On May 4, 2015 Department of Health and Human Services (HHS) Secretary Burwell announced that the Pioneer ACO program had saved the federal government $384 million and improved quality in its first two years and would therefore be expanded. HHS also released a 130 page independent program evaluation by L&M Policy Research that served as the basis for the Centers for Medicare and Medicaid Services (CMS) Actuary’s certification of the Pioneer program.

Burwell’s triumphant announcement was an intended shot in the arm for the troubled Pioneer ACO program, 40 percent of whose initial 32 members dropped out in the first two years. It also illustrated the yawning reality gap between DC policymakers and the provider-based managed care community. In reality, the Pioneer program badly damaged CMS’s credibility with the provider-based managed care community and sharply reduced the likelihood that the ACO will be broadly adopted.

If the Secretary’s goal of having 50 percent of regular Medicare’s payments come through “Alternative Payment Methodologies” by 2018 is to be met, that growth is unlikely to come from either the Pioneers or the larger Medicare Shared Savings Program (MSSP). Table I below shows why.

A Flawed System Rewarded Those In High-Spending Areas, Not Good Performers

As with the Physician Group Practice (PGP) Demonstration that preceded the Pioneers, savings among the Pioneers are highly concentrated in a fraction of the participants. According to the L&M report, the top eight performers saved Medicare $295 million, 78 percent of the consultant report’s claimed savings! In the PGP demo, two of the ten participants (Marshfield and the University of Michigan) generated almost 70 percent of the savings.

Table I: Pioneer ACO Top Savers


But look at what CMS paid out in performance bonuses to their biggest Pioneer savers: $295 million in savings generated only $31.4 million in bonus payments. Rewards were so meager that four of the eight top performers dropped out of the program, a fifth opted to defer calculating their bonus until the end of year three, and a sixth (Atrius) earned no bonuses in either of the first two years despite generating over $36 million in savings.

Table II: Pioneer Managed Care All-Stars


When one looks more broadly at the Pioneer cohort, the fifteen mature clinical enterprises with at least twenty years of managed care experience fared poorly (see Table II). Most of these managed care All-Stars were not strangers to managing the risk of Medicare patients, either sponsoring their own fully insured Medicare Advantage (MA) plans or contracting with MA carriers on a full risk basis.

Seven of the fifteen All-Stars dropped out of the Pioneer program after two years. The All-Stars earned only 7 bonuses in 30 possible program years and were paid a paltry $20.2 million despite L&M attributing over $185 million in savings to them (Table II). Healthcare Partners, which owned three of the Pioneer franchises (NV, CA, and JSA) dropped out of the Pioneers, but was purchased by DaVita in 2013 for $4.4 billion, a market validation of the strength of their model. (We added University of Michigan to this cohort because they were a surprise success story in the aforementioned PGP demo. University of Michigan was unable to replicate its PGP success in the Pioneer program).

The explanation for this curious outcome is that CMS used prior years’ spending benchmarks for calculating bonuses rather than comparisons of actual spending by the Pioneers to local non-participating beneficiary spending as in the L&M consulting report. The benchmarks CMS used penalized mature managed-care organizations operating in low utilization markets. Acres of low hanging fruit (meaning lots of previously unexamined and uncontrolled health care utilization) seemed to be a precondition for success in the Pioneer program. Metropolitan Boston, one of the nation’s lushest “cherry orchards,” generated 42 percent of the estimated savings for the entire Pioneer program, according to the L&M analysis. And the plummeting savings from year one to year two of the Pioneer program ($280 million to $104 million) raised questions about how rapidly the participants ran out of accessible fruit.

Implications For The Medicare Shared Savings Program

The performance of the Pioneer cohort has significant implications for the broader Medicare MSSP program, which has ten times as many participants. Most of the Pioneers were selected because they already had built and successfully used the expensive information technology (IT) platforms and utilization management systems essential to managing the care of older people.

New provider groups desiring to participate in the MSSP program faced millions in infrastructure spending to set up their ACOs and millions more in annual operating expenses, as well as the usual execution risks and losses associated with developing and operating a new business model. In 2011, the American Hospital Association estimated that an ACO with a single, 200-bed hospital could expect to spend $5.3 million in set-up costs, and then incur $6.3 million in annual operating expenses to participate in the MSSP. For a 1,200-bed, five-hospital system, set-up costs were estimated to be $12 million and annual operating expenses to be $14 million.

In recognition of the steep learning curve, MSSP was set up to provide three years of so-called “one-sided” risk (meaning no losses incurred by participants if they missed their performance targets) before flipping to real (e.g. “two-sided”) risk. This one-sided risk phase was intended as a “training wheels” period for the newer ACOs before they were exposed to financial penalties if they overshot their spending targets.

If managed care All-Stars were unable to generate bonuses reliably in the Pioneer program, this suggests that the capital and operating risks for the inexperienced participants in the Medicare MSSP program far outstripped potential rewards.

This appears to have been the case. The MSSP program’s financial performance, with 220 participants, was even more highly concentrated than with either the Pioneers or PGPs, with 3 percent of the participants generating 50 percent of the savings and 15 percent generating 85 percent of the savings. That means that the other 190 MSSP participants generated virtually no return on the hundreds of millions they spent on set-up costs and operating expenses for their ACOs.

Provider Economics Matter

A little context is important here to complete the picture. Based on our experience, the average American hospital in 2015 only covers about 90 percent of its expenses incurred in treating Medicare patients. So to participate in the Medicare ACO program, providers are being asked to spend many millions in capital and operating expenses for perhaps a one-in-six chance of reducing their Medicare losses by 1 or 2 cents on the dollar of actual spending.

That’s not a very appealing risk/reward relationship. The economics of Medicare’s ACO program greatly resembles Tom Sawyer’s famous fence painting project, where Tom talked his friends into paying him to let them paint his fence. It is telling that, thus far, CMS has not released analysis of the actual set-up and operating expenses of their ACO cohorts that would enable independent estimates of the return on investment experience so far.

Unless participation in the ACO program is made mandatory, which would provoke a firestorm of reaction from hospital and physician communities, it is unlikely that providers who do their homework will join future versions of this program in significant numbers. In a survey last fall, only 8 percent of the MSSP participants indicated that it was likely that they would renew their participation in the program as then configured. Problems extended well beyond benchmarks and financial losses to data quality and timeliness, inadequate risk adjustment methodology, overstretched and inexperienced staff, and constantly shifting policies. The newly issued final MSSP regulations are unlikely successfully to have assuaged these concerns.

There was in the May 4 HHS press release a sunny obliviousness to provider economics. CMS Innovation Chief Patrick Conway commented: “This success demonstrates that CMS can design and test innovative payment and service delivery models that produce better outcomes for the Medicare program, and beneficiaries.” If the models don’t make business sense for providers, they will never scale to the entire health system.

What Should HHS Do Now?

Medicare shared savings is not a new idea. CMS has been testing it for a decade, beginning with the Physician Group Practice demonstration in 2005. It is a reasonable forecast given the past decade’s experience that the ACO is not going to be a viable total replacement option for the regular Medicare program. A fall back position would be to leave the ACO as a contracting option for provider organizations in the high-cost Medicare markets, e.g. letting provider organizations compete to lower Medicare spending in their areas.

This would return the program to the original policy objective for the ACO proposed in Health Affairs by Elliott Fisher and his Dartmouth colleagues in 2007 — not containing cost but rather reducing variation. Alternatively, CMS could offer assistance to the handful of breakout superstars in the early ACO years (Memorial Hermann, Beth Israel Deaconess, Steward, Montefiore, etc.) to enter the full-risk Medicare Advantage market where they get to keep 100 percent of the savings they generate and share them with beneficiaries in the form of lower out-of-pocket expenses.

With over 17 million Medicare beneficiaries voluntarily choosing MA thus far, and enrollment growing at more than 10 percent annually despite three years of CMS payment reductions in real dollars, it is increasingly clear the future of managed Medicare lies in the MA program, not with directly contracted shared savings models. More thought should be given to how to capitalize on MA’s expansion to save money both for the Medicare program and for beneficiaries.