subject is mixed. A second explanation suggests that having more patients covered by ACO-like contracts may strengthen incentives. Providers with few patients covered by ACO-like contracts face a problem colloquially described as “having a foot in two canoes.” Initiatives that generate savings in the care of a provider’s ACO patients will reduce income from its fee-for-service patients — an effect that led some integrated networks to fail during the managed care era. Until providers reach a tipping point in the number of patients covered by risk-based contracts, it will be hard for organizations to generate substantial savings.

We believe that all ACOs would benefit from having more patients covered by ACO-like contracts, something we routinely hear from providers. Lack of downside risk may help explain the performance of integrated delivery systems: complex existing systems might require stronger financial incentives to change their behavior. In contrast, the increasing proportion of physician-group and hospital-coalition ACOs achieving savings over time suggests that sharing in downside risk may not be necessary for these types of ACOs to generate cost savings if they are given enough lead time.

Economic explanations suggest that the key issue is motivation: providers know how to save money, but they need financial motivation to change their behavior. An alternative set of explanations draws on organizational theory and literature and suggests that providers either don’t know how to achieve savings (at least initially) or that they have to focus on other things — such as organizational, start-up, and compliance work — before they can implement the changes necessary to save money.

One organizational explanation for the progress that different kinds of ACOs have made toward generating savings relates to knowledge: when joining an ACO, independent physician practices with little knowledge of new care models may have ideas about how to save money, but they may need time to work out what to do and how to do it. Such groups may need to explore care management models or learn how to improve care transitions. This explanation may be particularly apt for outpatient-physician–practice ACOs and some hospital-coalition ACOs that are new to the game.

Another organizational explanation is that the complexity of ACOs affects their performance. First, most ACOs are not preexisting organizations; they are collections of independent providers, such as a community hospital and local private-practice physicians who decided to pursue an ACO contract together. These new organizations must accomplish a great deal of foundational work, such as forming a board; determining how to share bonuses; and completing tasks that require internal coordination, such as reporting quality measures. Then they must agree on and implement a joint strategy for pursuing cost savings. In these ACOs, we expect that progress toward generating savings may be delayed, because setting up a functional ACO will initially require attention and effort.

Second, organizational complexity probably challenges integrated systems. Although all ACOs must learn how to change ingrained practices among clinicians, the large, complex, and diverse networks of providers in an integrated delivery system present an added challenge for clinical transformation. Overcoming the inertia in these systems may be like trying to turn a large battleship. In contrast, outpatient-physician–practice ACOs are small and have a simple organizational structure, which probably makes them more nimble and able to more quickly implement (or discard) new initiatives — and generate savings. Newly formed coalition ACOs must spend time developing organizational infrastructure, but these groups aren’t encumbered by existing organizational processes.

We believe these explanations support taking a nuanced approach to encouraging, motivating, and rewarding providers. For all providers, moving a greater share of patients to value-based payment arrangements may support efforts to transform care and remove the barrier of conflicting incentives. Larger, integrated systems may benefit from a wider breadth of incentives and from stronger incentives that include sharing in downside risk. Shifting too quickly to risk sharing for newly formed ACOs, however, could backfire: among a subset of ACOs, it may be critical to allow providers enough time to settle into new organizational forms and develop and implement strategies in a deliberate and thoughtful manner. ACOs are as diverse as the U.S. health care system, and saving money in health care is notoriously hard. Developing policy approaches that accommodate this diversity will be important for payment and delivery reform to achieve its potential.


Elliott Fisher developed the concept of “accountable care organization,” proposing that health care providers across a continuum could improve quality and more effectively control costs by sharing accountability for patient care. It sounded like a reasonable approach and it rapidly gained traction as various entities, with public and private support, implemented them even though there was little understanding as to exactly what they were.

There are now a plethora of reports, including this one, that confirm that, whatever they are, they are smoldering along and really not delivering on their promise of higher quality at lower cost. Most of these reports also suggest that the grand experiment should continue because “there must be a pony in there somewhere.” Unfortunately, they are largely ignoring the diversion of effort and resources to these entities while not accounting for the greater economic losses from the administrative excesses nor for the impairment resulting from an increase in physician burnout.

This certainly does not mean that we should discontinue efforts to improve quality and reduce wasteful spending, but there are more effective ways to do this. One of the more important measures would be to enact a well designed single payer national health program. The administrative savings alone is far more than ever could be achieved through any iteration of an accountable care organization. Also, establishing a health care financing system that places the patient first creates a financing infrastructure that automatically drives quality.

Dr. Fisher and his colleagues deserve credit for their ongoing efforts to encourage changes that would improve value in health care through higher quality while reducing wasteful services. But we have enough experimental data now that we should no longer defer enactment of single payer merely because we hope that accountable care organizations will fix our system. They will not. Enough minds have cranked out enough ideas that we know that there is no magical accountable care infrastructure in our future that will produce the dramatic changes we all wish would occur.

Let’s get on with enacting an improved Medicare program that covers everyone. Under such a system we can continue to strive for higher quality and reduce care that is not beneficial, but in a system that brings greater satisfaction to patients and to their health care professionals.

Stay informed! Visit www.pnhp.org/qotd to sign up for daily email updates.