What is a hospital? Future roles and prospects for success: the business of medicine: a course for physician leaders.

As hospitals consolidate and take on more financial and clinical risk, they face numerous obstacles. While the past can provide answers to solving many of the challenges, some issues are new and require innovative approaches. This article, from a speech delivered to The Business of Medicine: A Course for Physician Leaders symposium presented by Yale-New Haven Hospital and the Medical Directors Leadership Council at Yale University in November 2012, discusses the models for these hospital organizations and the pitfalls they will face in coordinating care. The insights will help these systems overcome potential problems and enhance their chances of success.

Defining a hospital has shifted from using individual, often idiosyncratic, state licensing laws to understanding how these institutions are aggregating into systems to deliver care. Compared to traditional stand-alone hospitals, these systems are more complex organizationally, more geographically dispersed, provide more services along the continuum of care, and accept some financial risk for the provision of care. The term for these systems 20 years ago was Organized (Integrated) Delivery Systems (ODSs); they are now called Accountable Care Organizations (ACOs). ODSs did not achieve their potential, and many of the challenges they failed to resolve remain for ACOs. In addition, legislation and market forces have raised some new concerns for ACO success. This paper will explore the origins of these systems, terms of ACO participation, the historic (but still relevant) reasons many ODSs failed, and new difficulties ACOs face. Through elucidation of these issues, ACOs can better anticipate and address the problems, thereby enhancing their likelihood of success.

Origins and Definition

In the 1990s, a number of researchers conducted studies on what were termed Integrated, or Organized, Delivery Systems. An ODS is defined as: “A network of organizations which provides or arranges to provide a coordinated continuum of services to a defined population and is willing to be held clinically and fiscally accountable for the outcomes and health status of the population served. An ODS will own or be closely aligned with an insurance product” [1] .Their formation was motivated by four aims:

Economies of Scale. The major savings from this benefit come from sharing support functions like payroll, accounting, logistics management, and volume purchasing (though this latter advantage can also accrue from membership in a group purchase organization, or GPO). Sharing equipment, like centralized computers and vehicles, can also achieve this goal.

Economies of Scope. When two or more different services can be produced using common resources at a lower cost than if created individually, then economies of scope exist. For example, a radiology department can produce a variety of diagnostic and therapeutic services using the same staff and equipment. Another way to look at scope is when system members are geographically close (a relative term that will vary case by case), they can diversify the types of services each offers without the expense of duplication. For example, one facility may provide high level neonatal care, while another might offer invasive cardiac treatments. This advantage is not always possible, since expensive diagnostics (like CT and MRI scanners) often need to be available on-site in the event of an emergency. As well, system hospitals may be located across a large geographic area, requiring each to be relatively self-sufficient with a diverse portfolio of services.

Capture Populations. It is important for systems to develop patient loyalty so that no matter what services are needed, the organization stays top of mind. They can accomplish this goal by developing strong brand recognition through enhanced quality reputation, easy access, and a portfolio of desired services.

Market Power over Payers. This advantage is a separate outgrowth of the previous one, but can be an independent motivator for system formation. With convenient access across a wide geographic area and a large, loyal patient base, insurance companies will need to include these systems in their networks. The one caveat systems face in this strategy is the possibility of antitrust (see below).

Despite the success of such systems as Geisinger Health System, Intermountain Healthcare, Cleveland Clinic, Mayo Clinic, and others, for reasons explained below, ODSs did not fulfill the promise of transforming the way care was delivered in the United States. These efforts, however, were mostly private initiatives. The Patient Protection and Affordable Care Act (ACA) created a public program to use these ODSs (now called Accountable Care Organizations, or ACOs) to address large regional variations in cost and quality of care for Medicare beneficiaries. The ACA’s goal is to reward quality performance and operational efficiency rather than sheer volume of services.

In addition to addressing Medicare issues, other provisions in the ACA create a number of constraints on the profitability of private payers, e.g., regulated premiums, reduced ability to charge different rates to older and higher risk members, mandated benefits, limits on members’ out-of-pocket payments, and bounds on profit margins. Consequently, these commercial plans are also very interested in shifting more risk to providers. The result is private ACO arrangements, whereby the health delivery system takes substantial clinical and financial risk for patient care.

Since private payer contracts can vary considerably, the next section focuses on requirements of Medicare ACOs, although the same principles can apply to private contracts. According to the Centers for Medicare and Medicaid Services (CMS): “ACOs are groups of doctors, hospitals, and other health care providers, who come together voluntarily to give coordinated high quality care to the Medicare patients they serve” [2].

Eligibility

The ACA allows an ACO to be comprised of:

Group practices of statutorily defined professionals;

Networks of individual practices of such professionals;

Partnerships or joint venture arrangements between hospitals and professional organizations;

Hospitals that employ these professionals; or

Other Medicare providers and suppliers as determined by the Secretary of HHS.

CMS identifies providers using their National Provider Identifier (NPI) and Taxpayer Identification Number (TIN) ― either Employer Identification Number or Social Security number. While specialists can contract with multiple organizations, primary care doctors can only join one ACO. Members of a group with one TIN must join the same ACO regardless of their office locations. The only way to avoid this situation is if each group member has a different TIN.

Operational Requirements of Medicare ACOs

According to CMS [3], to participate in the Shared Savings Program (explained below), an ACO must:

Accept responsibility for at least 5,000 Medicare Fee-For-Service beneficiaries.

Sign an agreement with CMS to participate for a period of at least 3 years.

Have in place procedures and processes to promote evidence-based medicine, beneficiary engagement, and coordination of care.

Report quality measures to CMS and give timely feedback to providers for continual improvement of care to beneficiaries.

Publicly report certain aspects of their performance and operations (CMS must also publicly report certain quality data).

Recognizing that not all systems are currently structured to accept risk arrangements, CMS allows an ACO to choose one of two program tracks. Both tracks initially use fee-for-service costs as a benchmark for assigned patients. “The benchmark is an estimate of what the total Medicare Fee-For-Service Parts A and B expenditures for ACO beneficiaries would otherwise have been in the absence of the ACO, even if all of those services were not provided by providers in the ACO” [4]. The first track allows an ACO to share 50 percent of the Medicare savings for the duration of their initial agreement. The second track allows ACOs to receive 60 percent of the savings, but they are also at risk for losses which must be paid within 90 days of the close of the accounting period. This second track (for so-called Pioneer Groups) was proposed to attract more established organizations (like Geisinger Health System, Intermountain Healthcare, Cleveland Clinic, Mayo Clinic, and others) that had refused to participate in the first round of proposals [5]. While many groups, such as some Premier Alliance members, did sign up for this second track, many of the targeted medical groups did not [6].

With respect to quality targets and bonuses: “Under both models, if an ACO meets quality standards and achieves savings and also meets or exceeds a Minimum Savings Rate (MSR), the ACO will share in savings, based on the quality score of the ACO ... As with shared savings, the amount of shared losses will be based in part on the ACO’s quality performance score” [7]. Unlike the ACO-specific financial targets, the 33 quality performance benchmarks will be derived from national targets. Payments will be based on a sliding performance schedule, with maximum payment at the 90th percentile of the benchmarks. Initially, the quality bonuses are based on pay-for-reporting, but change over a couple years to pay-for-results.

These quality measures are classified into four domains:

Patient/caregiver experience (7 measures)

Care coordination/patient safety (6 measures)

Preventive health (8 measures)

At-risk populations: diabetes (6 measures), hypertension (1 measure), ischemic vascular disease (2 measures), heart failure (1 measure), and coronary artery disease (2 measures)

CMS requires the ACO to explain how its method of sharing the savings with its providers will further the organizational goals, but does not specify the amounts or formulas for such distributions.

Member Assignments

Medicare beneficiaries do not choose to belong to a Shared Savings Program. Instead, CMS identifies primary care providers (general practitioners, family practitioners, internists, or geriatricians as well as nurse practitioners or physician assistants) who belong to an ACO and then assigns their patients to that ACO. This assignment is based on claims that use primary care billing codes linked to the provider’s NPI. If beneficiaries do not see a primary care provider, then they are assigned to an ACO based on whoever is providing primary care services, such as a cardiologist, oncologist pulmonologist, rheumatologist, etc. “Providers participating in an ACO must notify beneficiaries that they are participating in an ACO, and that the provider is eligible for additional Medicare payments for improving the quality and coordination of care the beneficiary receives while reducing overall costs or may be financially responsible to Medicare for failing to provide efficient, cost-effective care. The beneficiary may then choose to receive services from the provider or seek care from another provider that is not part of the ACO. A provider may not require a beneficiary to obtain services from another provider or supplier in the same ACO, as beneficiaries maintain the freedom to choose which providers they see” [3].

By mid-2012, 154 organizations were taking part in the hared Savings Programs covering 2.4 million Medicare members [8]. At that time, 8 million to 14 million people were receiving care from non-Medicare ACOs, and 15 million non-Medicare patients received care from ACOs that participate in the Shared Savings Programs [9]. Another 106 Medicare ACOs were added in January 2013 [10].

Financial Impact

For the period 2012 to 2015, CMS estimates aggregate median federal savings of $470 million with median aggregate bonus payout of $1.31 billion. ACO startup costs are estimated to range from $29 million to $157 million, with annual operational costs of $63 million to $342 million.

Other Models

In addition to the statutory (that is, mandated by the ACA) Shared Savings Program explained above, the Center for Medicare and Medicaid Innovation (CMMI) has established two other programs. The first is the Advance Payment Model for rural providers and physician-owned organizations for which startup costs are a significant barrier to participation in the Shared Savings Program. CMS will provide advance funding to help them start their ACOs.

The second, non-statutory model is the Pioneer ACO (mentioned above). This scheme was established to attract the well-developed, existing systems that refused to participate in the Shared Savings Programs. Despite the refusal of these large groups, by December 2011, 32 organizations did sign up. Compared to the Shared Savings Program, Pioneer ACOs differ in the following ways:

They must generally be responsible for 15, 000 members (instead of 5,000).

CMS assigns the beneficiaries prospectively, rather than seeing whom they choose for primary care.

While in the first 2 years of the agreement they are in a fee-for-service-based shared savings arrangement, the bonus and risks are higher and based on historical utilization patterns of the assigned beneficiaries.

Starting in the third year of the agreement, they can transition to a population-based, full risk arrangement, which can be extended to years 4 and 5. This method is a per-beneficiary per-month payment rather than a fee-for-service arrangement. Transition to this basis depends on achievement of minimum savings targets during the first 2 years.

They are required to negotiate similar payment arrangements with other payers by the end of the second year. These other arrangements must account for more than 50 percent of ACO revenues.

By the end of 2012, at least 50 percent of the ACO’s primary care physicians must have met meaningful use requirements for electronic health records.

While the explanation above outlines the general descriptions for payments, Pioneer ACOs have many options [11].

All ACOs share the following features:

The same quality performance assessment, quality measures, modes of data collection, and timing of data submission and reporting.

The ability of beneficiaries to opt out.

The ability of beneficiaries to choose providers outside of the ACO network without penalties.

Primary care physicians can only join one ACO.

Barriers to Success of the ACO model

Conceptual Problems from Research Findings

Both CMS and private payers have great expectations for the success of ACOs. However, pilot research findings described in Table 1 call into question the soundness of the ACO concept.

Table 1

Medicare Demonstration Projects Using Value-Based Purchasing [12].

Demonstration	Date	Description	Results
Acute Care Episode Demonstration	2009–ongoing	Bundled payments that cover all physician and hospital services for selected cardiac and orthopedic procedures.	No published results yet.
Gainsharing and Physician Hospital Collaboration Demonstrations	2008–ongoing	Models allowing gainsharing between hospitals and physicians. Incentive payments tie to quality and efficiency improvements.	No published results yet.
Home Health Pay-for-Performance Demonstration	2008–2009	Potential for shared savings for home health agencies that had the highest quality scores or the largest improvement in quality scores.	Preliminary findings from the first year show little or no effect on Medicare spending and on patient outcomes.
Participating Heart Bypass Center Demonstration	1991–1996	Certain facilities were paid a single bundled payment for heart bypass surgery. Centers were chosen based on prior efficiency and high quality.	Savings were about 10 percent of fee for service spending. Patient outcomes were similar to comparison group.
Physician Group Practice Demonstration	2005–2010	Physician group practices eligible for bonuses if they lowered spending on their Medicare patients. Bonuses also dependent on quality metrics.	Little or no effect on Medicare spending. Small improvement in process measures.
Premier Hospital Quality Incentive Demonstration	2003–2009	Hospitals eligible for bonuses if their quality scores exceeded a certain threshold.	No effect on Medicare spending. Small improvements in process measures, no effect on mortality.

Since the Physician Group Practice (PGP) Demonstration is the one most often cited as the design rationale for ACOs, a few comments about it are warranted. At the beginning of the study in April 2005, 10 large, nationwide group practices participated in a Medicare fee-for-service comparison with local control groups. If a demonstration group achieved more than a 2 percent savings for both Medicare Parts A and B compared to the control population, CMS paid a bonus of 80 percent of that savings. However, three features were very different from the model to which ACOs aspire. First, the results were calculated with a case-mix adjustment. The CMS ACO model does not use this adjustment. Second, CMS only conducted a fee-for-service appraisal, not the global risk-based arrangement to which the ACO model eventually aims. Third, quality performance targets were only for process measures, not outcomes.

The significant findings of this demonstration project are: “The four PGPs earning performance payments in the second year either were affiliated with an academic medical center (Dartmouth-Hitchcock and University of Michigan) or were free-standing physician groups (Everett and Marshfield). No performance payments were earned by the five PGPs that are part of integrated delivery systems (systems that include hospital ownership but are not affiliated with academic medical centers) or by the physician network (Middlesex) that is sponsored by a hospital affiliate. The majority of the savings at all sites occurred in outpatient, not inpatient, services. RTI [the firm CMS hired for the analysis of the project’s performance] hypothesized that the presence of a hospital was ‘a potential deterrent to achieving savings ... since these systems may be unable to reduce avoidable admissions or use lower cost care substitutes without affecting their inpatient revenue’” [13]. Note the significantly negative evaluation for systems involving hospitals.

In addition to the above findings, two additional major factors cloud the concept’s validity. First, with the risk-based bonus payment method (higher payments for sicker patients, as mentioned above), CMS incentivized the study groups to enhance data capture and coding to boost the basis for their payments. With these higher payments, the cost savings compared to the control sites appeared more substantial. As a MedPAC study [14] pointed out: “At nine of the 10 PGP sites, patient risk scores grew faster than at the comparison sites … Because the increased risk scores of patients at the PGP sites may be due to improved detection and coding of acute and chronic conditions, the evidence that the PGP demonstration has reduced the costs of care during its first two years is not definitive.” Also, Colla et al. [15] noted that even by the conclusion of the study: “Estimates were sensitive to the risk-adjustment method.”

The second confounding factor in interpretation of the results is that the cost improvement trends for the PGPs as well as control groups began at least a year before the study commenced.

Despite lack of proof of concept for ACOs, the goal of enhanced accountability for cost and quality over the continuum of care is a desirable one. Perhaps speculating that the execution and not the concept has been flawed, CMS proceeded with the ACO projects. To improve the possibility of success as well as elucidate further potential problems, the following discussion provides an analysis of previous issues faced by ODSs as well as newer ones that ACOs may encounter.

Previous Issues

Consolidation (Vertical and Horizontal Integration)

One of the assumptions behind ACO proposals is that better integrated systems will be more efficient and thus lower costs. Past integration attempts, however, have not proven to be successful. As Burns and Pauly [16] point out: “There are no general results from economic theory that vertical integration leads to greater efficiency or market power on the part of the firm.” In this case, vertical integration refers to such activities as hospitals’ purchases of physician practices. As far as horizontal integration (such as hospitals merging with one another) is concerned, direct patient care and other personnel-intensive services may not be scalable. More recent reviews of horizontal integration [17] have found that hospital consolidation generally results in higher prices; hospital mergers in concentrated markets can significantly increase prices, often exceeding 20 percent; hospital competition improves quality of care; and physician-hospital consolidation has not led to either improved quality or reduced costs.

Continuum of Care (Disease Management, Care Coordination, Case Management)

Proponents of integration believe ACOs are well-positioned to improve quality and lower costs for populations with specific chronic conditions and expensive, acute diseases. While this notion is attractive, there is little systematic evidence to support the effectivness of disease management programs. For example, in a large study of such efforts, RAND Corporation researchers found that “although disease management seems to improve quality of care, its effect on cost is uncertain” [18,19]. Further, programs such as those geared to lower readmissions have not consistently proven successful [20]. More importantly, these continuum of care and readmission reduction programs do not require an ACO structure to succeed.

Importance of Primary Care Physicians (Patient-Centered Medical Home [PCMH] as the Core of an ACO)

In order to succeed, ACOs need a core of primary care physicians who will coordinate care. In the case of Medicare patients, internists are necessary. Already in short supply, the situation is getting worse. This problem was highlighted in a study sponsored by the American College of Physicians and the American Board of Medical Specialties that found: “The shortage of primary care physicians may reach 46,000 FTEs by 2025 … greater than shortages predicted for any other specialty area” [21]. This shortage may be exacerbated by the requirement that a primary care provider can be a member of only one ACO. Further, the success of Patient Centered Medical Homes (PCMH), which are the essential core function of an ACO, has yet to be proven. In their systematic literature review, Jackson et al. [22] found that the “PCMH holds promise for improving the experiences of patients and staff and potentially for improving care processes, but current evidence is insufficient to determine effects on clinical and most economic outcomes.”

Start-Up Costs

As mentioned above, startup costs for ACO formation are significant, averaging $1.7 million per organization according to CMS. Funding is limited to the advanced payment model, so only relatively wealthy systems will be able to participate.

Alignment of Incentives

Successful ACOs must reward all decision makers for providing cost-effective, high quality medicine — a shift from so-called volume to value. However, according to physician placement firm Merritt Hawkins [23], 73 percent of search assignments featured a salary with production bonus, while only 35 percent offered bonuses with a quality-based component. Further, hospital administrators are also still largely paid for volume performance [24]. Clearly, hospitals have not made the appropriate shift in incentive payments necessary for ACO success.

Mixed Organizational Models (Ownership, Joint Ventures)

By and large, hospitals have unsuccessfully tried a variety of organizational models in their integration efforts. According to Burns and Pauly [25]: “Previous reviews suggest that hospitals’ efforts to partner with physicians using physician-hospital organizations or the acquisition of physician practices have not promoted cooperation, improve quality, contain costs, or integrated clinical care.” The exceptions seem to be organizations established and run by physicians, such as the Cleveland Clinic, Geisinger Health System, and the Mayo Clinic.

Same Payment Models

While traditional ACO models will initially be paid on a cost-saving basis using fee-for-service data, eventually all will move toward global payments based on capitation. ODSs employed these payment models without much success. ACOs have generally not yet adapted to make them work. One of the unanswered questions is how ACOs will distribute payments among the professional and institutional providers. Another unresolved problem is the system’s ability to collect and analyze the data needed to handle these payment models.

Information Systems

While current information systems are a vast improvement over those of the past 2 decades, at least two major problems remain. First, for ACOs to operate efficiently and effectively, systems across the enterprise must be able to “talk to one another,” an issue called interoperability. Despite the promises of Health Information Exchanges, this requirement had not yet been realized to any significant extent. Second, before information systems can enhance efficiency, the underlying processes that they automate must be evaluated and often reengineered. Too often, this latter process has not been accomplished [26,27].

Culture

Hofstede [28] defined culture “as the collective programming of the mind that distinguishes members of one group or category of people from another.” While mergers are executed only after financial issues are extensively vetted, few combining institutions assess cultural compatability. Perhaps the most famous example of cultural incompatibility causing merger failure was the union of University of California at San Francisco (UCSF) and Stanford University Medical Centers in 1997. The reason for its dissolution in 2000 was succintly stated by Nathan Nayman of the Hospital Council, a trade organization representing northern California hospitals: “Comparing Stanford and UCSF is like comparing apples and oranges. The two hospitals had radically different institutional cultures, which made the merger impossible in the end” [29].

Legal Issues: Antitrust

Antitrust concerns have always threatened integration efforts; however, from 1994 to 2007, the Department of Justice (DOJ) failed to prevent any nonprofit hospital mergers. Richman [30] noted that “confidence that nonprofit hospitals’ market concentration does not lead to higher prices largely drives judicial sympathy for nonprofits in merger cases, and in turn a tolerance of nonprofits’ market power.”

That trend changed on August 7, 2007, when the Federal Trade Commission (FTC) unanimously ruled that Evanston Northwestern Healthcare’s (ENH) 2000 acquisition of Highland Park Hospital violated Section 7 of the Clayton Act by creating a highly concentrated market, thereby increasing hospital prices and harming consumers. The change in attitude toward non-profits is summarized by the FTC’s conclusion that: “ENH’s non-profit status did not affect its efforts to raise prices after the merger, and ... does not suffice to rebut complaint counsel’s evidence of anticompetitive effects” [31]. Even though ENH claimed it spent more than $120 million on integration improvements with the extra charges, the FTC said quality improvements must result from cost-saving efficiencies, not higher prices.

Since then, the FTC has been successful in a number of challenges to horizontal and vertical integration of hospitals and medical groups. As Carlson [32] pointed out: “The collision of old-world antitrust enforcement and new ideas like accountable care, bundled payments, value-based purchasing and patient-centered medical homes has ratcheted up the uncertainty over healthcare needs.” Although the DOJ and FTC have developed joint antitrust guidelines for the ACO Shared Savings Program, the interpretation and application of these rules are still unclear. A further wrinkle on this issue came in 2013 when the U.S. Supreme Court ruled [33,34] that even a government-owned hospital is not exempt from antitrust when it seeks to purchase a local for-profit hospital.

Newer Challenges

Patients Not “Locked In” as They Were with HMOs

In the past, systems assumed financial risk using a model akin to Health Maintenance Organizations (HMOs). Among the features that make HMOs viable are two that apply to ACO success: pre-assignment and provider choice restrictions. HMO members are assigned prospectively to the providers caring for them. The Shared Savings Program described above assigns ACO members retrospectively; thus, physicians and patients do not know where to preferentially direct care as the need occurs. Also, unlike HMO members, ACO enrollees are free to seek care wherever they like without an increase in their out-of-pocket expenses. In short, if health care systems do not know who is their responsibility and cannot control utilization, they cannot be held clinically or (more importantly) financially responsible. These two features are important objections that caused the aforementioned large systems to decline to participate in the program. That this problem is significant is highlighted by the fact that Advocate Health Care (an Illinois partnership between a large multispecialty physician group and hospital system) reported that “nearly half of the ACO population’s hospital use occurred outside Advocate’s 10 hospitals” [35].

Quality Measurement & Management/Evidence-Based Medicine

Previous models rewarded systems mostly for producing financial savings. ACOs will also be evaluated on quality metrics. The good news is that these measures are much better defined than a decade ago. The downside is that many systems have not yet adapted to Medicare’s value-based purchasing, which includes patient satisfaction reports as well as clinical assessments. Further, the standards have not been clearly defined, and some targets are unrealistically high. In early 2013, the latter problem caused all Pioneer ACOs to request that CMS and CMMI delay application of outcome measures for another year [36]. The governemnt declined to do so.

More Insurance Risk

As mentioned above, in the past, ODSs assumed insurance risk by accepting global rates and capitation for specified services. The new arrangements, however, require assumption of even more elements of risk — including non-acute services such as home health and skilled nursing facility care. Past experience has shown that four elements can challenge these systems’ ability to profit by taking insurance risk: low enrollment, inadequate funding, adverse selection (enrollment of sicker patients), and lack of expertise managing insurance products. With respect to low enrollment, Medicare has put a lower membership limit on ACOs; the 5,000 members that Medicare requires should be sufficient to mitigate size-related risk. The initial shared-savings mode of payment provides fee-for-service compensation at Medicare rates, diminishing the payment adequacy problem (at least until capitation is phased-in). However, the latter two risks remain as significant barriers to ACO success.

With regard to adverse selection, unlike Medicare Advantage plans, CMS does not compensate ACOs for caring for sicker patients. What could go wrong is highlighted by the experience of the Greater Marshfield Community Healthplan, a Medicare HMO prototype in the early 1980s [37]. The plan was based around tertiary care center Marshfield Clinic in rural central Wisconsin. Because patients were sicker than at other demonstration sites around the country and capitation was based on rural Wisconsin rates, the plan folded after experiencing significant losses.

The last problem is that insurance risk and management are not core competencies of these hospital-based organizations. Outsourcing will not completely help since the ACO needs to understand and oversee companies that perform these services on their behalf.

More Internal Provider Negotiations on Division of Payments and Shared Risk

As long as payers compensate ACOs by historic fee-for-service rates, the ACOs will continue to pay employed and partnered physicians using a fee schedule (often based on Medicare rates). When the ACO assumes full risk and is only receiving global payments or capitation, the problem of dividing the money will emerge as an important issue for the system to address. Further, since the enterprise is at increased risk for losses, which part of the system will bear, the burden of those shortfalls must be worked out. This latter issue was one of the core difficulties in the above-mentioned UCSF-Stanford breakup. According to Stanford University President Gerhard Casper, “One of the largest problems was administrators’ failure to achieve physician buy-in. Faculty members at both institutions resisted the merger from the beginning, refusing to combine their practices and share financial risk” [38].

Conclusion

American health care is fragmented, duplicative, and wasteful. Because hospitals have been at the center of our acute care system for so long, policy makers believe they are qualified to solve these problems by making them responsible for populations across the continuum of care. While it is not impossible for ACOs to achieve the promise of their ODS forerunners, as the above analysis demonstrates, they must first overcome fundamental problems in both theory and practice. In addition to a shift from the acute care model to one of continuity, these systems must also revise their financial models to reward quality as well as volume and learn to embrace economic risk. In other words, they will need to learn to profit from empty beds as much as filled ones. The major question that remains is: Do hospitals have to reorganize to achieve better performance?