Replacing Medicare ACOs with a Better Integrated Care Option

The Affordable Care Act (ACA) is best known for its provisions aimed at expanding insurance coverage. But, as former president Barack Obama and others often noted while pushing for its enactment, the law’s authors also wanted to slow the pace of rising costs—or “bend the cost curve,” as they put it. Among the more high-profile of the law’s provisions aimed at cutting costs are those establishing accountable care organizations (ACOs) in the Medicare program.

The ACO effort, along with other “delivery system reforms” in the ACA, is based on the correct view that Medicare has substantial leverage over the workings of the entire medical care delivery system in the United States. Medicare is the largest payer of medical claims in most markets because of the high use of services by the elderly. Consequently, the nation’s vast network of hospitals, clinics, physician practices, labs, hospices, and other providers take their cues heavily from the financial incentives embedded in Medicare’s complex payment rules and regulations. Those promoting Medicare ACOs want to change the signals being sent by the Medicare program to promote more efficient models of care delivery. Their hope is that efficiency gains from widespread enrollment in Medicare ACOs will lead to higher-value care for other insurance enrollees, too.

While well intentioned, the ACO concept is unlikely to meet the lofty goals set for it because it relies too heavily on regulation by the Centers for Medicare and Medicaid Services (CMS) and not enough on competition and beneficiary choice.

What Are ACOs?

Since the enactment of the program in 1965, Medicare’s dominant model for providing coverage to enrollees has been fee-for-service insurance (FFS). Under FFS, physicians and other providers submit claims for reimbursement for the many different kinds of services rendered to patients. Over the course of a year, CMS makes more than 1 billion payments for FFS claims. It has become clear that building Medicare on a foundation of this kind of insurance—the prevalent model back in 1965—has resulted in fragmented and disorganized service delivery for many millions of senior citizens. Employers long ago moved away from the unmanaged FFS model for their employees, but Medicare has been much slower to change because many of the program’s key features are written in law.

In the early 1980s, Congress did create an option for beneficiaries to enroll in a private insurance plan instead of FFS, which is now called Medicare Advantage (MA). Most MA plans are HMOs (health maintenance organizations), and thus “manage care” on behalf of their enrollees by contracting with a participating network of hospitals and physicians. In 2016, nearly one-third of all Medicare beneficiaries were enrolled in MA plans.

But enrollment in an MA plan necessarily means leaving the traditional, government-managed Medicare program for enrollment in a private insurance plan. Some policymakers have wanted to give Medicare enrollees the benefits of managed care within the structure of the traditional program. ACOs are intended to serve that role.

The authority for ACO participation in Medicare has two sources. The first is the provision of the ACA that created the Medicare Shared Savings Program (MSSP). CMS has issued extensive regulations laying out the requirements for ACO participation under the MSSP authority. There are now three different tracks for MSSP ACOs. In Track 1, the ACOs assume no financial risk for missing their financial performance and quality goals but can earn a bonus for meeting or exceeding them (the so-called one-sided model). In Track 2, the ACOs assume financial risk, which means they can be penalized financially for missing their benchmarks for spending or quality of care. They also have the potential to earn somewhat higher bonuses based on their performance. In Track 3 there is also financial risk, and there is a new system of designating participating beneficiaries in the ACO prospectively rather than retrospectively. Track 3 ACOs have the potential to earn higher bonuses than those in Track 2.

In addition to the statutory MSSP, CMS also initiated, through its demonstration authority, a “Pioneer” ACO program. The Pioneer program was intended to provide an accelerated approach for experienced integrated care systems to begin managing the care of Medicare beneficiaries at a faster pace than was expected to occur with the MSSP. The expectation was that the Pioneer ACOs would move rapidly toward some kind of “population-based” or per-enrollee payment.

More recently, as the Pioneer program experienced substantial plan withdrawals, CMS initiated a new demonstration program, dubbed the “Next Generation” ACO model. This model is also for more advanced integrated care systems and was designed to address some of the flaws identified by the original Pioneer participants. In particular, Next Generation ACOs are allowed to contact beneficiaries to confirm their participation in the ACO, although beneficiaries retain the right to see any qualified provider they want, including those not aligned with the ACO. Beginning in 2017, Next Generation ACOs can also opt to receive payment in a modified capitation arrangement that breaks more cleanly from the FFS payment model.

The Early Results for ACOs Have Been Mixed, at Best

ACOs have not produced the kind of results the ACA sponsors had hoped they would. When the Pioneer ACO program began in 2012, there were 32 participating ACOs. The program closed at the end of 2016 with just eight participating plans. Plans dropped out of the program for several reasons, including benchmarks that the plans say penalized regions with already low costs, inadequate risk adjustment, and flaws in the design of the program that hindered the ability of the ACOs to adequately manage the use of services by ACO participants.

The MSSP ACO program has grown since its launch, and there are now a total of 433 participating plans. But the vast majority—411—of these ACOs are in Track 1, where they are eligible for bonus payments and cannot be penalized. Only 22 of the MSSP ACOs are in Tracks 2 or 3. Providers seem willing to participate in ACOs so long as there is no financial risk. Very few providers have been willing, at this point, to take on the responsibility of managing the care of the ACO enrollees within a fixed budget.

In August 2016, CMS announced program results for the MSSP ACO program for 2015. Overall, about half of ACOs were able to keep spending for the Medicare beneficiaries assigned to them below the benchmarks set by CMS, and the other half spent more than their target. The net savings was just $429 million (less than 0.1 percent of total Medicare spending). Moreover, the ACOs that showed savings were concentrated in regions with high per capita costs in the traditional FFS program. In fact, the ACOs that were eligible for bonuses actually spent more, on average, per beneficiary than those that failed to meet their targets. Finally, when the bonus payments were taken into account, the overall MSSP ACO program increased Medicare spending, rather than decreased it.

Some studies have shown that the savings from ACOs are greater than can be assessed by looking strictly at the cost experience for the beneficiaries assigned to the ACOs. This is because physicians and other providers of services adjust their practice patterns for all their patients, not just for those assigned to an ACO. Still, even when some potential “spillover” savings is included in the calculation, the overall savings from ACOs have been, at best, very modest to date.

The Fundamental Problem Is Assignment and Passive Beneficiary Enrollment

The fundamental problem with the ACO model, both as constituted in the ACA’s MSSP and as established through CMS’s demonstration authority, is that it tries to avoid direct engagement with the beneficiaries. Under the various versions of ACOs put forth to date, the beneficiaries are not given the opportunity to explicitly enroll in an ACO as they would in an MA plan. Instead, they are “assigned” to ACOs based on their use of physician services. Medicare administrators comb through the massive Medicare claims database, and any beneficiary whose primary doctor has joined an ACO is, by default, considered to be an enrolled member of the ACO, too.

Medicare’s program administrators have reiterated repeatedly that the ACO program is designed this way because the beneficiaries are not being asked to leave the traditional FFS Medicare program, and thus the beneficiaries are not giving up any of their rights under traditional Medicare. Most importantly, this means that Medicare beneficiaries who are assigned to ACOs are allowed to see any physicians they want to, even those who have no relationship to the ACOs.

The architects of the ACO program thought beneficiary assignment to ACOs would be a way to grow managed care within Medicare rapidly. Physicians and hospitals would choose to join ACOs because of the potential for bonus payments, and their patients would come along with them automatically. There wouldn’t be the need for any explicit beneficiary enrollment process because the beneficiaries wouldn’t lose (or gain) anything financially from enrollment in an ACO.

But it is not possible to run an effective managed care plan without some level of active engagement by the enrolled population. ACO results have been disappointing to date because of the explicit unwillingness of the authors of the MSSP to allow ACOs to use the same tools that integrated care systems use in the private sector to control the delivery of services to patients. Integrated care systems serving employer-sponsored coverage are not open network plans; for the most part, their enrollees are required to stay within the integrated care system to receive full coverage of their services. When beneficiaries receive services outside integrated networks, they generally must pay more of the bill themselves. That’s how managed care plans are able to get control of the care process and eliminate waste and inefficiency.

The use of beneficiary assignment to ACOs has created the illusion of large enrollment in the program, but the truth is that many of the beneficiaries assigned to ACOs are not aware that the assignment has occurred. They may have received a communication from their ACO stating that they have been assigned to it, but receiving such a communication is not the same thing as giving consent to the ACO’s procedures. The absence of a genuine enrollment process means the beneficiaries have little commitment to the ACOs they have been assigned to. This is why many of the participating ACOs experience substantial “leakage”—that is, many of the beneficiaries assigned to the ACOs see specialists outside the ACO network.

Because many of the physicians participating in ACOs are not entirely sure who among their Medicare patients are in their ACO, it is difficult for physicians to take the necessary steps to control their patients’ overall cost of care to meet the ACO’s performance goals.

Beyond the lack of an effective beneficiary enrollment system, there are other problems with the current ACO program. These problems include an extensive series of data reporting requirements that make it an overly expensive undertaking for providers, financial benchmarks that favor higher-cost areas, and significant delays in the availability of useable data for improving ACO performance. These shortcomings have also contributed to the program’s lackluster performance during its early years of implementation.

An Effective Beneficiary Choice Model

Medicare already provides beneficiaries with a choice of coverage. They have the option to enroll in an MA plan or remain in the traditional program. The ACO program should be modified to become another option that beneficiaries may select during initial enrollment into the program and annually, alongside the MA plans and unmanaged FFS. Amending the ACO program to meet this goal would include several steps.

Relax Certain Regulations

Hospitals and physicians should be allowed to form provider-driven managed care networks that do not require an insurance license. These plans, perhaps renamed Medicare provider networks, would be the successors to the current ACOs. They would be required to meet certain financial and quality standards to participate, but the extensive requirements placed on ACOs would be scaled back.

Direct Price Competition with MA Plans for Beneficiary Enrollment

Medicare provider networks would compete directly with the MA plans and unmanaged FFS for beneficiary enrollment. The networks would be paid using today’s FFS payment rules, but the payments would be made to the network administrator, not the individual providers of services. The network administrator would then have the authority to distribute payments to its participating providers according to its own payment formula. Medicare provider networks would be required to provide to program administrators the estimated cost reduction amounts that the networks anticipated for the year and the amount that would be shared with the beneficiaries in the form of reduced Medicare premiums to encourage their enrollment in the plan. The reduced premiums paid by these beneficiaries would be deducted from the aggregate FFS payments paid to the networks during the course of the year. Thus, the government would be assured it would spend no more than it does under current law. If a network could cut costs well below unmanaged FFS, it could offer a substantially lower premium to the beneficiaries compared to unmanaged FFS and thus increase its enrollment. Medicare provider networks that promised to hold down costs but then failed to do so would suffer financial losses. The networks would need to compete for beneficiary enrollment with MA plans that also offer lower cost sharing and higher benefits compared to unmanaged FFS.

Reform of Supplemental Insurance

Beneficiaries enrolled in Medicare provider networks would also be allowed to secure supplemental insurance beyond Medicare that fills in most of their cost sharing, while beneficiaries opting to remain in unmanaged FFS would face some restrictions on what kinds of supplemental plans they could purchase. In particular, those in unmanaged FFS could not acquire supplemental coverage that entirely filled in the deductibles and cost sharing required in traditional Medicare. The law was recently amended to prohibit coverage of the deductible for physician and other outpatient services, starting with purchases of such coverage in 2020, but much more should be done to ensure unmanaged FFS has effective cost sharing attached to it.

Unmanaged FFS insurance only makes sense if the enrollee in the coverage is required to pay something out-of-pocket at the point of service. Otherwise, there would be no limit to the use of care because the insurance plan (or plans, in case of Medicare plus supplemental coverage) would pay the entire bill for every service, no matter how expensive or unnecessary. A Medicare provider network, however, would more closely resemble an HMO in terms of its financial incentives. A network of this kind would suffer financially if it did not have effective control of the use of services by its enrolled population. It makes sense, therefore, to give the enrollees in these networks more latitude to lower their cost sharing by securing expansive supplemental coverage if they would like to, while restricting this choice to some degree for those enrolled in unmanaged FFS.

These reforms would ensure that low-cost, high-value, and provider-driven managed care options would be given ample room to compete and thrive in Medicare. If such plans could deliver better care at lower costs than the other alternatives currently available in Medicare, including MA plans, then they would be a very attractive option for Medicare beneficiaries. With actual beneficiary enrollment, Medicare provider networks would have the leverage that comes from committed enrollees to drive real and lasting changes in how health care is delivered to their patients.


The Medicare ACO program is unlikely to meet its goals because it is based on the flawed assumption that CMS can nudge the Medicare program toward higher-value care without directly engaging with the program’s beneficiaries. Use of services in the unmanaged FFS program is high in large part because the beneficiaries, especially those with expansive supplemental coverage, have little reason to not use services that a physician suggests might improve their health. The Medicare ACO program, as currently structured, does not alter the existing incentives in any substantial way, which makes it very unlikely that it will have a significant effect on program costs.

A more effective approach would make changes to the incentives beneficiaries now face so that they have sound reasons to enroll in lower-cost and higher-value arrangements. The results would be far better for the government—and for the beneficiaries, too.