By Kip Sullivan, October 2010
The “accountable care organization” (ACO) is the latest fad in American health policy. It remains an unknown concept to the vast majority of the public, including most doctors, but it is all the rage among health policy analysts as well as lawmakers who sit on heath policy committees in Congress and in state legislatures.
Although the assumptions used by ACO proponents to justify ACOs have been around since the dawn of the HMO movement, the ACO label is relatively new. It was invented late in 2006 during a discussion at a public meeting of the Medicare Payment Advisory Commission (Medpac). The seminal article announcing the concept appeared in December 2006. By 2009 the ACO had become so fashionable among congressional Democrats it was mentioned in all three draft health care “reform” bills prepared by Democrats during the first half of 2009 (two of those bills originated in the Senate and one, the Tri-Committee bill, was written in the House). The ACO movement’s crowning achievement to date is the inclusion of ACO provisions in the final “reform” legislation – the Patient Protection and Affordable Care Act (PPACA) http://frwebgate.access.gpo.gov/cgi-bin/getdoc.cgi?dbname=111_cong_public_laws&docid=f:publ148.111.pdf
– which President Obama signed in March 2010. These provisions, which appear in Section 3022 of PPACA under the title, “Shared savings program,” authorize the Centers for Medicare and Medicaid Services (CMS) to create an ACO “program” by no later than January 1, 2012.
The definition of “ACO” bears a striking resemblance to the definition and history of “HMO,” a term coined in 1970. As was the case with the HMO, the ACO has been promoted primarily for its alleged value as a cost-cutting tool. Like the HMO concept, the ACO concept is vague and has multiple definitions which vary depending on who you ask. Like the HMO, the ACO is defined as an entity that will be “held accountable” for providing comprehensive health services to a defined population. As was the case with the HMO, “accountability” for cost will allegedly be achieved by shifting some or all of the insurance risk now born by insurance companies and public programs like Medicare to providers, and “accountability” for quality will allegedly be achieved by subjecting providers to report cards. As was the case with the HMO, Kaiser Permanente is often held up as an example of what an ACO might look like. The HMO and ACO are both said to have “enrollees.”
The history of the ACO also bears a striking resemblance to the history of the HMO. As the HMO was associated with the name of a single health policy entrepreneur (Dr. Paul Ellwood), so the invention of the ACO is associated primarily with one man – Dr. Elliot Fisher, director of the Center for Health Policy Research at Dartmouth Medical School. As the HMO industry was stimulated by federal legislation (notably 1972 federal legislation that allowed HMOs to participate in Medicare, and the 1973 HMO Act which subsidized the formation of HMOs for the non-elderly), so the first ACOs will be those authorized by federal legislation (PPACA) that requires Medicare to establish ACOs.
The principle difference between HMOs and ACOs, at least for the foreseeable future, will be their size. Whereas HMOs, like most insurance companies, generally have enrollees in the hundreds of thousands, the ACO has so far been defined as having a much smaller number of enrollees, possibly as few as 5,000 (that’s the minimum number of Medicare beneficiaries who must be in an ACO according to PPACA’s Section 3022). The other major difference between HMOs and ACOs, at least for the near term, will be the extent to which they bear insurance risk. Whereas HMOs function like insurance companies (they bear 100 percent of the risk that the premiums they charge will not be enough to cover all necessary services for their enrollees), ACOs will bear little or no insurance risk for the first few years. However, judging from published papers by Elliot Fisher and other proponents of ACOs, proponents want ACOs eventually to bear all insurance risk, just as HMOs have.
How do ACO proponents define an ACO?
Because the ACO concept is so vague, explaining it is difficult. In this section, I will illustrate the problem by quoting two authoritative sources on what an ACO is and listing several of the important questions these quotes don’t answer. The average reader will find this section frustrating. The only way to make sense of the discussion of ACOs is to go beyond reading the vague descriptions of ACOs offered by their proponents. To get a better sense of what ACOs might look like and how they might affect the US health care system, one must discuss the motives and goals of ACO proponents. I will do that in the section that follows this one.
In this section I quote from two authoritative sources: A recent paper by prominent ACO proponents, and Section 3022 in the recently enacted federal health care “reform” law which authorizes the Department of Health and Human Services, through CMS, to set up ACOs.
Here is a recent example of an ACO definition taken from a 2010 article co-authored by Elliot Fisher:
ACOs consist of providers who are jointly held accountable for achieving measured quality improvements [note that “measured quality improvements” is synonymous with report cards] and reductions in the rate of spending growth. Our definition emphasizes that these cost and quality improvements must achieve overall, per capita improvements in quality and cost, and that ACOs should have at least limited accountability for achieving these improvements while caring for a defined population of patients.
ACOs may involve a variety of provider configurations, ranging from integrated delivery systems and primary care medical groups to hospital-based systems and virtual networks [how can a network in the real world be “virtual”?] of physicians such as independent practice associations. All accountable care organizations should have a strong base of primary care. Hospitals should be encouraged to participate, because improving hospital care is likely to be essential to success. But in contrast to others’ definitions, we believe that this need not be an absolute requirement for all ACOs (Mark McClellan et al., “A national strategy to put accountable care into practice,” Health Affairs 2010;29:982-990).
This definition contains the ingredients common to virtually all ACO definitions, notably, language depicting a (poorly defined) group of providers being “held accountable” (by unidentified means by unidentified parties) for “measured improvements” (measured at an unknown cost to providers and the measurer) in the “cost and quality” of health care delivered to a (poorly defined) “population.” This is, of course, very similar to language used to describe HMOs over the last four decades. Note that in the quote above Fisher et al. are now saying ACO’s may or may not include hospitals, but ACOs do have to have “a strong base of primary care.” As we shall see in the next section, that was not always Fisher’s definition of an ACO. His original definition of an ACO required hospitals to be at the center of ACOs.
The second authoritative source I’d like to quote is Section 3022 of PPACA. (You can read the entire section by clicking on the following address and putting 3022 in the “find” box http://frwebgate.access.gpo.gov/cgi-bin/getdoc.cgi?dbname=111_cong_public_laws&docid=f:publ148.111.pdf.) The description of an ACO goes on for about three pages. Here are excerpts. I have emboldened words and phrases that cry out for definition:
Not later than January 1, 2012, the Secretary [of the Department of Health and Human Services, HHS] shall establish a shared savings program … that promotes accountability for a patient population and coordinates items and services under parts A and B, and encourages investment in infrastructure and redesigned care processes for high quality and efficient service delivery.
Here are other excerpts from Section 3022 describing the providers and enrollees in ACOs:
[T]he following groups of providers of services and suppliers which have established a mechanism for shared governance are eligible to participate as ACOs under the program under this section:
(A) ACO professionals in group practice arrangements.
(B) Networks of individual practices of ACO professionals.
(C) Partnerships or joint venture arrangements between hospitals and ACO professionals.
(D) Hospitals employing ACO professionals.
(E) Such other groups of providers of services and suppliers as the Secretary determines appropriate.
The ACO shall include primary care ACO professionals that are sufficient for the number of Medicare fee-for-service beneficiaries assigned to the ACO….
The Secretary shall determine an appropriate method to assign Medicare fee-for-service beneficiaries to an ACO based on their utilization of primary care services…
These excerpts tell us almost nothing about what an ACO is. They do tell us something about the aspirations of ACO proponents. They want to improve quality and lower cost, and they think ACOs will do this if “ACO professionals” band together under “shared governance” and accept “accountability for a patient population.” Who will the patients be? We don’t know. We know only that they will come from the fee-for-service Medicare program, and they will be “assigned” to the ACO by the HHS Secretary based on a formula that has something to do with which primary care “professional” the patient sees. Who will the providers be? We know only that they have to include primary care “ACO professionals,” a term which is defined to include but is not limited to physicians.
Perusing the remaining subsections of Section 3022 does little to improve our understanding of ACOs. Most of these provisions may be characterized as descriptions of reporting requirements or capacities ACOs have to have in order to qualify as ACOs. Here are two examples from Section 3022:
The ACO shall be willing to become accountable for the quality, cost, and overall care of the Medicare fee-for-service beneficiaries assigned to it.
The ACO shall demonstrate to the Secretary that it meets patient-centeredness criteria specified by the Secretary, such as the use of patient and caregiver assessments or the use of individualized care plans.
Of course, reporting requirements tell us little or nothing about the entity doing the reporting.
Other requirements in Section 3022 merely describe requirements for participating in the program that have nothing to do with how patients are taken care of, such as these:
The ACO shall enter into an agreement with the Secretary to participate in the program for not less than a 3-year period….
The ACO shall have a formal legal structure that would allow the organization to receive and distribute payments for shared savings … to participating providers of services and suppliers.
Judging from the way Elliot Fisher and other proponents talk about ACOs, Medicare ACOs under PPACA will be created by a two-step process: Providers will decide to form ACOs and will submit to CMS a list of primary care doctors who are in the ACO; CMS will then “assign” to the ACO those Medicare recipients who got a plurality of their primary care from those primary care doctors in recent years. But, I repeat, Section 3022 does not say how ACOs will be created. That will be up to CMS.
Let us now contemplate some of the questions these quotations from the Fisher et al. paper and from PPACA leave unanswered.
What does it mean to say patients will be “assigned” to an ACO? Does it mean the patient must only see providers in that ACO? If so, shouldn’t the patient have some say in whether he or she wants to accept “assignment” to the ACO? If the patient is not going to be notified that he or she has been assigned to ACO X and given the opportunity to object, and if the patient is not going to be required to see only the providers in that ACO, how will it be possible for ACO providers to be “held accountable” for patients who visit providers outside of ACO X? If the ACO doesn’t include a hospital, how are the ACO “professionals” in the ACO supposed to be held “accountable” for hospital care? PPACA is totally silent on these obvious questions. As we shall see in the next section, Fisher and other ACO proponents are aware of these questions, but they have no answers to them.
How will CMS “hold ACO professionals accountable” and what will that cost? ACO advocates support “pay for performance” programs – programs in which someone (in today’s system, that “someone” is usually an insurance company) issues report cards on doctors and hospitals and rewards or punishes providers based on the grades they get on the report cards. But existing research indicates P4P schemes have little effect on cost or quality, and that report cards can actually harm patients, especially sicker patients.
And now, to really confuse things, let me ask this question: How is an ACO supposed to serve non-Medicare enrollees? As difficult as it is to understand ACOs under Medicare, we can at least grasp that Medicare ACOs will deal with one payer. ACOs serving non-Medicare enrollees will have to deal with multiple insurance companies. How is that supposed to work? Section 3022 is, of course, silent on this issue (because it deals with Medicare only). Strangely, Elliot Fisher and other proponents of ACOs have offered no solutions to this problem either.
But let’s imagine a “solution” to this problem. Let’s imagine that insurance companies band together into the equivalent of regional monopolies for the purpose of developing common rules and contracts with ACOs, and that through these contracts the insurance industry eventually shifts all or some insurance risk to ACOs. In other words, ACOs now insure their “enrollees” just the way insurance companies insure their policy-holders. If that happens, we will have the equivalent of two layers of insurance companies – the existing layer, and a layer of ACOs under the current layer. Would it be fair then to ask, at long last, who needs insurance companies? If administrative costs of the current system are high, won’t the administrative costs of a two-layer multiple-payer system be even higher?
History of the ACO concept
The preceding section should make one thing clear: Defining an ACO in any detail is impossible.
And yet the ACO experiment is under way.
When human beings witness grandiose schemes going awry (wars that were supposed to be easy to win, ships that weren’t supposed sink, collateralized debt obligations that weren’t supposed to bankrupt the banking system of the Western world), people often ask, “What were they thinking?” Given the grandiose claims made for ACOs, and the numerous fundamental questions about them that ACO advocates can’t answer, it is appropriate to pose that question about the ACO proposal now, even though it is too early to say the proposal has already failed.
The ACO concept – the arguments made for it and the vague descriptions of it – is not new. It has intellectual roots in the managed care movement that extend back decades. As I have already indicated, the description of the ACO and the justification for experimentation with it are very similar to the HMO concept and the justifications used for unleashing HMOs upon the populace forty years ago. But the “ACO” label is only four years old. In an endnote to a 2009 paper about ACOs in Health Affairs, Elliot Fisher and colleagues stated that Fisher coined the label in the course of a “discussion” with Glenn Hackbarth, then and now the chairman of Medpac. (Elliot Fisher et al., “Fostering accountable health care: Moving forward in Medicare,” Health Affairs, 2009;28:w219-231[Published online 27 January 2009]).
Robert Berenson, who is the vice chair of Medpac, confirmed Fisher’s statement in a useful paper about the history of the ACO concept. Berenson and his co-author stated, “Together, the Medicare Payment Advisory Commission …, the organization that advises Congress on payment and related policies for Medicare, and Fisher provided the impetus for the current concept and interest in ACOs…” (Kelly Devers and Robert Berenson, “Can Accountable Care Organizations improve the value of health care by solving the cost and quality quandaries?” October 2009, Urban Institute, 2)
A search of the transcripts of Medpac’s public meetings reveals that this discussion occurred at Medpac’s November 9, 2006 meeting. The agenda for that meeting contains an item entitled, “Elliot Fisher on hospital staffs.” Fisher was on Medpac’s agenda because Medpac had been instructed by the Deficit Reduction Act of 2005 to propose to Congress methods of paying doctors under Medicare Part B that would replace the Sustainable Growth Rate (SGR) formula. As a Medpac staffer put it at Medpac’s October 5, 2006 meeting, “The report must discuss disaggregating the current national (SGR) target into multiple pools using five alternatives: Group practice, hospital medical staff, type of service, geographic area, and physician outliers.” http://www.medpac.gov/transcripts/10_06_MEDPAC_all.pdf
Congress is interested in a substitute for the SGR that breaks doctors up into “pools” that are smaller than the entire population of US doctors who treat fee-for-service Medicare patients (which is the “pool” to which the SGR applies). The theory is that if smaller groups of doctors are subject to spending caps, the caps will work better than the SGR does. In short, Fisher had been invited to the November 9, 2006 Medpac meeting to comment on the possibility of using the hospital medical staff “pool” as a substitute for the SGR.
When his turn came to testify, Fisher proposed something he called “Extended hospital medical staffs,” or EMHSs. The EHMS, according to Fisher, was concocted using his Medicare database at Dartmouth Medical School. Here is an excerpt from Fisher’s testimony in which he describes the EHMS. Notice how much it resembles the Six Degrees of Kevin Bacon game, a game in which party-goers try to “connect” Kevin Bacon according to which movies Bacon and the other actors have been in. http://en.wikipedia.org/wiki/Six_Degrees_of_Kevin_Bacon
[T]hinking about inpatient hospital stays and how to foster accountability among physicians for the services provided during hospital stays, we’ve tried to build a model that extends that to all beneficiaries and all physicians within the American health care system. (p. 287)
Let me briefly describe the general approach we’ve taken to assigning patients. If a physician works in an inpatient setting, we assign them to the hospital where they provided care to the greatest number of Medicare beneficiaries say saw. If they get no inpatient work, we assigned him to the hospital where the plurality … of their patients they billed for were admitted. So if you touch a patient, you identify all the Medicare beneficiaries they touch, and you see which hospitals they go to. It turns out, not surprisingly, that you can assign virtually all physicians billing Medicare to a hospital. (p 288) http://www.medpac.gov/transcripts/1108-09Medpac%20final.pdf
Fisher said that applying his rules for creating EMHSs to Medicare claims data for the three-year period 2002-2004 would create 5,000 EMHS’s across the country.
Fisher’s statement that he can invent rules for assigning patients to doctors and doctors to hospitals is no more or less logical or useful than the statement by the inventors of the Kevin Bacon game that they can assign a Kevin Bacon number to virtually any actor. Thus, Elvis Presley (who never met Kevin Bacon) has a Kevin Bacon number of 2 because he appeared in the movie “Change of Habit” (1969) with Ed Asner, and Ed Asner appeared in the movie “JFK” with Kevin Bacon. As one of the inventors of the Kevin Bacon game observed about the origins of the game, “It became one of our stupid party tricks, I guess. People would throw names at us, and we’d connect them to Kevin Bacon.” http://en.wikipedia.org/wiki/Six_Degrees_of_Kevin_Bacon
Does Elliot Fisher’s formula for connecting patients to doctors and doctors to hospitals to form an EMHS have any more integrity or usefulness than the rules of the Kevin Bacon game? At the November 2006 Medpac meeting, Fisher offered contradictory comments on this most fundamental issue. In his testimony he stated, “It turns out 62 percent of physicians [who treat Medicare patients] perform inpatient work.…. That looks pretty good.” Compared to what? Is a Bacon number of 3 “pretty good”? Fisher then said, “These [EMHSs] are coherent groups [which] really are practicing within a local environment.” “Coherent” compared to what? Is the list of actors which gives an actor a Kevin Bacon number of 7 more or less “coherent” than a list which gives an actor a number of 4?
But during his testimony Fisher made statements indicating even he wonders whether his computer-generated constructs are any more useful than a Kevin Bacon number. As I have already noted, he observed that “it is very hard … to think about how a primary care physician is going to influence the practice of a cardiologist across town” (p. 287). During the question-and-answer phase that followed Fisher’s testimony, Medpac commissioner Robert Reischauer observed that Fisher’s proposal might work if doctors, hospitals and other health care providers were already working together in large, cooperative groups, but the problem is they are not. In fact, Reischauer observed, there is much conflict between providers. Reischauer said Fisher seems to think the health care industry resembles “a whole bunch of well behaved sheep eating on a prepared lawn” when the proper metaphor would be “the Serengeti … with some lions and some hyenas and elephants … not only eating the grass but also each other.” Reischauer then observed, “[If] you draw a line around this group” that was artificially created by Fisher’s algorithm, “That doesn’t create an organization” (p. 335). Fisher’s response was useless. He replied that his EMHS proposal “encourages people to start having a conversation,” and he alleged it would “move the system in a different direction.” Then, no doubt realizing that his answers were not helpful, he concluded, “But you know,… I am a naïve epidemiologist.” (p. 338)
Fisher had an equally inadequate answer to a question about how ACOs could be held accountable for anything when they had no power to force their patients to see only ACO providers. When Fisher was specifically asked by a commissioner how an ACO in the north could be forced to pay for the health care of “snowbirds” (Medicare recipients who travel in the south during the winter) during their travels in the south, Fisher offered this unhelpful response:
Snowbirds are going to be a challenge. We have the same problem in Vermont, where they all go south for a month…. But I do think … if patients assigned to – chose the physician or the group with which they were going to be affiliated, and then that group had some responsibility to communicate…. [I]f it were my patients in Vermont … I would have some responsibility to communicate with you. When I was practicing at the VA and my patients went south, it was always very hard to get the conversation going with the physicians in the other communities. But that’s, I think, what we need to do. And this would give someone the incentive to pay attention to that communication. (pp. 363-364)
It is difficult to imagine a more cryptic answer. Fisher simply didn’t answer the question whether it would be fair or effective to require a Vermont ACO to be forced to pay for the cost of medical care for one of its “enrollees” in Florida.
Notice as well how Fisher implied that Medicare beneficiaries would be given a choice to join the ACO to which Fisher’s algorithm assigns them. However, he does not come right out and say that. But if Medicare will be obligated to notify recipients of the ACO to which they have been assigned, and to give them a choice of accepting their assignment or staying out of the ACO (or joining another one), and substantial numbers of recipients refuse to join their assigned ACO, how will the ACO concept work? Fisher had no comment on this issue then, and has apparently offered no solution to this problem in the intervening four years.
It was during the question-and-answer portion of this November 9, 2006 Medpac meeting that the phrase “accountable care organization” was invented. Medpac Chair Glenn Hackbarth first used the term rather than EMHS (p. 309). A few minutes later Fisher said to Hackbarth:
I love your notion of accountable organizations. It’s exactly the right thing we want to create. And I agree completely with applying it to all services. It should include the whole gamut of care so we get rid of the silos, because you look at the numbers of care transitions and you just see that these places are churning patients from hospital to acute care to nursing home back to the hospital. (p. 311)
It took awhile for the ACO title to catch on with other commission members at this meeting. At page 356 of the transcript we learn that commissioner Nancy Kane MD referred to “this accountable – whatever we’re calling it.” http://www.medpac.gov/transcripts/1108-09Medpac%20final.pdf
Despite the important questions Fisher’s testimony had left unanswered, the momentum for ACOs grew rapidly after the November 2006 Medpac meeting. Fisher and several of his colleagues published an article in Health Affairs a month later laying out the tables he presented at Medpac’s November meeting and presenting essentially the same arguments for trying to create ACOs that he had presented to Medpac (“Creating Accountable Care Organizations: The Extended Hospital Medical Staff: A new approach to organizing care and ensuring accountability,” Health Affairs 2007: w44–w57 [published online 5 December 2006; 10.1377/hlthaff.26.1.w44]
Fisher and several associates published numerous articles thereafter, most of them in Health Affairs, all making the same basic arguments and none answering the unanswered questions raised at the November 2006 Medpac meeting. As I reported in the first section, ACOs had become a staple of Democrats’ health care “reform” proposals by late 2008 or early 2009.
After Fisher’s presentation, Medpac instructed its staff to develop an ACO proposal for commissioners to discuss. Medpac devoted time to the ACO concept at several of its subsequent meetings http://www.medpac.gov/transcripts/1108-09Medpac%20final.pdf
and published chapters on ACOs in several reports to Congress.
It is not clear to me how the ACO concept shifted from the hospital-centric version proposed by Fisher in November 2006 to the primary-care-centric version – a version that does not require that an ACO include a hospital – which appears in Section 3022 of PPACA. As late as June 2009, Medpac was still asserting that an ACO had to include a hospital (see Endnote 1 in Chapter 2 of Medpac’s report of that date http://www.medpac.gov/chapters/Jun09_Ch02.pdf). But hospitals are not required by PPACA’s definition, and as we saw in the first section Fisher and his colleagues are no longer saying hospitals will play the role of Kevin Bacon in the construction of ACOs. Like Fisher, Medpac is now saying ACOs don’t have to include hospitals. At the last meeting of the Medpac commissioners, which occurred on September 13, 2010, a Medpac staffer described an ACO as an entity that must have “a core group of primary care physicians that serve at least 5,000 fee-for-service Medicare beneficiaries.” He stated explicitly that specialists and hospitals are optional. http://www.medpac.gov/transcripts/913-914MedPACfinal.pdf
As confusing as the Kevin-Bacon version of ACOs was (doctors were linked to hospitals by arbitrary rules, and patients were linked to doctors by arbitrary rules), it at least had the virtue of helping us grasp who would constitute the core of an ACO’s providers and patients. Now even that much definition has been erased. In an October 5, 2010 article, Kaiser Health News summarized the ACO as “just a concept for now.”
Until the policy entrepreneurs and politicians who brought the ACO concept to center stage agree on some Kevin-Bacon-style rules that set limits on how separate (in terms of distance, shared patients, and shared rules of governance) providers can be and still be part of an ACO, any “definition” of ACO must begin with, “It’s just a concept.”
The ACO is the managed-care movement’s new HMO. Despite the failure of the HMO and the managed-care tools pioneered by the HMO to work as advertised, believers in managed care remain convinced that something like the HMO is still the solution to the US health care crisis. They are driven to this conclusion by their inaccurate diagnosis of the crisis – that the fee-for-service system is the cause of high US health care costs. From this diagnosis they jump to the conclusion that the FFS method of paying providers must be replaced by some method of payment that shifts insurance risk to doctors. (Don’t bother asking why shifting all doctors to a salary is not on the table as an alternative to the FFS method. Only shifting insurance risk to doctors is on the table.)
But shifting risk to doctors and hospitals is feasible only if the provider group accepting insurance risk is large and the patients these providers treat can be forced to see only those providers. Individual providers, and even small groups of providers, cannot absorb substantial amounts of insurance risk; only large groups of providers can do that. The ACO is seen by its proponents as the vehicle for clumping providers together into entities large enough to bear insurance risk. This rationale for ACOs is rarely mentioned; ACO advocates prefer to argue that providers must be pushed into large chains of providers to improve their ability to “coordinate care” and otherwise take better care of patients than they do now.
If enough political support for the ACO is brought to bear and providers begin to think that ACOs in some form are inevitable, we will see another wave of mergers and consolidation among providers like the one that occurred in the 1990s (that wave was provoked by consolidation within the insurance industry and the widespread adoption of managed care tactics by insurers). Even though PPACA’s version of the ACO does not require hospital participation, hospitals and hospital systems are already making moves to acquire clinics and other hospitals in order to be in a position to create ACOs. For example, in February 2010 two competing hospitals in Omaha – the Nebraska Medical Center and the Methodist Health System – announced they were forming the Accountable Care Alliance. http://www.kaiserhealthnews.org/Stories/2010/October/05/accountable-care-organizations.aspx
The insurance industry is not happy about these early signs of another wave of consolidation among providers. It is already complaining about it to HHS and to anti-trust authorities. http://www.kaiserhealthnews.org/Stories/2010/October/05/accountable-care-organizations.aspx
Because the ACO so closely resembles the HMO, there is no reason to expect the ACO will perform any better than the HMO did. Like the HMO movement, the ACO movement will probably damage quality on balance and have no impact on costs, and may even raise costs. Costs may rise because administrative costs will go up (both for insurers and providers), and because providers will be larger.
Advocates of ACOs are among the most passionate advocates of evidence-based medicine. They should practice what they preach. They should encourage pilot projects which test clearly defined examples of the ACO concept (or better yet, some of the tools ACOs will allegedly use to improve care and cut costs), and when research on these pilots has demonstrated that ACOs are safe and effective, then and only then should ACO advocates promote widespread adoption of ACOs.