The independent physician practice, though declining, is still a significant medical practice model in the United States, as one physician out of three identifies him or herself as independent practice owner, partner, or associate. In 2016, 38% of physicians in the United States were still working in practices comprising five or fewer physicians. Recent trends show physicians, particularly primary care doctors, moving from smaller to larger group practices, with a simultaneous shift from private practice to employed settings. Medical practice is also evolving with continued technological advances resulting in less face-to-face visit-related work and more time spent completing required documentation.
As of January 1, 2017, half of the 480 organizations participating in the government’s Medicare Shared Savings Program (MSSP) — which offers upside potential and downside protection for the 438 Accountable Care Organizations (ACOs) in what the Centers for Medicare & Medicaid Services (CMS) refers to as Track 1 — reported to CMS that they are composed solely of networks of individual physician or small group practices.
Research suggests that physician leadership plays a key role in the success of ACOs. In 2015, smaller ACOs and those led by physicians were performing better than ACOs led by hospitals and other large medical organizations. In addition, estimated cost savings do not appear to be linked to financial integration with a hospital.
There are known advantages to small health care organizations, including the ability to provide the same quality of care at a lower total cost. Yet because physician-led ACOs are such a heterogeneous group, little is known about the formation of ACOs comprising mostly individual practices.
In this context, participating in physician-led ACOs may be seen by independent physicians as an attractive alternative to consolidation with hospitals or health systems, allowing physicians to provide high-quality care at a fair price while retaining their independence. However, little has been written about how independent physicians start ACOs and how they initially evolve.
This report describes the findings of a study that aims to understand the evolution of ACOs primarily formed and led by networks of independent physician practices and how they assume accountability for patients, including when they receive services from specialists and hospitals outside the ACO. We explore the benefits and challenges facing such ACOs and suggest actions that could support and sustain the independent physician-led ACO model.
A Sample of 13 ACOs Across the Country
From the lists maintained by CMS, we purposefully selected a national sample of Track 1 (i.e., upside risk only) MSSP ACOs that did not include a health system or hospital participant. However, we made exceptions for physician-led ACOs that included a small local hospital. Few ACOs in our sample included specialist practices as organizational members; the majority of physician-led ACOs we studied were composed of primary care practices.
We intentionally sought a mix of early ACO adopters (2012 and 2013) that had achieved shared savings, and late adopters regardless of their results. After being contacted by email, executives from 13 ACOs in eight states (Arizona, California, Delaware, Florida, Kentucky, New York, Pennsylvania, and Texas) agreed to participate in the study. Across the 13 ACOs, we interviewed 40 ACO insiders (executives and staff of participating practices and experts) by phone or in person between November 2016 and May 2017.
The 13 ACOs in our sample were, on average, smaller and better performing than the MSSP ACOs overall. The sections below describe what the interviewees believe are keys to their success, as well as their remaining challenges.
Results: Small Can Be Beautiful
ACO executives from the sample group reported that because of their limited resource availability, they targeted a small number of priorities with a focused approach. These ACOs also tended to implement changes to achieve accountability goals one at a time, building from one action to the next in a progressive sequence. As one CEO explained, they “were not trying to boil the ocean all at once and do everything.”
Due to their small size, physician-led ACOs are nimble and able to experiment. The ACOs in our sample tested specific interventions with a portion of their members before implementing them across all members. One executive noted that his ACO had “the ability to try a program — whether it’s with care coordination, disease management — completely mess it up, figure out that it’s the wrong direction, and try something else.”
The ACO was typically small to start (four or five people) and had experienced leaders, often with backgrounds in managed care. Most ACOs had regular contact with the practices that comprised it, using different modes of communication. Communications targeted physicians as well as other staff members, including the practice manager, regarding topics such as reporting or organizing care management activities. Through regular communication, the leaders of physician-led ACOs sought to create alignment between the ACO and the practices. One ACO executive said they were “just trying to align all of communications to reinforce when a committee is made, when sending out notifications or something, constantly trying to reinforce the same information and send it out.”
Despite the shared focus on primary care among ACOs, Patient-Centered Medical Homes (PCMH) and Comprehensive Primary Care Plus (CPC+), which are CMS programs, the study participants did not describe a clear pattern of progression from PCMH or PCP+ to the formation of an ACO or vice versa. Half of the study ACOs participated in Medicare Advantage, which addresses the same population as the MSSP, but respondents reported no real synergy between the two programs despite the programs’ similar objectives. One ACO executive explained that the Medicare Advantage plans were “pretty much managed by the insurance company,” and that they “have really not been able to manage the care in the way they wanted to.”
A More Systematic Approach to Care Management
The government’s MSSP was the first ACO contract for half of the ACOs in the sample. Commercial ACO contracts often followed physician practices’ agreement to participate in the federal MSSP, as quality metrics were similar (though not the same), and thus the incremental burden was low. Monthly care management fees paid by the commercial contracts were also used to finance ACOs’ care management teams.
Due to the financial incentives from Medicare, the inception of the ACOs in our sample happened concurrently with the development of a systematic approach to care management. All utilized the Medicare incentive for annual risk assessment visits, or wellness visits, while most used the chronic care management services fee provided by Medicare by developing a chronic care management program, consisting primarily of calls to patients conducted between office visits. Some also used a transitional care management services fee by implementing a transitional care management program to support patients after being discharged from the hospital, thereby providing evidence that the ACOs were informed of hospital stays. These programs, which were billed on a fee-for-service basis, provided a means for collecting data needed for ACO quality reporting, while also increasing physicians’ Medicare revenue.
Using data and reports provided by CMS on the total expenses of the patients attributed to the ACO, the ACO insiders reported that they were able to target patients with higher expenses to avoid hospital admissions, readmissions, and emergency room visits. To support patients between office visits, some physician practices changed their office hours or made same-day (within 24 hours) appointments available. Physician practices also instituted team-based care, so that nurses and medical assistants, as well as physicians, had direct contact with patients.
While all ACOs in the sample had some sort of care management program, the scope and services provided varied. Some ACOs chose to offer care management only through the physician practice, others through the ACO. Having a care management team within the ACO did not necessarily mean that patients from all physician practices in the ACO had access to care management. Rather, this depended on the physician practices’ size and human resources. Other actions within ACOs varied as well, including, for example, processes for scheduling follow-up appointments. In most cases, a care coordinator employed by the ACO was assigned to the physician practice and considered a staff member and had access to the practice’s electronic health record (EHR). One physician said that the ACO care coordinator was “that extra pair of hands helping to make sure the patient is taken care of.”
Small May Also Be Limiting
Access to capital was the most acute issue raised by ACO executives we interviewed. Each ACO found its own way to support operations: loans, physician contributions, venture capital, founders, etc. For ACOs not funded by larger companies, limited ability to make up-front investments impacted the development of technical operations, such as data analytics and health information technology (HIT). In these ACOs, the infrastructure for data analytics could be very basic in the initial years, so they relied almost entirely on quarterly reports from CMS. None of the ACOs included in our study had a unified EHR encompassing all its members, and interoperability within the ACO was reportedly a common obstacle that had not yet been overcome.
ACOs in our study partnered with a variety of outside entities to offer the array of services needed for patient care. However, the type of partners was variable and depended on the local market. ACOs often established partnerships with analytic or information technology vendors and occasionally with ancillary or post-acute service providers. These partnerships were not always successful, and one ACO in the study withdrew from the ACO program because it had failed to thrive.
The size of the ACOs was a compromise between the need for a diverse set of skills — clinical, IT, administrative, and financial — and the limited capacity of the ACOs to invest. As several executives expressed, their organizations were “very lean.” This proved particularly challenging when the ACO needed to adjust to a growing number of members and Medicare beneficiaries.
Little Clinical Integration Outside of the ACO
ACO insiders we interviewed described their ACOs’ ability to manage care outside the organization as dependent on the interoperability of HIT systems and the quality of the regional health information exchange (HIE) with local hospitals. HIE enabled automated alerts for physicians when patients were discharged from the hospital in five of the eight states included in our sample. In areas that lacked HIE, ACO insiders claimed that getting information when a patient was discharged was a difficult obstacle to overcome, as small organizations have less clout in dealing with hospitals.
In most cases, there was no formal agreement on discharge procedures with hospitals at the ACO level but, in general, hospitals were willing to work to improve hospital discharges because of the penalty for readmissions. Executives from four ACOs we interviewed were in the process of developing a hospital liaison function, making it possible for someone from the ACO to go into the hospital when their patients were admitted.
None of the ACOs had yet formed structured specialty referral networks, including most of the few ACOs in the sample that had specialist practices as member organizations. As a result, the management of specialty care referrals was left to the discretion of individual physicians. Some used the EHR to track referrals, and two had tools for referral management, but these were not utilized by the specialty practices, and they had no arrangement governing their use. Four ACOs identified establishing specialty referral networks as a top priority, explaining that this has important implications on the ACO’s ability to provide comprehensive patient care and care management. The remainder of the ACOs, however, had not yet reached this stage of development due to time constraints and competing priorities as described above.
While some ACOs had partnerships with post-acute facilities, and close to half were working to engage leaders of post-acute facilities, leverage among our physician-led ACO sample was limited because primary care physicians usually do not refer patients to post-acute care.
Clinical and functional integration (e.g., information exchange) with specialists and hospitals were not among the initial priorities for the first years of an ACOs’ existence, and, as a result, were generally poor.
Unresolved Issues of the U.S. Health Care System
Interviewees often complained about the disproportionate burden of coding and billing for the services provided to patients. The change from ICD-9 (International Classification of Diseases) to ICD-10 was mentioned several times as a challenging experience for the physician practices, as well as individual physicians. Interviewees reported that health care regulations and changes to regulations were difficult for physicians to keep up with. Documentation was considered very time-consuming, and EHRs were not seen as helping physicians in their daily work.
Some of the interviewees regretted that there was no accountability among patients and found it difficult to be responsible for the overall care of patients when patients were free to see any provider of their choosing.
More than half of the ACO representatives we interviewed cited challenges around benchmarks they perceived were constantly moving both during the year and from one year to the next. Executives from five ACOs with earned savings expressed some worry about the CMS “benchmark” making it increasingly difficult for them to perform well over the years. (For each MSSP ACO, CMS calculates a benchmark using risk-adjusted average per capita expenditures for Part A and B services for original Medicare fee-for-service beneficiaries, and the benchmark is recalculated every year.) The impact of risk adjustment, which is used to calculate beneficiary risk scores and to adjust the benchmark, could not be anticipated by any of the ACOs of our sample and was therefore seen as a “black box.”
Half of the practices in the sample described the lag time between performance and evaluation as a challenge, explaining that until reconciliation, which happens each September following performance the previous calendar year, the ACO does not know if it will achieve any savings. One executive called MSSP “a delayed gratification program.”
The perspective of downside risk programs was worrying for half of the ACOs. When renewing their agreements in 2017, only one of 10 ACOs chose a downside risk model.
MSSP Track 1 Model Fits with Values of Independent Physicians
The opportunity to receive value-based payments and the prospect of generating savings without assuming risk were key to engaging independent physicians in the MSSP. Physician leaders preferred to keep their financial involvement to a minimum, with no obligation to invest. Conversely, there was also no obligation among physician practices within the ACO to adopt one single EHR, and this was not on the agenda of any of the ACOs in this study.
The ACOs could choose their own form and pace of implementation. The small size of the ACOs seems to foster close relationships with the physician practices that comprise the ACO. Two-sided communication helped the ACOs to customize support to meet the specific needs of their practices and the local environment. As one CEO said, “doctors don’t really respond to big data, they respond to very local data.”
The ACOs’ role in developing care management seemed essential in the redesign of primary care practices: Care management fees were not limited to the ACO programs, but it seems that participation in the program resulted in practices utilizing care management.
The inception of the ACO seems more straightforward when practices have had some experience working together and benefit from the support of organizations such as an Independent Practice Association or a Management Services Organization.
Large-scale investment in sophisticated information systems and data analytics capabilities were not considered necessary to achieve savings in the ACOs’ first years. Limited internal resources imposed a greater need for partnership to overcome the lack of access to capital, analytics capabilities, and HIT interoperability. Resource constraints were also a driver for local partnership to increase impact outside of the ACO.
There remains room for improvement in clinical integration across the continuum of care, as impact outside ACOs has been moderate to date. However, it is unclear whether the MSSP ACO model alone could fill this gap, as it may be only transitional.
Conclusion: Success Keys for ACO Leaders
Lessons learned through the early experience of successful MSSP ACOs may be useful for other ACO leaders. As they begin their ACO journey, the following are practical areas that leaders can control and manage irrespective of regulations and policy:
- For the first years of the ACO, building the ACO leadership and staff should be the main objective, as there is no immediate need for massive investment in infrastructure or IT.
- Establishing close relationships between practices and ACO leaders allows the ACO to develop programs to address problems and needs of participating practices. This is particularly important at the beginning to build trust and engagement.
- ACOs should not attempt to do everything at the same time, but rather to identify with the ACO member practices a small number of priority issues and implement them. A step-by-step approach, starting with specifying roles and responsibilities at the primary care level, seems particularly suitable for small organizations, at both the level of practice and of the ACO as an organization. Putting basic organizational elements in place then enables ACOs to fulfill care management functions like annual awareness visits and chronic care and transitional care services.
Conclusion: Next Steps for Policymakers
The information collected through our interviews also suggests that there are a series of challenges policymakers can work to address in support of the adoption and successful implementation of physician-led ACOs. Adapting the MSSP to facilitate participation and involvement, particularly among small, independently run practices comprising physicians who do not want to consolidate, could contribute toward improving care quality at a lower cost. The following are suggestions from the authors:
- Maintaining financial reserves to protect against downside risk presents a major challenge to small, physician-led ACOs. To support the existence of these organizations, some requirements, such as the reserve fund (which requires ACOs that assume double-sided risk to reserve at least 1% of CMS fee-for-service expenditures as a repayment mechanism), could be adapted according to the size of the ACO.
- The number and complexity of rules governing the health care system is also a substantial barrier. CMS could reduce burden on providers by simplifying its regulations, such as the billing codes, and by making less frequent changes to existing rules and codes. Both stabilizing the rules and regulations and making them easier to understand would help sustain provider participation in the program.
- Different rules for different Medicare programs (Medicare Advantage, traditional fee-for-service, and ACO) are costly to follow and may limit the engagement of small organizations across programs. Synchronizing the rules of these programs to the extent possible could promote participation among smaller organizations and would benefit all.
- Absence of HIT interoperability persists, yet this is a requirement for providing comprehensive patient care. If a primary care physician can’t exchange information with a specialist or hospital, properly managing referral and discharges becomes very challenging. There is no indication that market forces alone will result in such interoperability, making this an area where CMS could exert a positive influence.
- There is no explicit roadmap for ACO implementation. The process is incremental and happens over time, with each lesson learned used to refine the next steps. Thus far, ACO leaders have mostly been learning on their own, but common patterns can be observed, and newcomers would benefit from resources that share best practices and implementation tools. CMS could undoubtedly contribute to the development of such resources.
Such programs should be reinforced to support and encourage independent physicians’ participation. The whole health care system could benefit from maintaining organizations led by independent physicians.