Uncategorized

Pioneer ACO’s Disappointing First Year

On July 16, the CMS Innovation Center reported the first-year results for the Pioneer Accountable Care Organization program:  13 Pioneers, or about 40 percent of the participants, earned bonuses. The program saved Medicare a gross $87.6 million before bonus distributions, cutting the rate of growth in Medicare spending by 0.5 percent, from 0.8 percent to 0.3 percent annually.

However, nine of the 32 members dropped out and press reports hinted at a contentious relationship between the Pioneers and a well meaning but green and overtaxed CMS staff.  It was not an auspicious beginning for a program whose advocates believed would eventually replace regular Medicare’s present payment model. There immediately followed a blizzard of spin control from ACO “movement” advocates stressing the need for patience and highlighting first year achievements.

What was irritating about the Pioneer spin is it treated the ACO as if it were a brand new idea with growing pains. This studiously ignores a burned out Conestoga wagon pushed to the side of the trail: the Physician Group Practice demonstration CMS conducted from 2005-2010. The PGP demo tested essentially the same idea — provider bonuses for meeting spending reduction and quality improvement targets for attributed Medicare patients.  The pattern of arrow holes and burn marks on the PGP wagon closely resemble those from the Pioneer’s first year, strongly suggesting more troubles ahead for the hardy, surviving Pioneers.

The PGP Precedent. Like the Pioneers, PGP participants were not ordinary community hospitals or freshly formed physician groups or IPA’s.  Rather, most were “high functioning” organized clinical enterprises, some with decades of global risk contracting or health plan operating experience.  Particularly in light of the degree of clinical integration and care management experience of its participants, the PGP results were extremely disappointing; only two of the ten participants were able to generate bonuses in each of the program’s five years, and one, Marshfield Clinic, earned half the total bonuses.  Managed care veterans like Geisinger Clinic and Park Nicollet earned bonuses in only three of their ten program years. Two other high-quality multi-specialty clinics had even rougher sledding, with Everett Clinic getting one year of bonus ($126,000) and Billings Clinic completely shut out.

The pattern in the first Pioneer year is remarkably similar.   While thirteen of the Pioneers earned bonuses, it appears from press reports that four of them generated 2/3 of the savings.   It is likely not coincidental that three of those four participants (Massachusetts General, Beth Israel Deaconess’ physician organization, and New York’s Montefiore) either run or practice at some of the most expensive hospitals in the country, in two of the country’s highest per capita Medicare spending markets.  Orchards full of low hanging fruit (e.g. very high levels of previously unexamined Medicare spending) appear to be an essential precondition of ACO success.


And just as in the PGP, the managed care veterans among the Pioneers got skunked.  The two most sophisticated physician-directed group practices among the Pioneers — HealthCare Partners in the West and Atrius Health in Massachusetts — generated no savings. Neither did Presbyterian Health System of Albuquerque, whose health plan has dominated its state for two decades.  Park Nicollet and Allina, both of Minneapolis and thirty-year capitation veterans, generated no savings.  University of Michigan, one of the surprise PGP successes, left the Pioneers after a marginal first year, as did Presbyterian and North Texas Specialty Physicians, the oldest and largest IPA in metro Dallas/Fort Worth.

Problem areas in the Pioneer model. Adverse selection combined with an inadequate risk adjustment methodology may have hurt these managed-care savvy Pioneers.  Organizations with great reputations for helping high-risk Medicare patients will differentially attract them.  One of the unfortunate learnings of the PGP was the vital importance of aggressive coding of patient acuity to offset selection effects.

And because they cannot proactively identify patients, particularly physician-centric Pioneer organizations will have trouble diverting them from hospitalization, or reducing the use of expensive out-of-network services.  It’s really tough to practice managed care without the patient’s knowledge or consent, or sharing some of savings with them — the fundamental flaw in ACO program design.

If the policymakers who decided to make the ACO the main event in payment reform had looked honestly at PGP’s results, they would have chosen differently.  Politics has played a prominent role in the ACO ‘movement”.  The ACO was blessed with enthusiastic, high-profile advocates such as Dartmouth’s Elliott Fisher and former CMS administrator Mark McClellan.  But crucially, left-of-center health policy types believed they’d found in the ACO a surefire shortcut to winding down Medicare Advantage and moving toward a single payer system.  Zeke Emanuel said as much in a New York Times blog posting:  “. . . thanks to the accountable care organizations provided by the new health reform act, a new system is on its way, one that will make insurance companies unnecessary.”

It is difficult to draw more precise conclusions about the first year Pioneer performance because all we have to go on is a press release, and a useful Health Affairs blog posting combing multiple press reports.  CMS has thus far refused to release the actual performance data on the 32 Pioneers (and also discouraged its participants from talking to the press).  No data on start-up expenses and on-going infrastructure costs was contained in the press release, so it’s impossible to tell what the ROI for participants is, so far, and how much savings they will have to generate to break even.

Another year of performance results from the Pioneers and first year data from the much larger cohort of regular Medicare Shared Savings Program participants would certainly be helpful.  But we’ve now got six full years of field testing of the Medicare shared-savings idea, and the prognosis for Medicare ACOs is not encouraging.  If the most sophisticated provider-based managed care enterprises cannot master this program, neither will most of the nation’s community hospitals, even with the help of an avid corps of consultants and policy cheerleaders.

And if regular Medicare costs continue growing at historic low rates (0.8 percent annual growth is hardly a burning platform), one might also reasonably ask how urgent cutting the nation’s rate of growth in Medicare spending is as a headline policy objective, compared to, say, markedly reducing Medicare expenditures in high-use areas, or controlling post-acute spending, the fastest growing and least well managed part of the health system.

Based on what we’ve seen so far, it is highly unlikely that the ACO is going to be a viable “total replacement” solution for fee-for-service Medicare.  While Shared Savings might function as a “Betty Ford Center” type payment option for high-cost hospitals or markets, episode-based payment for expensive acute services and “relationship-based” primary care payment models like the Patient Centered Medical Home, are more likely, if less comprehensive, payment reform options.  How to strengthen the role physicians might play in reducing unnecessary hospitalizations and post-acute costs is also an urgent policy question.

The growing Medicare Advantage program. Meanwhile, real live Medicare beneficiaries with a choice are choosing the politically incorrect Medicare Advantage program, whose enrollment continues growing at 10 percent a year and will exceed 15 million members by the end of calendar 2013. More than 40 percent of each cohort of baby boomers aging into Medicare, including this author, are selecting Medicare Advantage, because it’s a much better deal financially and logistically than regular Medicare.

Not only is Medicare Advantage not withering away, as the CBO incautiously predicted in 2009, it may eventually be half the program.  Most of the risk shifting in Medicare is going to take place in the MA program, mediated by health plans and their contractual relationships with providers.  Of course, providers already have the option of contracting on a full risk basis through MA.  Encouraging more provider-sponsored health plans would also help with the needed shifting of risk inside Medicare.

In health policy, numerous “silver bullet” solutions have had a transient high concept appeal only to be revealed later not to work.  Solving Medicare’s problems is most likely going to require doing a dozen smaller things right, rather than one big thing.  Let’s get on with the task of finding those dozen smaller things, and hope that evidence, not politics or wishful thinking, will show us the way forward.

Jeff Goldsmith is president of Health Futures Inc, which specializes in corporate strategic planning and forecasting future health care trends.

Goldsmith, Jeff. Pioneer ACO’s Disappointing First Year, Health Affairs Blog, 15 August 2013. Copyright ©2013 Health Affairs by Project HOPE – The People-to-People Health Foundation, Inc.

15 replies »

  1. Good to hear that, the reaction to the rule was overwhelmingly negative. With the exception of some consumer groups, most comments expressed extreme disappointment in the proposed ACO rule as written and called on CMS to perform major revisions.

  2. Great post Jeff; informative and insightful. I’ve been skeptical of the ACO hype for a while. Also, especally appreciate the way you connect the back-end policies and structure to what happens on the front-end between clinicians and patients.

    About this phrase: “One of the unfortunate learnings of the PGP was the vital importance of aggressive coding of patient acuity to offset selection effects.”

    As a geriatrician, I feel that a similar thing happens to me as an individual provider. More complex patients are likely to select a geriatrician. Even if it were feasible for me to code to match the acuity, wouldn’t I have to invest in aggressively coding acuity in order to cover my time and efforts for these patients? Any thoughts on how to facilitate the match between a patients’ needs and the providers efforts?

  3. ACO’s are, for the most part, managed care without the risk and without the patient’s consent and sharing in the savings. Those are crucial design flaws. The risk (that is, a global budget comprised of the premiums of enrolled members) is what forces providers to make uncomfortable choices and restrict or eliminate low value care. The patient’s consent is crucial because so much of health cost, and the health risks that generate them, are based on the patient’s behavior and co-operation with treatment.

    ACO’s merely layer another set of reporting requirement on top of the existing fee for service system. The core incentives to do more and make more are playing in the background. Some win and some lose under those incentives, and the shared savings rewards are too small and too late (e.g. 18 months later) to have a material effect on provider behavior.

    And as many argued in the comment thread to Painter’s post, what REALLY drives provider behavior are values, not incentives. You can tinker with the incentives all you want. If you don’t alter the core values that motivate, particularly, physician behavior, you change nothing.

    If you want to see two contrasting views of this issue, see my debate with Tim Ferris of Mass General at Connected Health a couple of years ago: http://healthcare.partners.org/streaming/CCH/symposium2011/Imperial/ImperialThursday14.html

    Provider health plans are needed because of the high degree of consolidation of many regional health insurance markets. They differ meaningfully from ACOs in that they must compete with health plans, and therefore contract for and manage care at risk.

    I want my doctor to work for me, and with his/her colleagues, not be a pawn at the end of an asymmetrical commercial health plan network contract. I’d rather give my money to a provider sponsored plan, particularly if I already use their services, than to a health insurer. Some of the finest health plans in the US are sponsored by or work in tight collaboration with physician groups.

  4. Thanks for your response, Mr. Goldsmith. Glad to be able to learn from you, and I look forward to reading your blueprint when I get a moment to do so.

    Dr. Painter’s article is excellent, but I’m surprised to see you cite it next to your recommendations above (also excellent, by the way), which are… payment reforms to realign incentives.

    It seems inconsistent to say that the ACO payment reforms “don’t do it,” as you put it, but then to recommend policy changes that are payment reforms that somehow “do it.” What am I missing?

    Fundamentally, though, none of your recommendations seem incompatible with ACOs, are they? Provider-sponsored health plans are kind of like the most basic form of ACO, and the payment reforms could be incorporated without issue. Am I looking at this wrong?

  5. Good points all around, but certainly agree that the current resource requirements for playing in the burgeoning ACO game are daunting and limiting as Goldsmith describes. As an example, to overcome the lack of PGP effectiveness due to missed severity coding, armies of “documentation experts:” or overlay software continue to add-up, while detracting from the fundamental focus on cost-containment (presumably to get 10x worse with ICD-10).

    Accountability in cost and quality as espoused in theory IS the answer, but agree that focus on some of the ideas describes by Goldsmith above make sense as a logical start.

  6. Talked about this at the end of the posting.

    – We need more provider sponsored health plans that take full risk capitation and participate in Medicare Advantage (e.g. managed care that beneficiaries actually choose).

    – We need to pay primary care physicians and some specialists who manage life changing clinical conditions like diabetes, mental health, congestive heart failure on a subscription basis rather than per visit/per test, lowering their overhead and encouraging long term relationships with patients

    -We need to purchase expensive clinical interventions on a “complete clinical solution” basis- an all-inclusive, severity adjusted payment that covers all clinical and hospital services including recovery/rehabilitation. Patients need multiple provider clinical options for expensive care (cardiac intervention,
    cancer care, most surgery), and should be able to avoid out of pocket expenses or even receive cash payments for selecting high value choices.

    These latter two options substitute single payments (either per month or per episode) for the present FFS payment framework, reducing administrative costs and complexity, and returning time to clinicians for patient care. I’d leave the unpredictable stuff, including emergency care and diagnostic visits on a fee basis with copays graded to income.

    For a look at these policy options in greater detail, please look at the Policy Blueprint I produced for the Physicians Foundation this spring: http://www.physiciansfoundation.org/uploads/default/PF_Blueprint_Report_-_May_2013.pdf

  7. The take home message for me – a practicing physician – is that if big, experienced organizations with large IT staffs, etc. etc. can’t make this work, there is no hope for me.

  8. The ACO has been the main event in payment reform, and blossomed into a billion dollar industry. It was flawed in its conception as well as execution.
    It represents a huge diversion of energy from the main policy challenge we face: significantly lowering the cost of healthcare. As we’ve suggested, lowering the rate of growth of Medicare spending is not the right policy target.

    It isn’t a straw man argument. It’s been promoted as the replacement for present Medicare FFS payment. As exhaustively and brilliantly argued in a recent THCB post and thread, realigning incentives “doesn’t do it”. https://thehealthcareblog.com/blog/2013/08/06/aligning-physician-incentives-doesnt-do-it/ This post is worth reading before accepting what “most say”, as you put it. Unfortunately, the crowd is wrong.

    Docs and hospital execs are not laboratory rats, and our health system is not suddenly going to become affordable by finding the right operant conditioning schedule, particularly one that imposes so many additional documentation requirements and associated costs. We need to put our policy energy elsewhere.

  9. Mr. Goldsmith, thanks for your perspective. You’re referring to “Pioneer” spin as if you’re not spinning twice as fast in the other direction.

    Almost half of Pioneer ACOs shared savings, and all improved quality. Sure, nine are leaving the Pioneer ACO program, but seven of those are leaving it to join the Medicare Shared Savings Program; they’re downgrading, not dropping out. I think those are much more auspicious results than you’re highlighting.

    And casually dismissing a 0.5% drop in the growth of Medicare spending is cavalier. It may seem small but, because it compounds over time, ends up being quite significant.

    You knock ACOs for being “low-hanging fruit pickers.” But picking previously-unpicked LHF is a good thing, right? It was left there for a reason (because the previous incentives, like FFS, gave hospitals reason not to pick it), and it was picked for a reason (incentives changed, and it made sense to pick).

    It’s a straw-man argument to insinuate that ACO supporters believe they are the last great hope for health care cost reduction. Most say that realigning incentives for providers holds the promise of increasing quality and reducing cost.

    The program accomplished both. Not so disappointing, in my opinion.