CMS recently proposed regulations for Accountable Care Organizations which are extensive and complex. This article reviews the proposed regulations and discusses the prospects for success.
The 2010 Patient Protection and Affordable Care Act included a provision to promote the formation of Accountable Care Organizations (ACOs). These organizations will be eligible to share in the savings to Medicare if they are able to reduce costs and provide high-quality care. The law allows a wide variety of organizations to become ACOs, even networks of providers that are small compared with major integrated delivery networks. The Center for Medicare and Medicaid Services recently proposed regulations, which are extensive and complex. They impose significant regulatory requirements on these new organizations, ranging from the structure of the organization to quality standards for qualifying for any shared savings. There are a number of challenges to ACOs and it is uncertain whether they can achieve the goals Congress had in mind or even whether many healthcare provider organizations will be interested in participating in the program. The potential for shared savings may be too small to justify the additional costs and regulatory burdens of becoming an ACO. In addition, the incentives to physicians may be inadequate to encourage behavior that reduces cost while maintaining quality. The article reviews the proposed regulations and discusses the prospects for success of ACOs.
(Am J Manag Care. 2011;17(8):560-568)
Last year’s major healthcare reform bill, the Patient Protection and Affordable Care Act (ACA),1 provided for a new form of healthcare organization called Accountable Care Organizations, or ACOs. The concept of ACOs has been advocated by a number of healthcare experts for some time.2-6 In addition, the Center for Medicare and Medicaid Services (CMS) conducted a demonstration project during 2005-2009 to test the concept. The results of this demonstration, which are discussed below, were not completely encouraging. However, they did not dissuade Congress from authorizing CMS to implement a national program. ACOs now have new prominence because the ACA includes a section intended to promote their formation.
Recently, CMS proposed regulations describing how hospitals and other providers can create an ACO and setting out the requirements to qualify for the program.1,7 The CMS has also published rules for ACOs to apply for waivers from certain federal laws in order to operate effectively.8 The reaction to the Proposed Rule was overwhelmingly negative. Major healthcare organizations, such as the American Hospital Association (AHA), the American Medical Association (AMA), and others criticized the proposal as too burdensome and costly considering the limited incentives offered by the program. I discuss some of these comments in this article in connection with particular provisions of the proposal.
The ACA and the proposed regulations promote the formation of ACOs by providing for a creative form of reimbursement that is intended to encourage providers within the ACO to reduce costs while maintaining quality. In particular, an ACO may receive payments for shared savings if it can reduce costs for Medicare beneficiaries below a specified benchmark. These payments in turn can be shared with physicians and other providers in order to encourage more efficient healthcare delivery. But while ACOs have the potential to reduce costs, in large part by influencing physician behavior, as I discuss below, a number of challenges may prevent them from providing significant benefits.
The ACO potentially has 2 advantages over other forms of organization. First, the ACA provides for considerable flexibility in how ACOs are organized. They need not be as large or integrated as an Integrated Delivery Network (IDN). Consequently, they can be developed more quickly and with fewer resources. The ACA provides that the following arrangements may qualify as an ACO1:
1. ACO professionals in group practice arrangements. (“ACO professionals” are defined as physicians, physician assistants, nurse practitioners, and clinical nurse specialists.)
2. Networks of individual practices of ACO professionals.
3. Partnerships or joint venture arrangements between hospitals and ACO professionals.
4. Hospitals employing ACO professionals.
5. Such other groups of providers of services and suppliers as the Secretary of Health and Human Services (HHS) determines appropriate.1,7
The second potential advantage is that the ACA provides for a unique form of payment to an ACO: participation in “shared savings,” which are defined as a percentage of the difference between a benchmark determined by the Secretary and the costs to Medicare for services provided to beneficiaries who are assigned to the ACO.1 Although this type of reimbursement has features that resemble capitation, it is different in important respects. Capitation reimbursement pays the same amount for each enrollee in a plan, whether an individual uses services or not. An ACO is generally guaranteed at least the standard Medicare reimbursement levels for beneficiaries who require services. Under some circumstances, an ACO can experience a loss in reimbursement if its expenditures for beneficiaries are too high. Thus, total Medicare payments are higher for a population that experiences a greater need for services, while total payments are fixed under a capitation approach. As a result, ACOs face considerably less financial risk than a Health Care Organization (HCO) that depends on fixed capitation payments. (As used here, “HCO” is a generic term for an integrated network of healthcare providers, ranging from a single hospital affiliated with a physician group to a large IDN, which may include dozens of hospitals, affiliated physician groups, and other types of providers.)
General Requirements to Become an ACO
CMS’s Proposed Rule sets out the requirements for participating in the program, the formulas for shared savings, and other relevant regulations.7 The proposed regulations are complex and lengthy. In part, this is the result of the analytical challenges inherent in devising a rational way to calculate shared savings. In addition, it appears that CMS envisioned each ACO to be a model HCO, which incorporates all the ideal qualities of efficiency, high quality care, and sensitivity to patients. The complete regulatory scheme is quite daunting, so much so that it is likely to discourage many organizations from participating in the program.
In order to qualify for participation in the Shared Savings Program, an ACO must agree to a series of conditions.1 It must:
1. become accountable for the quality, cost, and overall care of the Medicare fee-for-service beneficiaries assigned to it;
2. enter into an agreement to participate for at least 3 years;
3. have a formal legal structure that allows it to receive and distribute shared savings;
4. include primary care professionals sufficient for the beneficiaries assigned to it (at least 5000);
5. provide HHS with required information;
6. have a management structure that includes clinical and administrative systems;
7. adopt processes to promote evidence-based medicine and coordinate care; and
8. meet patient-centeredness criteria specified by HHS.
Specific Requirements for ACOs
The statutory requirements are amplified in the proposed regulations. The following discussion summarizes the principal requirements:
Three-Year Agreement. A participating ACO must agree to comply with certain requirements for 3 years.7 As discussed below, an ACO can choose to participate under Track 1 (the “one-sided” lower-risk, lower-reward model) or Track 2 (the “two-sided” higher-risk, higher-reward model). After the first 3 years, all ACOs must operate under the Track 2 model.
Creating a Separate Legal Entity to Serve Certain Functions. An ACO can be composed of several distinct legal entities, such as hospitals and physician practices. However, it must create a separate legal entity, such as a non-profit corporation, to serve certain functions, such as distributing shared savings.7
Governance Structure. While the statute and the Proposed Rule are flexible regarding the type of legal entity required, an ACO must have a governance structure that meets CMS’s requirements. The ACO’s participants must have at least 75% of “control” of the ACO.7 As a technical matter, it would seem that this 75% control requirement could be met by giving 1 physician practice, out of many participating in an ACO, the right to name, say, 8 of 10 board directors. However, the explanation of the Proposed Rule also states: “[W]e propose that the ACO must demonstrate a mechanism of shared governance that provides all ACO participants with an appropriate proportionate control over the ACO’s decision making process.”7 This statement suggests an additional requirement, that each participant must have a share in the control of the ACO that reflects its degree of participation in the ACO. In addition, beneficiaries are to play some role in governance as well.
The idea of proportionate control is not a traditional one in HCOs or any other corporate structure, nor is “customer” participation in governance. Control of a traditional corporation is vested in a board of directors, which is chosen in a variety of ways, including by shareholders in a for-profit corporation. There is no requirement that service providers affiliated with a corporation, eg, physician groups affiliated with an HCO, share proportionally in its governance. Mandating a particular scheme of governance could undercut the efficiency of the ACO and require a major revision of existing governance structures of HCOs that want to qualify as ACOs. There is certainly a value in ensuring the accountability of the leadership of an ACO, but “proportionality” is an unwieldy and unnecessary requirement. Note, for example, the suggestion made by Delos M. Cosgrove, MD, on behalf of the Cleveland Clinic, in his letter (May 26, 2011) (Cleveland Clinic Comment) to Donald Berwick, MD: “We recommend that CMS drop its specific Board composition and relationship requirement.” Note also the suggestion made by Rick Pollack, on behalf of the AHA, in his letter (June 1, 2011) (AHA Comment) to Donald M. Berwick, MD: “We urge CMS to moderate its proposed governance requirements to allow providers to use their current governance process…as long as they can demonstrate how they achieve shared governance on care delivery policies.”
Leadership and Management Structure. The ACO must satisfy CMS that it has a particular management structure that accomplishes the goals that CMS envisions. For example, there must be a full-time senior-level medical director physically present at one of the ACO participant facilities who is a board-certified physician.7 An ACO’s participants must have a meaningful investment in the ACO but the investment need not be financial. It can be a “meaningful human investment (for example, time and effort) in the ongoing operations of the ACO....”7 There must also be an “infrastructure” (eg, information technology) that enables the ACO to collect and evaluate data.8 One of the most demanding—and probably unrealistic—requirements is that 50% of primary care physicians participating in the ACO meet the CMS standard for “meaningful use” of information technology.7
Giving Certain Notices to Beneficiaries. These notices include informing beneficiaries that their ACO providers are participating in the ACO, that they have a right to use other providers, and that they can opt out of CMS sharing data on the services they receive.7
Taking Responsibility for the Care of at Least 5000 Beneficiaries. All ACOs must have responsibility for at least 5000 beneficiaries, although CMS contemplates that some ACOs will be much larger.8 A beneficiary does not “choose” an ACO, at least not directly, and an ACO does not “choose” a beneficiary. Instead, the Proposed Rule states: “The term ‘assignment’.... refers only to an operational process by which Medicare will determine whether a beneficiary has chosen to receive a sufficient level of the requisite primary care services from physicians associated with a specific ACO so that the ACO may be appropriately designated as exercising basic responsibility for that beneficiary’s care.”7 In general, CMS determines whether a patient is assigned to an ACO based on whether a patient is receiving a “plurality” (not a majority) of his or her primary care services from participants in the ACO.7 This determination is made “retrospectively,” that is, it is based on use of services during the 3-year period before the ACO agreement begins by beneficiaries that “would have” been beneficiaries of the ACO if they had kept using the same primary care providers during the agreement period that they used during the benchmark period.7 “Performance period” and “performance years” are CMS’s terms for the period of the ACO’s agreement when determinations of shared savings are made.7
Complying With Extensive Reporting Requirements. These include reporting information about participating providers, submitting extensive information in the original application and in the event of significant changes, and reporting information related to quality of care and making public disclosures.8
Adopting Procedures to Achieve “Evidence-Based Medicine,” “Patient Engagement,” “Coordination of Care,” and “Patient-Centeredness.” These terms are not self-defining and there are obviously a wide variety of approaches to achieving these goals, depending on how they are defined. The explanation of the Proposed Rule states: “Evidence-based medicine can be generally defined as the application of the best available evidence gained from the scientific method to clinical decision-making.”8 The Proposed Rule also includes several criteria for determining whether the “patient-centeredness” requirement is met. These include, for example, making beneficiaries and their families partners in caregiving and ensuring patient involvement in governance. The ACO must also have a “beneficiary experience of care survey” in place and a plan for using the results of the survey.
Submitting All Marketing Materials to CMS for Advance Approval. States CMS: “We believe there is a potential for beneficiaries to be misled about Medicare services available from an ACO or about the providers and suppliers from whom they can receive those services.”8 For this reason, all marketing materials must receive advance approval from CMS. Note that this is a more burdensome requirement than the one imposed by the Federal Trade Commission (FTC) on advertisers generally. The FTC does not require advance approval of advertisements and marketing materials, but challenges advertising that it considers to be false or deceptive. This advance approval requirement puts a significant administrative burden on CMS and also risks substantial delays in marketing campaigns while CMS performs its review. In his letter to Donald Berwick (June 3, 2011) (AMA Comment), Michel D. Maves pointed out: “[The] requirement that CMS approve all ACO marketing materials is an unprecedented, unnecessary, inappropriate, and unworkable requirement.” Additionally, it does not seem likely that ACOs have a greater potential to engage in deceptive advertising than advertisers in general. Therefore, it is difficult for CMS to justify this unusual advance review requirement.
Complying With Other Legal Requirements. The Proposed Rule requires that the ACO have a compliance plan to ensure compliance with all legal requirements, including the Anti-Kickback Statute, the False Claims Act, etc.8 One of the most significant requirements is intended to prevent conflicts of interest by mandating that ACOs require members of their governing body (not all ACO participants) to disclose relevant financial interests.7 A waiver procedure proposed by CMS would allow ACOs to be exempt under some circumstances from the Physician Self-Referral Law, the Anti- Kickback Statute, and other restrictions.7
Complying With Antitrust Standards. Obviously, CMS and the antitrust agencies that participated in developing the Proposed Rule are concerned that the cooperation of multiple providers within an ACO might harm competition. An ACO would already be subject to federal and state antitrust laws without any new requirement, but the Proposed Rule and a Policy Statement issued by the FTC and Department of Justice provide for a special antitrust review procedure.9 The antitrust agencies already have in place a Health Care Policy Statement, which addresses the degree of integration required to be analyzed under a “rule of reason,” which takes into account market share and other facts that affect competition, rather than a “per se” analysis, which focuses only on the nature of the agreement. Arguably, the antitrust procedure in the Proposed Rule includes a more flexible standard for rule of reason analysis.
In general, the required review is based on the ACO’s market share for a “common service” if 2 or more participants in the ACO offer the same service. If the total market share of all the ACO participants for that service is below 30% in the ACO’s geographic area called the Primary Service Area (PSA), the ACO is eligible for a “safe harbor” for that service and no review is required.9 For example, assume that: 1) 2 group practices that are participants in an ACO offer cardiology services, and 2) the combined market share of these 2 group practices for cardiology services is less than 30% in the relevant PSAs. Under those circumstances, there is a safe harbor for cardiology services and no special review is required. However, if the market share is above 50%, the ACO must seek an antitrust review and obtain approval from the agencies.8 If the share is between 30% and 50%, the ACO has the option to seek a review.9 In that intermediate zone, the Policy Statement provides that, if the ACO avoids certain types of conduct, it is more likely to obtain clearance from the agencies.9 For example, conduct that should be avoided includes discouraging payers from directing patients to use providers outside the ACO and contracting with physician specialty groups on an exclusive basis.9
Note that the standard of review—safe harbor, optional review, or required review—is based on a particular service, not the overall percentage of patients who are assigned to the ACO in a particular geographic area. Even if the market share under the above analysis is less than 30% for all services but 1, review is required for the remaining service.9 Note also that, except for the “dominant provider” limitation discussed below, no review is required if there is only 1 participant in the ACO that offers a particular service.9 To take the cardiology services example, assume that there is only 1 group practice within the ACO that offers cardiology services, but its market share in the PSA is 75%. There would be no mandatory review. However, to come within a safe harbor, “The ACO participant (a ‘dominant provider’) must be non-exclusive to the ACO to fall within the safety zone. In addition, to fall within the safety zone, an ACO with a dominant provider cannot require a commercial payer to contract exclusively with the ACO or otherwise restrict a commercial payer’s ability to contract or deal with other ACOs or provider networks.”9
The antitrust agencies are obviously worried that the consolidation of hospitals and physician practices within an ACO can threaten competition, and therefore a special review process is needed. However, like the requirement for advance approval of marketing materials, this new review procedure may end up delaying the operation and formation of ACOs and imposing unnecessary burdens and costs on ACOs as well as the agencies. Some commenters argued that, if there is to be a mandatory review threshold, the one proposed by CMS and the antitrust agencies is too low. In contrast, the health insurance industry tended to take the position that the antitrust standards should be stricter. That is presumably because health insurers are worried that consolidation may allow an ACO to have enough market power to drive down health insurance premiums.
As a practical matter, many HCOs with even a modest degree of integration will exceed the 50% threshold in at least 1 area, triggering an expensive mandatory review. For example, the AHA found that: “[I]n the overwhelming majority of these areas, hospitals that sought to partner with any other hospital likely would exceed the 50 percent review threshold…and, thus be faced with costs of potentially several hundred thousand dollars in order to defend their ACO application before one of the Agencies.” In light of these difficulties, CMS and the antitrust agencies should consider scrapping the special review process altogether in favor of issuing more specific enforcement guidance and offering to give expedited advisory opinions.
Meeting Quality Standards. Section 1899(b)(3)(A) of the Act requires the Secretary to determine appropriate measures to assess the quality of care furnished by the ACO. As a first step, the Proposed Rule identifies performance criteria and reporting requirements but does not set specific quality standards.8 The Proposed Rule describes 5 “domains” of quality measures. Those standards will be developed in future rule making. The Proposed Rule includes 65 measures to calculate the “ACO Quality Performance Standard.”8 In the first year, the ACO would meet the quality performance standard by full and accurate reporting. In subsequent years, the ACO would have to meet certain minimum quality standards, which will be announced by CMS. If it does not meet the minimum quality performance standard for at least 1 domain, it would not be eligible for shared savings.7 In addition, the level of shared savings allocated to an ACO will be reduced if it falls below a perfect score on the quality standards. This is termed the “performance score approach” by CMS, to be distinguished from the “threshold approach,” which would allow maximum shared savings if all minimum quality standards are met.7 The effect may be that it is impossible for most ACOs to achieve the maximum possible savings. As noted in the AHA Comment, “[T]he manner in which CMS proposes to use the ACO’s quality score to reduce the sharing rates makes these maximums unattainable. This is because, unless an ACO scores 100 percent on all 65 proposed quality measures, which may not be possible, the quality score will always serve to reduce the sharing rate from the stated 50 and 60 percent.”
The most significant element of the ACO program is the potential for ACOs to participate in “shared savings,” that is, some portion of the savings to Medicare from their efficient delivery of care. That possibility is the incentive for organizations to participate in the program in the first place. The question is whether the incentive is strong enough, given the administrative burden, implementation costs, and financial risks that the ACO must assume.
Although the idea of “shared savings” sounds simple enough, implementing the idea in a rational way is quite complicated. First, CMS must establish some baseline for comparison of the Medicare costs of the ACO’s beneficiaries to determine whether the ACO actually reduces costs. Second, CMS must decide how much of the savings the ACO can keep and develop formulas for rewarding desirable activities and penalizing undesirable ones. Third, CMS must develop a way to monitor the success of the ACO in lowering costs and meeting quality standards. All of these goals present considerable analytical and practical challenges.
Two ways of identifying “benchmark” expenditures, ie, a rational baseline for measuring whether the ACO reduced expenditures, were considered by CMS. In theory, CMS could compare the costs of caring for 5000 ACO beneficiaries in, for example, the Des Moines, Iowa, area with 5000 different beneficiaries with roughly the same demographic and health characteristics in a different city. Even though that approach was used in a CMS demonstration project,10 it raises a host of conceptual and measurement problems. Consequently, the only options considered by CMS were to compare the costs of the ACO beneficiaries who “would have” been assigned to the ACO had they continued to receive care from the same providers during the performance period that they used during the benchmark period (CMS’s Option 1) with the costs to Medicare during the benchmarking period for persons who “actually” were assigned to the ACO during the performance period (Option 2). There is a large overlap between these groups but they are not the same since, for example, some beneficiaries who would have been assigned to the ACO changed providers or died. CMS believes both approaches are “legally permissible” but it proposes Option 1.7 Several commenters criticized this approach. For example, the AMA recommended that CMS abandon a retrospective attribution approach and allow patients to volunteer to be part of an ACO. The ACO would then know up front the beneficiaries for which they are responsible.
The identification of beneficiaries and calculation of their costs during the benchmarking period does not end the analysis because CMS must then adjust the costs during the benchmarking period for inflation and other factors, including beneficiary characteristics.7 In order to adjust for health risks, CMS proposes to use its hierarchical condition categories (HCC) prospective risk adjustment model that was developed for the Medicare Advantage program.7 Additional adjustments are made for other factors, such as geographic adjustments, past bonus payments, and penalties for providers.7
The Minimum Savings Rate (MSR)
The statute requires that there be some minimum amount of savings before any sharing takes place. CMS calls this minimum savings the “minimum savings rate” (MSR).7 CMS says one factor in coming up with a rational MSR is that the smaller the number of beneficiaries in an ACO, the more likely that cost reductions will be the result of random variations. Thus, CMS says that the more beneficiaries served by an ACO, the lower the MSR can be.7 An additional consideration is whether an ACO chooses the “one-sided” or “two-sided” model, which are discussed below. Since ACOs following the two-sided model are high risk/high reward, CMS permits a lower MSR for these ACOs. The applicable MSRs follow a sliding scale depending on the size of the ACO, ranging from 2% to 3.9% for the onesided model to a flat 2% for those adopting the two-sided model.7
The Net 2% Rule
In the Proposed Rule, CMS expressed concern about sharing savings from the “first dollar” once an ACO reaches the MSR. For example, assume an ACO exceeds its required MSR of 3%. It could in theory begin to share in all savings. However, CMS was concerned that, in the case of small ACOs, the savings could reflect only random variation. Its concern is greater in the case of the “one-sided” model since there is no accountability for losses. Thus, according to CMS, the ACO has “less motivation to eliminate unnecessary expenses and may be more likely to be rewarded as a result of methodological requirements.”7 Thus, CMS proposes that for most ACOs using the one-sided model that reach the required MSR, shared savings would be based only on savings over 2% of their benchmark.7 For example, if the benchmark is $12,000 and the required MSR is 3%, no shared savings are permitted unless the savings exceed $360. However, the shared savings are based only on $120, the savings in excess of 2% of the benchmark. An exception is proposed by CMS for smaller, physician-driven ACOs and ACOs caring for underserved populations, which are exempt from this “net 2% provision.”7 The high-risk, high-reward ACOs operating under the two-sided model are exempt altogether.8
Track 1 (the “One-Sided” Model) and Track 2 (the “Two-Sided” Model)
The Proposed Rule creates 2 models for ACO participation. Track 2, the “two-sided” model, has greater financial risks than Track 1, the “one-sided” model,” but it has the potential for greater rewards.8 An ACO chooses 1 track for 3 years. The shows the differences between the 2 tracks.7
Percent of Shared Savings
The statute does not specify the percentage of savings that an ACO can retain. Obviously Congress intended that CMS should get to keep a significant share of the savings, since a major purpose of the ACO program is to reduce Medicare costs. The ACO gets to keep the rest in order to provide an incentive to reduce costs, but how are these savings to be divided? Subject to limitations discussed below, CMS proposes that the shared savings would be 50% for ACOs under the Track 1 model that meet quality standards and 60% for ACOs under the Track 2 model that meet quality standards.7 These percentages can be adjusted upward for treating certain groups of patients, for example, beneficiaries who receive service from rural health clinics.7 They can also be capped by the performance payment limits discussed below and adjusted downward for failure to meet certain quality standards.
Performance Payment Limits
The Act provides that the Secretary is to establish limits on the total amount of shared savings that may be paid to an ACO.1,7 In the explanation of the Proposed Rule, CMS says that the cap should be high enough to “provide a significant opportunity for ACOs to receive shared savings generated from quality improvements and better coordination and management” of services, but not so high that it creates “incentives for excessive reductions in utilization.”7 On the theory that many health researchers believe that the rate of unnecessary healthcare services is more than 10%, CMS proposes to place the cap at 7.5% for Track 1 ACOs and 10% for Track II ACOs.7 These caps are based on the applicable benchmark. To take some examples, assume that the benchmark for a Track 1 ACO is $10,000 and average expenditures for the beneficiaries in the ACO are 10% below the benchmark ($9000). Assume further that the maximum shared savings rate for the ACO is 50% because the ACO does not qualify for a bonus for treating underserved patients. Leaving aside potential adjustments, the ACO will receive $500, or 5% of the benchmark per patient. To take another example, assume that a Track 1 ACO is able to reduce expenditures to 20% below the benchmark, or $8000. Based on the shared savings rate, the ACO would receive $1000, or 50% of the savings, but the 7.5% cap would limit the payment to $750. To make the financial picture even more daunting, CMS proposes to withhold 25% of the shared savings to insure that ACOs can repay Medicare any share of losses from exceeding their benchmark. As noted in the AMA comment, this has the effect of penalizing ACOs that do well in the first 2 years compared with ACOs with no savings in their first 2 years.
The Prospects for ACOs
The concept of ACOs is creative and there may be significant potential in this form of HCO. They are certainly worth a try. However, there are several potential weak nesses in the ACO paradigm:
1) The degree of integration required to become an ACO is far lower than that of a conventional IDN. This very permissive approach to ACO organization was no doubt an attempt by Congress to encourage as many organizations as possible to participate in the program and to avoid limiting it to established IDNs. However, it is unlikely that an organization that does not provide services at all basic levels of care and is too small to achieve the economies of scale of even a midsize IDN can reduce costs to Medicare significantly.
An ACO that offers a narrow range of services will find it difficult to steer its patients toward efficient ways of receiving care, because the statute guarantees patients the right to go to any provider. This right is restated in the Proposed Rule.7 While this “right” reflects prevailing political values, it makes it impossible to ensure that beneficiaries go to doctors or institutions that use evidence-based treatment protocols, meet minimum quality standards, or follow any of the requirements imposed on the ACO to provide high-quality, low-cost care. It is quite possible that a beneficiary can be assigned to an ACO because his primary care physician participates in the ACO, but still receive unnecessary, duplicative, or low-quality specialty care (all billed to Medicare) from physicians outside the ACO. The same is true if the beneficiary is an inpatient in a facility, such as a nursing home, that is not owned by (or under the control of) the ACO. Even if the beneficiary is an inpatient in a facility whose services are not covered by Medicare, the quality of care may be such that the beneficiary is forced to be readmitted to an acute care hospital, driving up costs to Medicare. These costs are then attributed to the ACO even though it had no control over the care at the other facility. A number of commenters argued that CMS should strengthen the relationship between beneficiaries assigned to the ACO and the providers in the ACO. See, for example, the AHIP Comment, “[B]eneficiaries should only be assigned to an ACO if a minimum threshold requirement is met, as well as the plurality standard.” The likely result is that ACOs can be successful only if there is a very high degree of integration.
2) Determining the amount of shared savings presents a number of analytical problems. The statute provides that the “benchmark” is per capita Medicare expenditures “adjusted for beneficiary characteristics and such other factors as the Secretary determines appropriate and updated by the projected absolute amount of growth in national per capita expenditures for parts A and B services under the original Medicare fee-forservice program.…”1 It is easy to calculate average per capita Medicare expenditures for identified beneficiaries. The difficult part is deciding how to make adjustments for inflation, beneficiary characteristics, and other factors. As discussed above, CMS proposes to calculate a benchmark based on costs to Medicare during a 3-year benchmarking period for beneficiaries who would have been assigned to the ACO if they had continued to use the same primary care providers. The proposed rule reflects a very serious and thoughtful effort to make this calculation. However, the approach proposed by CMS is quite complex, depends upon a number of assumptions, and requires an extensive data analysis. Given the small margins involved, a small overestimate of the benchmark (eg, 3%) could eliminate all potential shared savings.
3) Those HCOs that have made major efforts to reduce costs for Medicare beneficiaries in the past do not “get credit” for these cost reductions. They must reduce costs even further because their benchmark is based on the costs to the patients they have served during the benchmarking period. For example, the Cleveland Clinic Comment noted: “[W]e are concerned with measurements against the individual ACO’s adjusted benchmark, rather than a national or regional benchmark. We believe that being measured against one’s own performance is biased against providers who already manage care effectively and biased in favor of those who historically have done a poor job of care management.” A related problem is that HCOs that have been very inefficient may have a greater potential for shared savings than HCOs that have invested heavily in becoming more efficient. This anomaly follows from the fact that shared savings are based on the cost history of patients served by the ACO’s own primary care physicians.
4) Medicare reimbursement levels result from a combination of statutory formulas, historical baselines, and complex cost analyses (often using old cost data). Sometimes Medicare reimbursement rates seem reasonable, but often they seem wildly arbitrary. Many HCOs claim that Medicare reimbursement does not cover average costs for Medicare patients. They often assume that they must rely on other payers to make up a shortfall. In this reimbursement environment, it may be too much to expect even very efficient ACOs to reduce Medicare costs significantly. Of course, the issue is whether the ACO’s cost per beneficiary is below its benchmark, not whether the cost to treat a Medicare beneficiary is below its own costs. A very efficient ACO might lose only $500 per Medicare beneficiary, while a less efficient HCO might lose $1000. The problem is that, even if the efficient HCO is eligible for “shared savings,” it is still losing money on Medicare beneficiaries. Why then would it make a major effort to expand its Medicare patient pool?
On the other hand, the possibility that HCOs don’t make much money, and may even come out in the red, on Medicare patients, can make the shared savings paradigm more plausible. Consider a simple example in which an HCO is able to prevent a readmission to one of its hospitals by effective post-operative outpatient monitoring. The HCO prevents a readmission that would have cost Medicare $10,000. If we focus on these savings alone, assume that the HCO receives $5000 in shared savings. On the other hand, it has lost $10,000 in revenue. Why would an HCO have the incentive to prevent the readmission? Presumably, it would be concerned with the effect on its net, rather than gross, revenue. If the hospital only breaks even if the patient is readmitted, then the prospect of shared savings really is an incentive. If the hospital actually loses money on the readmission, the possibility of shared savings is an even stronger incentive. The problem is that this analysis may apply to the HCO itself, but it does not apply to individual physicians who participate in the ACO, as I discuss below.
5) Presumably, the bottom-line consideration of HCOs considering the prospect of participating in the ACO program is whether the potential for shared savings (discounted by the risk of shared losses) is greater than the costs of qualifying for the program and meeting all its regulatory requirements. CMS estimates the start-up costs for an ACO, including investment and operating costs, for the first year to be about $1.8 million.7 It estimates the median savings to Medicare for the first year to be $100 million, based on an assumption of 75 to 150 ACOs.7 If we assume that the median number of ACOs within this range is 112.5, the savings for a median ACO would be less than $1 million. In other words, ACOs lose money the first year. The savings go up each year and CMS estimates the total net savings to the federal government in the first 3 years to be $510 million.7 If we assume that the ACOs are rewarded with shared savings in the same amount (it could be a little more), and there are 112 ACOs, an average ACO would be rewarded with about $4.5 million over 3 years, or $1.5 million per year. The shared savings are modest, even if the ACO is fairly successful in reducing Medicare expenditures, while the regulatory burdens are significant.
Moreover, these estimates of the costs of establishing an ACO may be a substantial underestimate. The AHA retained a consulting firm to estimate the costs of developing the capabilities to be an ACO that meets CMS’s requirements. The consulting firm estimated that the combined start-up and first-year operating costs are $11.6 million for a small ACO and $26.1 million for a large ACO. In its comment, the AHA concluded: “The discrepancy between the investment required and the potential for reward imposes significant business risk for an ACO.”
6) Finally, it is not clear that ACOs do much about changing incentives to physicians to lower costs. There is no doubt that financial incentives can increase the alignment of physician behavior and the financial and quality goals of a healthcare organization.2-6 The question is whether the financial incentives inherent in the ACO program are significant enough to have any impact on physician behavior. The results of a CMS demonstration project were not encouraging. In the demonstration, 10 physician groups could receive shared savings if they lowered average Medicare care costs for assigned patients.10 While it is true that the 10 groups received a total of shared savings of $31.7 million, or $3.17 million on average, these numbers are misleading. Five of the 10 groups received nothing. Of the 5 groups that did receive shared savings, 1 group (Marshfield Clinic in Marshfield, Wisconsin) received $16.2 million, or over half the total.
If Marshfield is ignored as an outlier, the remaining 9 groups received an average of $1.7 million. In addition, the program provided for the groups to receive 80% of the shared savings, rather than the 50% to 60% proposed by CMS. These were large group practices ranging from 232 to 1291 physicians. Assume that the average size of the groups was about 760 physicians and that group allocated all of the savings to the doctors and did not keep any for administrative overhead or other purposes. In that case, an average physician (again leaving out Marshfield) was eligible for about $2240. It is not likely that an ACO would allocate all of the shared savings to its physicians. Moreover, this modest allocation of shared savings has to be seen in the context of the physician’s total patient load.
Based on the assumption that an ACO received shared savings of $1.5 million each year, an ACO with 10,000 beneficiaries would get $150 per patient. Assume that the ACO decides to allocate 50% of this amount to physicians and 3 physicians qualify for a reward for an average patient. In this example, a single physician is eligible to receive $25. Another example is in the AHA Comment, which assumes savings of $40 per patient. CMS then withholds 25% of that. If we assume, as in the example above, that the ACO keeps 50% of the savings and divides the remainder among 3 physicians for each patient, and we deduct 25%, a physician is awarded $5 per patient. It is impossible that these levels of reward could change physician behavior in any significant way. In fact, a physician who performs a single unnecessary physician service per patient would come out substantially ahead.
It remains to be seen whether many existing or newly formed healthcare organizations will find the potential benefits of the program worth its financial costs and regulatory burdens. Based on the potential financial gains, it seems unlikely. The program is intended to go into effect in January 2012, and it is likely that many HCOs will take a wait and see attitude before making a major commitment to qualify for the program during the next year and a half. On the other hand, CMS may decide to revise the regulations before they are put in final form, and the requirements may not be so formidable. Also, some existing organizations, such as large integrated delivery networks, may believe that they can operate significantly more efficiently than less integrated providers and will be willing to try the program at least for a 3-year period.
If ACOs are not the solution to the problem of healthcare costs, then what is? In order to answer this question, it is helpful to look closely at the relationship between the healthcare organizations and physicians. Presumably, the vast majority of physicians do not deliberately order unnecessary tests, perform unnecessary services, or demand that the hospital purchase unnecessary drugs and devices. In general, however, they are not financially accountable for expenditures by hospitals and they feel free to exercise their clinical discretion. Hospital administrators are typically non-physicians, who do not feel that they have the expertise to overrule physicians. Even if they are physicians, their specialty may be different from that of the treating physician. And, even if they have the same specialty, they may feel that that they are not familiar enough with the details of the case to question the treating physician’s judgment. At the same time, the treating physicians may threaten, expressly or by implication, that they will terminate their relationship with a particular hospital and take their patients elsewhere. Hospitals, who usually want more, not fewer patients, are not inclined to take that risk. It is doubtful that the shared savings potential of ACOs can alter this basic dynamic.
Author Affiliations: From Correia & Associates, PC, Washington, DC; Washington College of Law, American University, Washington, DC.
Funding Source: None reported.
Author Disclosure: The author reports no relationship or financial interest with any entity that would pose a conflict of interest with the subject matter of this article.
Authorship Information: Concept and design; acquisition of data; analysis and interpretation of data; and drafting of the manuscript.
Address correspondence to: Edward W. Correia, JD, MP, Correia & Associates, PC, 1101 30th St, NW, Ste 500, Washington, DC 20007. E-mail: email@example.com.
1. Patient Protection and Affordable Care Act, Pub L. No. 111-148, 124 Stat. 119 (2010).
2. Devers K, Berenson R. Can accountable care organizations improve the value of health care by solving the cost and quality quandaries? Robert Wood Johnson and Urban Institute. Timely Analysis of Immediate Health Policy Issues series. 2009:10.
3. Shortell SM, Casalino LP, Fisher E. Implementing accountable care organizations. Advancing National Health Reform policy series. http:// www.law.berkeley.edu/files/chefs/Implementing_ACOs_May_2010.pdf. Published May 2010.
4. DeVore S, Champion RW. Driving population health through accountable care organizations. Health Aff (Millwood). 2011;30(1):41-50.
5. Goldsmith J. Accountable care organizations: the case for flexible partnerships between health plans and providers. Health Aff (Millwood). 2011;30(1):32-40.
6. Lierberman SM, Berko JM. Building regulatory and operational flexibility into accountable care organizations and “shared savings.” Health Aff (Millwood). 2011;30(1):23-31.
7. US Department of Health and Human Services Centers for Medicare & Medicaid Services. Medicare shared savings program: accountable care organizations. Fed Reg. 2011;76(67):19,528. http://www.gpo.gov/fdsys/pkg/FR-2011-04-07/pdf/2011-7880.pdf. Accessed July 11, 2011.
8. US Department of Health and Human Services Centers for Medicare & Medicaid Services. Waiver designs in connection with the Medicare shared savings program and the innovation act. Fed Reg. 2011;76(67):19,655. http://www.gpo.gov/fdsys/pkg/FR-2011-04-07/pdf/2011-7884.pdf. Accessed July 11, 2011.
9. Federal Trade Commission, US Department of Justice Antitrust Division. Proposed statement of antitrust enforcement policy regarding accountable care organizations participating in the Medicare shared savings program. Fed Reg. 2011;76(75):21, 894. http://www.gpo.gov/fdsys/pkg/FR-2011-04-19/pdf/2011-9466.pdf. Accessed July 11, 2011.
10. Iglehart JK. Assessing an ACO prototype — Medicare’s physician group practice demonstration. N Engl J Med. 2011;364(3):199.