Keywords

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Introduction

The Affordable Care Act substantially raised the national profile of a new healthcare delivery system financing model called accountable care organizations (ACOs). Put very simply, ACOs are groups of healthcare providers that join together and agree to be financially and clinically accountable for patients who seek most of their care from them. While the ACO is considered a relatively new approach, its origins can be traced back to a much earlier period in American healthcare history.

Brief History of the Medicare Program

By the early 1950s in the United States, several attempts to institute a major national health insurance program had surfaced and fizzled. Significant health insurance proposals had emerged under the Theodore Roosevelt, Franklin Roosevelt, and Harry Truman presidencies, but partisan fighting and consistent opposition from the American Medical Association, which viewed the nascent efforts as “socialistic” and financially detrimental to physicians, ultimately stymied efforts [1].

By the 1950s however, the political climate was ripe for consideration of a more limited national insurance program restricted to the elderly – a group widely accepted to be underinsured and burdened by poverty and sickness. At that time, almost half of the elderly population in the United States lacked health insurance [2]. To make the proposal more palatable, proponents initially suggested that the insurance program cover only hospitalization services [1].

After years of political maneuvering and interim steps on the path to a national program, in 1964, supporters of a federal Medicare program had secured both the presidency and a majority in the House and Senate; this advantageous political climate enabled program supporters to propose a more far-reaching program than one restricted to hospitalization only. The resulting Social Security Act amendments in 1965 included a so-called “three layer cake” [3] of programs. One layer “Part A,” consisted of coverage for hospitalization; another layer, “Part B” was a voluntary program that required beneficiaries to pay premiums in return for coverage of physician visits; and the last layer, a joint federal-state initiative focused on the poor, would become the Medicaid program [1].

Medicare FFS Versus Managed Care History and Payment Mechanics

At its inception, Medicare was predominantly a “fee-for-service” (FFS) program, meaning that the federal government would pay a fee to healthcare providers for each service rendered. In the Part A program, institutions were paid based upon the costs they incurred and in the Part B program, physicians were paid “allowed charges,” defined as the customary, prevailing charge for such services [4].

Some decades earlier, an alternative to FFS medicine had emerged, primarily for employed populations: the concept of prepaid health plans, an early type of managed care plan. During the Great Depression, several doctors – Drs. Michael Shadid, Donald Roos, H. Clifford Loos, and Sidney Garfield – had all developed subscription like models in their respective geographies through which workers and their families would pay the doctors a set monthly fee in return for medical care, when needed [5]. Ultimately, individual fiefdoms of doctors banded together to create entire pre-paid networks of medical providers, the early prototype of what would become Health Maintenance Organizations (HMO). Famous group practices that launched such pre-paid plans included Kaiser Permanente in California, Group Health Association in Washington D.C., Group Health Cooperative in Washington state, and the Health Insurance Plan of New York. In these cases, the entity managing healthcare finances was also responsible for actually delivering clinical care [1].

While these plans were able to enroll sizable numbers of beneficiaries in relatively confined geographic areas, medical societies, including the American Medical Association (AMA), prevented widespread adoption of such plans, expressing concerns that business staff would interfere with medical practice and that medical professionals taking on business responsibilities would engage in improper contracting practices [5]. Consequently, when the Medicare program was established, the notion of managed care was still relatively limited; in the early 1970s there were under 50 HMOs nationally [6].

Managed care in the Medicare program expanded in the 1970s, when the government implemented demonstration programs that provided prepayments to HMOs, organizations responsible for operating networks of providers available to deliver a comprehensive set of medical services to beneficiaries [7]. Much like any other kind of budget, prepayments set a financial ceiling and then deferred to plans and providers to determine an appropriate allocation of funds underneath that ceiling. The prepayments provided the benefit of allowing the government to proactively budget for patient healthcare costs, instead of waiting for costs to accrue on a FFS basis. The prepayment demonstrations also coincided with a broad national effort to expand HMOs–the HMO Act of 1973, which provided $375 million in funds to support the expansion of HMOs, through grants, contracts, and loans. The Act also required employers to provide an HMO option to employees [8].

By the mid-1980s, the Medicare risk based contracting demonstrations became a permanent fixture of the Medicare program, in part due to a growing body of research indicating that HMOs reduced healthcare costs [7]. A famous randomized controlled trial that compared participants in the Group Health Cooperative of Puget Sound, one of the early prepaid physician groups, to individuals seeking care FFS found seemingly impressive impacts –the expenditure rate for all healthcare services was 25 % less among those receiving services from the Cooperative compared to the FFS group [9]. Of note, patients enrolled in the HMO product were less satisfied than their FFS counterparts, perhaps signifying that individuals highly valued the unrestricted FFS provider networks [7].

Cost Pressures and Medicare Managed Care Expansion and Contraction

Over time, managed care became a conceptually bigger part of the Medicare program, particularly as cost pressures became more acute. Post 1965, the cost of the Medicare program far exceeded any predictions. There were several reasons for the outsized cost growth including an initial pent-up-demand for healthcare among the elderly who gained coverage under Medicare; a Part A payment structure that encouraged hospitals to provide a high volume of services; and the fact that increases in beneficiary payments (premiums) paid to the Part B program were tied to inflation and yet Part B cost growth far exceeded inflation growth [2]. By the mid 1990s, Medicare’s share of the federal budget had more than doubled since the program began and was the third largest component of the federal budget [10].

Policymakers initiated several major payment reforms to mitigate cost growth, such as the inpatient prospective payment system and the physician fee schedule, both of which set out to bring order to Part A and B payment policies and reduce incentives to inappropriately increase volume of service use among beneficiaries [7]. Another increasingly attractive cost containment tool was managed care.

HMOs were appealing because of studies like the one described earlier that suggested that HMOs could deliver care more efficiently, at lower cost. Reflecting this belief, initially prepayments to HMO plans (also called capitated payments) were pegged to 95 % of the average FFS Medicare costs in the county where the plan was operating [11]. In setting the payment rate at 95 % of the expected FFS costs in a given county, policymakers thought they would be preemptively achieve cost reductions.

The Balanced Budget Act of 1997 (BBA) entrenched the presence of HMOs and other types of managed care plans in the Medicare program even further, formally establishing the Medicare Part C program, known as Medicare Advantage today [12]. Beneficiaries could elect to enroll in a managed care plan, Part C, or remain in traditional FFS Medicare, Parts A and B. In addition to HMOs, the BBA allowed the Medicare program to include a number of other types of managed care plans that had proliferated in the private market and offered various types of networks and approaches for managing utilization of healthcare services.

At the same time as attempting to expand the program, policymakers also sought to right-size payments to managed care plans. Despite best efforts to build cost reductions into capitated payments, in practice, HMOs did not reduce costs. Initially, Medicare HMO payments were tied to average FFS costs incurred by beneficiaries, both healthy and sick, in a given county. The payments assumed that HMO plans would enroll an average cross section of beneficiaries, both healthy and sick. But to the extent that HMO plans enrolled a relatively healthy population, they would essentially be overpaid because the underlying payment reflected costs associated with some sicker beneficiaries. This scenario is exactly what happened– sicker beneficiaries tended to remain in Medicare FFS, while healthier beneficiaries enrolled in managed care; some estimates suggested that the government overpaid managed care plans by as much as $2 billion [13].

The BBA attempted to reign in some of this inappropriate spending by limiting payment increases in geographic areas with relatively high HMO prepayments [7]. Facing this reduction, a number of managed care companies withdrew from markets entirely, thereby involuntarily dis-enrolling sizable numbers of patients. In other cases, to make the new economics work, Medicare managed care plans curtailed benefits available to beneficiaries, reduced payments to providers, or instituted additional steps before beneficiaries could access care, like requiring primary care providers to serve as “gatekeepers,” to specialty care. Practices like these, which were widespread in the private market as well, resulted in a significant public backlash against managed care [14]. By the early 2000s, 12 % of Medicare beneficiaries were enrolled in managed care plans instead of FFS; the managed care enrollment rate had actually declined since the passage of the BBA [7].

Subsequent legislation, like the 2003 Medicare Modernization Act (MMA), increased payments to Medicare managed care plans to revive the role of private plans in Medicare and alleviate the cost pressures that had precipitated the earlier backlash. Enrollment in Medicare managed care did in fact rise after its passage, tripling between 2004 and 2013 [11]. In adjusting the payment to plans, however, the MMA further eroded the short-term prospect of managed care as a cost containment tool in the Medicare program. One analysis found that Medicare spent an additional $922 on average for Medicare managed care enrollees compared to comparable beneficiaries in Medicare FFS, leading to extra payments in excess of $5.2 billion by 2005 [15].

Provider-Based Accountability: A Throwback

By the early 2000s, the viability of managed care as a cure-all for reducing Medicare expenditures had diminished, but the cost pressures facing the Medicare program had not eased. In addition, despite the growth of managed care plan enrollment after the MMA’s passage, the majority of Medicare beneficiaries were not enrolled in health plans, but rather remained in the program’s traditional FFS program. Seeking to experiment with non-HMO/managed care models to reduce costs and improve quality in the FFS context, in 2005, the Centers for Medicare and Medicaid Services launched the Physician Group Practice (PGP) demonstration.

This demonstration allowed ten large physician group practices, six of which were multi-specialty practices and one of which was a physician-hospital organization [16], with at least 200 participating providers to access savings relative to a pre-determined spending benchmark associated with Medicare FFS beneficiaries who sought care from their providers. Savings were also tied to provider performance on 32 quality metrics [16].

Sidebar: PGP Demonstration Outcomes

The PGP demonstration ran from 2005 to 2010, with a 2-year extension after 2010. Results from the demonstration were positive from a quality perspective – all of the participating ten groups met nearly all of the quality metrics (29 out of 32 metrics) – but the financial outcomes were more modest. In order to access shared savings, demonstration participants had to both meet quality outcomes and achieve a minimum savings rate of 2 %. Half of the demonstration participants saved more than 2 % more than halfway into the demonstration [17, 18].

Around the same time as the PGP demonstration, the term “accountable care organization” began to enter the healthcare lexicon. One of the first explicit national discussions of an ACO model emerged at a November 2006 meeting of the Medicare Payment Advisory Commission (MedPAC), an independent congressional agency tasked with advising Congress about issues pertaining to the Medicare program. MedPAC had been directed by Congress to examine alternatives to the Sustainable Growth Rate (SGR) system, which was intended to adjust physician payments on the basis of changes in input prices, growth in Medicare FFS enrollment, and increases in physician service volume compared to national economic experience [19]. Over time, the SGR system had created a system that dictated physician fee cuts that in the words of one expert, far “exceed[ed] the magnitude of the willingness to cut fees” [19]. During the meeting, Dr. Elliott Fisher, Professor of Medicine at Dartmouth Medical School, surmised that part of the solution would involve an attempt to “strengthen local organizational accountability for the decisions that drive higher costs and worse quality [20].”

Dr. Fisher outlined a process for creating virtual organizational structures (as opposed to established physician groups – the basic organizational unit in the PGP demonstration) to take on financial and clinical accountability; he suggested that such organizations were important because small groups of providers could not significantly influence cost and quality outcomes and that there were relatively few large multispecialty practices in the United States. To create such structures of providers – “extended hospital staff” – he commented that nearly all physicians could be attributed to a hospital either by virtue of being employed by the hospital or because a majority of the patients the physician saw were admitted to a particular hospital when seeking inpatient services.

Second, most beneficiaries could be assigned to a “predominant care physician,” either a primary care provider or a specialist that accounted for most of the care they would receive in a given time period. Because of these linkages, he argued, medical groups consisting of diverse arrays of physicians and an anchoring hospital, could reasonably be held accountable for the cost and quality outcomes associated with attributed beneficiaries [20].

By first creating loose organizational structures, borne out of imputed physician relationships to particular hospitals and beneficiary ties to those providers, Dr. Fisher and other meeting participants moved the national dialogue closer to the current incarnation of accountable care organizations. In many ways, by conferring financial and clinical responsibility upon a single organization, the ACO model resembled the early pre-paid physician group practices and HMOs, without the network limitation features that had led to a managed care criticisms in the 1990s [21].

The Affordable Care Act and ACOs

Nearly 300 pages into the Affordable Care Act text, drafters picked up the thread from the PGP demonstration and the MedPAC discussion in a section titled “Encouraging Development of New Patient Care Models.” While the provisions in this section generated less public attention – and controversy – in the lead up to the law’s passage than provisions pertaining to the health insurance exchanges, collectively, its implications were arguably just as sweeping [22].

Section 3021 of this portion of the act established a Center for Medicare and Medicaid Innovation (CMMI) to experiment with innovative payment and service delivery models focused on reducing Medicare and Medicaid program expenditures, while preserving or ideally enhancing the quality of care provided to beneficiaries. The guiding principle behind CMMI’s initiatives is a framework known as the triple aim. The triple aim, developed by the Institute for Healthcare Improvement, a Massachusetts-based non-profit dedicated to advancing health care systems throughout the world, includes the following tenets [23]:

  • Improving the patient experience (including quality and satisfaction)

  • Improving the health of populations

  • Reducing per capita cost of health care

In IHI’s formulation, these three aims collectively maximize the performance of health systems: The Act appropriated no less than $10 billion dollars [24] between fiscal years 2011 and 2019 for the fledgling center to meet this call to action.

The Act also enumerated the center’s portfolio of activities, which included the promulgation of accountable care organization (ACO) models. In the Act’s formulation, outlined in section 3022, Medicare Shared Savings Program (MSSP) ACOs would be comprised of various groups of providers with shared governance structures; that would be “willing to become accountable for the quality, cost, and overall care of the Medicare FFS beneficiaries” assigned to these groups. The Act also held open the possibility for the Secretary of Health and Human Services (HHS) to test a novel variation on the MSSP that would enable highly integrated delivery systems, rather than health insurance companies, to take on partial capitation [22]. While the law delineated the broad financial parameters of the program, it did not go into great depth about how the program would be operationalized or clinical expectations.

Detail arrived a little over a year later when the Centers for Medicare and Medicaid Services released what is known as a (NPRM) codifying section 3022 of the law [25]. Once Congress enacts laws, federal agencies, like the U.S. Department of Health and Human Services, derive authority from the enacted law to issue regulations that detail how the agency intends to implement its provisions [26]. Before regulations are finalized, agencies must seek public input on a proposed version of the regulation [27].

CMS issued its proposed rule on ACOs in April of 2011 [25]. Among other areas, the NPRM sought public input on the idea of creating two ACO options, so as to encourage the broadest possible range of provider groups to participate. Option one, the MSSP, included a one-sided model through which groups of providers that sufficiently managed beneficiaries’ expenditures underneath a pre-determined threshold could share in those savings. The model was considered one-sided because participating providers could only gain financially or, at worst, remain neutral, but they would not bear any financial losses as a result of the program [25].

Capitalizing on language in the ACA enabling the Secretary of HHS to test a variation of MSSP, the NPRM also detailed specifics of a “two-sided model” that HHS would offer, called a Pioneer ACO, that would allow organizations with more experience managing financial risk to take a bigger cut of any savings reaped, but also to be accountable for a portion of losses, if incurred [25]. The second option was geared toward systems that already had years of experience taking on financial risk [28].

By early 2012, the two programs had officially launched, with the Pioneer ACO program beginning in January 2012 and the MSSP program starting in April 2012 [29]. Several of the physician groups that had participated in the PGP demonstration elected to participate in a transitional program that aligned with MSSP parameters or the Pioneer ACO program [30]. ACOs were no longer mythic “unicorns” as some healthcare commentators had jokingly termed the much talked about but yet to be implemented model [31]. The text below outlines key features of the two programs.

Key Features of Medicare Pioneer and MSSP ACO Programs

Provider Participation and Length of Programs

Groups of healthcare providers and hospital systems can join together to form a Medicare ACO. Critically, participants in either the MSSP or Pioneer program must have a Medicare-enrolled Tax identification number (TIN); ACOs may comprise a single TIN or multiple TINs [32]. Specifically, physician group practices, provider group organizations (PPOs), independent physician associations (IPAs), employed staff in medical organizations, joint ventures between hospitals and physician organizations, as well as some critical access hospitals, rural health clinics, and federally qualified health centers (FQHCs), can apply to participate in the Medicare ACO programs [33].

While groups of providers with multiple TINs can apply collectively as a single ACO, the ACO must also have a single governing body that can contract with CMS. While CMS does not strictly define a minimum or maximum number of participating providers, applicant ACOs are expected to represent certain minimum thresholds of Medicare beneficiaries aligned with their providers. MSSP programs are expected to be accountable for 5,000 beneficiaries whereas Pioneer programs in non-rural areas must be accountable for 15,000 beneficiaries (see subsequent section for detail on how patients are “aligned” with Pioneer and MSSP programs) [34].

In 2011, CMS issued a request for applications to the Pioneer ACO program [35] and by the end of the year, CMS selected 32 organizations nationally to participate in the 3-year initiative, with an option at CMS’ discretion to continue for two additional performance years if the program met its performance objectives [36]. In early 2014, CMS issued a request for information seeking feedback on a future Pioneer ACO solicitation and how the current cohort of Pioneer ACOs may evolve over time [37].

Under the MSSP program – a permanent program rather than a demonstration like the Pioneer – CMS has selected four cohorts of participants, two in 2011, one in 2012, and two in 2013, as well as recently closing a solicitation for 2014 applications [37]. MSSP agreements, like the Pioneer, cover three performance year periods [38]. To date, there are over 350 such ACOs, including some advanced payment ACOs, which is a variation of the MSSP program that includes some start up payments for ACO formation that are recouped out of shared savings, if achieved [38].

Patient Alignment and Engagement

Unlike a managed care or health plan model, beneficiaries do not enroll in an ACO. Rather, much like the PGP demonstration and the ACO concept as outlined by Dr. Fisher, defined populations of beneficiaries are aligned with particular ACOs. CMS has developed a methodology for analyzing individuals’ historical utilization of particular Medicare providers and then determining primary healthcare providers to whom these individuals appear to be linked [39]. The intention of this imputed connection versus an enrollment model, is that it enables CMS to designate a locus of care responsible for coordinating a beneficiary’s services, without in any way modifying the individual’s network of providers. Aside from improved care coordination, beneficiaries assigned to an ACO should not observe changes to their benefit package or network of providers.

To align beneficiaries, CMS examines 3 years of historical service utilization data among Medicare FFS beneficiaries and then determines ACO applicant providers from whom beneficiaries have received the preponderance of their primary care (as determined through a list of “qualifying” Evaluation and Management codes). While the methodology focuses on isolating relationships between beneficiaries and primary care providers, CMS does incorporate beneficiary utilization of certain types of specialists such as nephrologists, oncologists, rheumatologists, endocrinologists, pulmonologists, neurologists, neuropsychiatrists, and cardiologists [39].

A key difference between the MSSP and Pioneer programs is that in the Pioneer program, ACOs can choose to have beneficiaries aligned prospectively. At the start of a performance year, Pioneer ACOs choosing this option will know the universe of beneficiaries for whom they will be fiscally and clinically responsible. A prospective alignment model enables Pioneer ACOs to target high cost, high need beneficiaries at the beginning of the performance year and manage their care throughout the entire period [40, 41].

By contrast, CMS uses retrospective alignment in the MSSP program, which is the approach that was also used in the PGP demonstration; Pioneer ACOs can also elect to have a retrospective alignment methodology, though it is not publicly known if any Pioneer ACOs have selected this option. Under the retrospective approach, CMS presents participating ACOs with a preliminary list of attributed individuals and then updates this list quarterly based upon actual service utilization until finalizing the alignment at the end of the performance year [41]. Retrospective alignment necessitates a broader population health strategy because ACOs do not know whom they will be financially responsible for in advance.

Financial Model

Both the MSSP and the Pioneer program are shared savings programs. If ACOs manage beneficiary healthcare costs beneath an expenditure benchmark, while meeting defined quality expectations, they can share in or access a portion of the dollars under the benchmark threshold. Regulation drafters sought to devise a shared savings methodology that would safeguard against inappropriate activities to bring down costs, like setting up barriers to access or reducing the quality of services, and protect against ACOs unfairly benefitting from overall trends in the market (e.g. a general national decline in Emergency Room utilization).

For the Pioneer and MSSP programs, CMS uses 3 years of average Medicare Part A and B expenditures for ACO-aligned beneficiaries to develop a financial benchmark [41]. In the MSSP program, the benchmark continues for the duration of the participation agreement with CMS (3 years) and similarly in the Pioneer program, the benchmark remains in place for 3 years and is recalculated in the fourth year of the demonstration. At a very high level, both programs take steps to adjust benchmarks for differential risk profiles of attributed beneficiaries, acknowledging that different age, sex, and disability sub-groups may incur very different expenditures [41]. Additionally, like the PGP demonstration, both programs have minimum savings rates (MSRs) that ACOs must surpass before accessing any savings or experiencing any losses; these MSRs are meant to protect against minor variations in expenditures year over year [41]. Underneath these general commonalties, there are a few key differences between the MSSP and Pioneer benchmark methodologies [41, 42]:

  • Risk Levels: The Pioneer program involves greater levels of financial risk and savings opportunity in the initial years of implementation than the MSSP program and is therefore meant for organizations with prior experience executing ACO-like arrangements.

  • Population-Based Payments: In the third year of the Pioneer demonstration (2014), certain ACOs were eligible to transition to a population-based payment, which involves receiving a portion of the FFS benchmark in advance on a monthly basis, similar to capitated payments. MSSP ACOs cannot access this payment option.

  • Performance-based payment contracts: Pioneer ACOs are required to receive at least 50 % of their overall revenues through outcomes based payment arrangements such as shared savings deals; this requirement is premised upon the idea that if Pioneer ACOs substantially move their business model to such arrangements, it will better promote the triple aim. MSSPs do not have to meet this requirement, presumably because the model is focused on delivery systems with less risk experience.

Physician Payment

In the MSSP program and during the first 2 years of the Pioneer demonstration, physicians are paid as they usually are within the Medicare FFS program. However, as noted above, Pioneers that achieve certain levels of shared savings may receive population-based payments or pre-payments in the third year of the demonstration. With this flexibility, ACOs could theoretically choose to pay physicians differentially, though CMS has not indicated which if any ACOs had taken that step (physicians must also agree to participate in this payment structure). In the request for information released in early 2014, CMS solicited feedback from the field about evolving to ACO models with even greater levels of financial risk that would further enable ACOs to develop creative physician payment mechanisms [37].

Notably, the MSSP and Pioneer programs waive certain federal laws for the purposes of meeting the triple aim. Among other areas, ACOs are permitted to gainshare with participating providers [43]. Gainsharing, broadly speaking, is defined as delivery systems distributing savings accrued from cost reductions to healthcare providers who have helped generate those reductions. Typically, the Department of Health and Human Services has been wary of allowing such payments because such financial incentives could inappropriately induce physicians to limit patient care in order to cut down on costs [44]. In the context of the ACO program, however, such payments are expected to incentivize maximal coordination of patient care across settings, while quality performance standards safeguard against inappropriate reductions in care.

Quality Monitoring

In order to ensure that ACOs achieve cost reductions in a manner consistent with good clinical practice, CMS requires ACOs to meet several quality metrics, similar to the approach in the PGP demonstration. The 33 metrics in the MSSP and Pioneer program encompass a range of nationally accepted process and outcome metrics across the following four categories [45]:

  • Patient/caregiver experience

  • Care coordination/patient safety

  • Preventive health

  • At-risk population:

    • Diabetes

    • Hypertension

    • Ischemic Vascular Disease

    • Heart Failure

    • Coronary Artery Disease

For the most part, CMS selected measures from among those already used today in other CMS programs such as the Electronic Health Record (EHR) Incentive or the Physician Quality Reporting System programs. Even so, the ACO programs have offered an opportunity to advance the field of knowledge about these measures, in that a number have never previously been applied to a FFS population or have never been deployed nationally before. CMS is using findings from these programs to inform reasonable thresholds for quality performance [45].

Because of the experimental nature of several of the measures, shared savings are not immediately tied to actual quality performance by the ACOs. In the first year of the Pioneer demonstration and the first year of any MSSP initiation, ACOs are required to report on all quality metrics. In the second year, 25 of the 33 measures are “pay for performance,” or impact the amount of savings retained, and finally in the third year, 32 of the 33 measures are pay for performance [45]. Performance is based upon patient survey data, claims and administrative data from CMS, and then data the ACOs must directly collect and report upon [46].

Care Coordination

Part of the rationale – if the not the most significant reason – for initiating ACO models at the federal level was a recognition that Medicare FFS beneficiaries are often subject to fragmented care. The Medicare Payment Advisory Commission has found that Medicare FFS beneficiaries frequently receive duplicative medical tests, receive inconsistent medical information or even different diagnoses from providers, and seek care from “higher-intensity” settings, like the emergency departments, than is warranted by their condition [47].

The ACO programs seek to address this fragmentation by stimulating groups of providers to better coordinate care for groups of FFS beneficiaries across healthcare settings. The federal programs promote better coordination through enabling gainsharing amongst diverse providers, setting quality reporting and performance standards that embed cross-system collaboration, and requiring the establishment of governance structures that include representatives across a given delivery system or provider organization.

However, beyond those parameters, the ACO programs essentially defer to the participating providers to determine how to best coordinate care for beneficiaries – the models certainly do not call for particular clinical pathways or care management structures. With that latitude, ACOs have pursued a multiplicity of approaches to improve care coordination and reduce fragmentation. The following chapter provides a detailed case study of care management activities at Montefiore Medical Center, a Bronx, New York-based academic medical center that is implementing a Pioneer ACO model.

Other ACO Models

ACOs are not limited to the Medicare FFS program. ACOs have also proliferated nationally within Medicaid programs and amongst commercial payers. By some estimations, there are over 600 public and private payer ACOs nationally [48]. A number of state including Utah, Colorado, Oregon, and Minnesota have advanced models through their Medicaid programs designated to delegate financial and clinical risk to provider groups [49]. Managed care plans in some case have even advanced ACO like models, developing shared savings arrangements with contracted provider networks. As one example, Blue Cross Blue Shield of Massachusetts (BCBS) cultivated an alternative quality contracting (AQC) model through which it would provide a global budget to sub-contracted providers to manage all costs of their patients, while meeting quality targets. BCBS of Massachusetts worked closely with CMS in the development of the ACO programs, building upon lessons learned from the AQC model [50].

Conclusion

ACOs are not unicorns, but time will tell if they are in fact thoroughbreds – reliable cost-cutting, quality-enhancing programs, worthy of significant national expansion. Early information on the Medicare ACO programs is promising. Results from the first year of the Pioneer program showed that 13 of the 32 participating ACOs yielded $87.6 million in gross savings in 2012, translating into $33 million for the Medicare Trust Fund and shared savings amongst the Pioneers of $76 million (disclosure: Montefiore was the top financial performer amongst the Pioneer ACOs in the first demonstration year). Shared losses were more modest, totaling $4 million [51].

A subsequent independent analysis requisitioned by CMS that used a comparison group analysis instead of the benchmark methodology employed in the MSSP and Pioneer programs, also verified substantial savings associated with the two programs [52]. Little, however, is known about the infrastructure costs individual ACOs have incurred by instituting these programs or the structural features that increase the likelihood of clinical or financial success. More research and time is needed to fully appreciate the impacts of ACOs, both inside and outside the Medicare program. Notwithstanding that research gap – and the meandering path to our present day ACO models– what is evident is that the volume-driven FFS reimbursement framework once so foundational in the nation’s healthcare system is slowly becoming a relic of days past.