Editor’s Note: This is the fourth post in a Health Affairs Blog series on Medicare physician compensation and the Sustainable Growth Rate mechanism. Paul Ginsburg, Robert Berenson, and Mina Matin have contributed earlier posts, and in the coming days the series will feature posts by Frank Opelka, Eugene Rich, and Gail Wilensky.

Medicare expenditures for physician services continue to be impacted by the annual growth in volume of services per beneficiary. Growth in volume per beneficiary between 2000 and 2005 averaged more than 5 percent per year, with increases in imaging and certain procedures more than double that rate. Volume increases may be supply driven and/or driven by changes in physician practice patterns caused by perceived inadequacies of Medicare payment rates. This is especially true in specialties and practice sites where Medicare billing can be increased through the use of imaging procedures and other diagnostic and therapeutic technologies in the office setting.

However, the paradox for physicians who can do this is that such practice patterns increase the pressure on the Centers for Medicare and Medicaid Services (CMS) to reduce physician fees. Thus fee-for-service physicians are stuck in a vicious cycle caused by the inadequacies of the Medicare physician payment system. In addition, since not all physician specialties and practice sites can use technology to increase the volume of Medicare billing, the Medicare physician payment system, and particularly the payment update “Sustainable Growth Rate” (SGR) formula is creating increasing disparity in income among specialties. As described in the earlier post in this series by Mina Matin, Tom Bodenheimer, and Kevin Grumbach, the hardest hit are the primary care specialties.

The physician update (SGR) formula has basic problems, as outlined in Paul Ginsburg’s post. Essentially, it is a flawed volume control mechanism because all physicians in the country are part of the same volume control pool. Since the formula treats all physicians in all regions of the country alike regardless of their practice patterns, there is no incentive for individual physicians to moderate their actions.  

Finding alternatives to the current failed approach to managing the volume of physician services is of increasing interest to a broad and growing group of stakeholders. An effective “fix” must contain a mechanism to create incentives for the appropriate use of services for beneficiaries. 

The Opportunity 

A total solution to the problem of inappropriate increases in the volume of physician services remains to be developed. However, there is a current opportunity to investigate whether a subset of the national physician community, properly equipped and with proper incentives, could manage both the volume of services and the quality of services delivered to Medicare beneficiaries over time, within a fee-for-service setting. Several elements combine to create this opportunity. 

The first of these elements is the growing presence of advanced practice management techniques such as evidence-based medicine, practice protocols, and other systematic care processes in multi-specialty group practices. These organizations include well known groups such as the Permanente Medical Groups, Mayo Clinic, Geisinger Clinic, Virginia Mason Clinic, and others, which have adopted a range of techniques that serve to bring up-to-date medical science systematically to the practice of medicine. Such groups and such activities are not rare. In fact, there are over 600 multi-specialty group practices with more than 50 physicians in the United States. 

Second, recent improvements in hardware and software have created a practical, reliable electronic medical record (EMR). In fact, many of these same multi-specialty group practices are now involved in the deployment of these systems. The modern EMR provides a ready vehicle for the introduction of evidence-based medicine and systematic care processes at the point of care. It is also capable of producing vastly more and vastly cheaper clinical information about care delivery, for analysis and quality improvement. Thus, multi-specialty group practices have or are soon to have a new and powerful tool to both improve and monitor physician practice patterns. 

Third, recent work by John Wennberg and others, who analyzed the practice patterns of group practices in the United States, has suggested a methodology that can use Medicare claims data to link Medicare fee-for-service beneficiaries to group practices rendering care to these beneficiaries.  Since many multi-specialty group practices have a loyal and consistent patient base, it is possible, using Medicare claims data, to construct a “virtual capitated population” for such group practices. Such an analysis can be further refined by adjusting for care delivered to beneficiaries outside of the group practice environment, whether by physician referral or patient choice.  Such algorithms are already commonly employed by large integrated organizations to allocate pre-payment to care delivery divisions within the organization.  

As described above, one reason for the failure of the SGR formula is the size of the physician pool to which the formula is applied. In such a situation, there is a lack of incentive for physicians to employ volume control mechanisms, including the use of evidence-based medicine and systematic care processes. The existence of multi-specialty group practices utilizing practice management tools (i.e., evidence based medicine, systematic care processes, and EMRs) combined with the potential to link Medicare beneficiaries to these practices creates the possibility to provide volume management incentives via the annual Medicare physician payment update system to smaller physician pools, ones equipped with the tools to appropriately manage both volume and quality. 

The Proposal

The following elements could be combined as a proposal to CMS for a pilot project. 

1. Create one or more additional payment update “pools.” CMS would authorize, through the authority under the Medicare Modernization Act (MMA) section 646 or other authority, and on a budget neutral or “shared savings” basis, the creation of one or more physician payment update pools (PPUPs) for the purpose of calculating separate annual updates. The additional PPUPs initially would comprise physicians in multi-specialty group practices who care for Medicare beneficiaries on a fee-for-service basis. Medical groups could apply for and be accepted and maintained in a separate pool based upon meeting established criteria. Such criteria could include, as a minimum, demonstrated ability to use evidence-based medicine and other systematic processes of care; electronic medical information capabilities; the use of systematic quality of care improvement techniques; responsible physician compensation practices; and the willingness to be part of a collective, transparent monitoring and improvement process. The infrastructure for such work could be provided by organizations such as the American Medical Group Association, National Committee for Quality Assurance, or other organizations. 

2. Create a separate update calculation. CMS would create a separate update calculation for the PPUPs based on the development of “virtually enrolled” patient populations, as described above, for participating group practices.  These populations, when combined, would create a pool of Medicare beneficiaries for whom the rate of increase of the volume of services in any give year could be assessed. As described above, the Medicare claims database would be used to create adjustments for care delivered to Medicare beneficiaries outside of the participating group practice environment. 

Based upon the difference between the annual volume of services provided to beneficiaries in the practices within the PPUPs and the concurrent national physician annual volume, the participating practices could receive a separate annual payment update. Computed Medicare savings could be fully “passed through” to the groups or savings could be shared with the Medicare program through a predetermined formula. The hypothesis is, of course, that the smaller physician pool(s), combined with the assets of organized practices, would serve to strengthen incentives within participating groups to be good resource stewards and avoid services of uncertain value, leading to relatively higher payment updates over time for those groups.  


It is likely that such a pilot project as described above could produce several benefits for the Medicare program, including: 

  • Reduction in the provision of services of uncertain value to Medicare beneficiaries served by the participating medical groups. 
  • A demonstration of how the volume of professional services could be managed within the traditional fee-for-service Medicare program through the application of incentives to smaller pools of participating physicians. 
  • A Medicare payment environment in which appropriate volume related behavior is positively rather than negatively reinforced, encouraging physicians over time to aggregate into more efficient and effective practice organizations. 

Significant Issues 

There are a number of significant issues and questions that arise when considering such a proposal. These will need further analysis and contributions by others if the proposal is to evolve further.  Among these issues and questions are the following: 

Can the “virtual enrolled patient populations” be created easily and accurately enough to of be of practical use? This is primarily a technical question which could be addressed initially by John Wennberg and Elliot Fisher of Dartmouth’s Institute for Health Policy and Clinical Practice, and others together with CMS. As noted above, many integrated care organizations currently use similar algorithms for such calculations. 

Would this demonstration have a significant adverse impact on the update for physician services for physicians not participating in the PPUPs? The public policy value of the pilot project, properly constructed, should outweigh such concerns. The pilot offers the possibility over time for physicians, appropriately organized and supplied with proper financial incentives to break the current despised “SGR cycle” of poor updates and annual pressure to increase service volume to maintain income. However, it is likely that the pilot project would elicit objections from some parties who are legitimately concerned about the general adequacy of projected physician payment updates in the near term. 

Would the pilot project create inappropriate and adverse incentives for physicians participating in the project that could result in the withholding of necessary services for Medicare beneficiaries? This is very unlikely. Medical groups would still be receiving fee-for-service payments from CMS based upon services provided, and therefore they would continue to have a financial incentive to provide services. Conversely, the group will have a strong incentive to manage the volume of services appropriately in order to remain within the new update pool. (A mechanism will need to evolve to remove poorly performing groups from the pool.) Managed appropriately, the “balanced” incentives inherent in the pilot project should over time improve both the appropriateness of services and the quality of services rendered. 

Should the pilot project be applied to hospital services as well as physician professional services? This has been suggested by some. It would add a level of complexity to the pilot project because it would require melding of Medicare Part A and Part B payment streams. However, there is a reason to consider such a design, to accommodate the medical groups that own and operate hospitals as part of a single not-for-profit economic entity. To not do so would effectively eliminate those groups from the pilot project because their hospitals would remain under a volume-driven Medicare revenue environment, creating conflicting financial incentives within the system. In addition, the development of such a design over time would serve to minimize conflict between non-related hospitals and physician groups that might wish to participate in such a pilot. 

Should there be one national separate target pool or multiple ones? In its March 2007 special report on the SGR, MedPAC discusses a variety of target payment pool options. There could be one national pool of participating groups, a set of regional pools, or “pools” as small as one institution. Essentially, the larger the alternative payment pool, the less would be the administrative complexity for CMS, but the weaker the incentive would be for any one group to manage volume.  (Again, there would likely evolve a mechanism for excluding malperforming groups from such a pool).  The smaller the pool size, the more complex the work for CMS would become, but the stronger the incentives and accountability for the participating groups would be.

Could this model be expanded beyond existing multi-specialty group practices? This issue was discussed in depth at MedPAC early in 2007 and led to the “Hospital Medical Staff” alternative. This option builds on a proposal by Fisher and colleagues that, if provided with appropriate financial incentives, hospitals and their medical staffs could, over time, come together into more “accountable” delivery systems, capable of qualifying for inclusion in an alternative Medicare payment system such as described above. There are several cultural, governance, and regulatory obstacles to this option, but it would provide for much greater “reach” of this form of Medicare payment reform. In addition, expanding this incentive model would, over time, likely foster closer physician-hospital “clinical integration”, as envisioned by the American Hospital Association (AHA), with resulting improvement of care coordination.