Editor’s Note: This is the seventh and last post in a Health Affairs Blog series on Medicare physician compensation and the Sustainable Growth Rate mechanism. Paul Ginsburg, Robert Berenson, Mina Matin, Jay Crosson, Frank Opelka, and Eugene Rich have contributed earlier posts.

One of the advantages of coming last in a series of blog postings is that I can comment on some of the ideas that have been raised in the previous postings as well as share some of my own thoughts on the choices facing the Centers for Medicare and Medicaid Services (CMS) and the Congress. One thing is clear. Policy analysts, members of Congress, congressional staff, and much of the physician community are unhappy with the current reimbursement strategy, which combines the use of the Resource-Based Relative Value Scale (RBRVS) with a Sustainable Growth Rate (SGR). The RBRVS sets relative prices, but it is the SGR, through its impact on the conversion factor, that converts relative weights to absolute dollars that sets the absolute reimbursement rates. 

Most of the focus has been on the SGR — what’s wrong with it, how to fix it, or whether to do away with it all together. By linking the growth in Medicare spending for physician services to the growth in the economy, and then tying changes in physician fees to any excess growth in physician spending that exceeds the growth in the economy, Medicare has enforced a much more rigid relationship between growth in physician spending, mostly fueled by changes in the volume and intensity of services provided by physicians, and their subsequent fees under the RBRVS. At least one important area of concern that has already been mentioned is the dubious wisdom of holding one sector of Medicare spending — physician services — to this very tight growth rate while other sectors of Medicare are not required to maintain such a relationship. Thus, the recommendation that has surfaced on occasion is that if expenditure targets are to be used for physician services, it would be better to use expenditure targets across all Medicare services.

What that does, however, is freeze the relative shares of Medicare spending that exists at a moment in time, with little or no assurance that the relative shares represent the best distribution of spending in Medicare. Expenditure targets for all of Medicare could force spending on Medicare to stay within specified growth rates — if the targets were actually enforced — but expenditure targets per se do nothing with regard to improving quality, clinical appropriateness, or any of the other goals that have been set for Medicare. 

Initially the SGR looked attractive to many physicians since physicians received larger increases in fees than would have otherwise occurred. That was because economic growth was still robust in the late 1990s and the growth in volume and intensity of services remained relatively modest. It was hardly surprising, however, that this positive effect of the SGR would not remain indefinitely, since either a slowdown in the economy or an increase in the volume and intensity of services would negatively affect the allowable rate of fee increases. With both occurring throughout much of this decade, the SGR has produced successive years of predicted fee reductions, most of which have been overridden by the Congress — the reduction in 2002 being the exception. 

The primary problem with the SGR is that while it can control total spending by physicians (assuming it is actually implemented), it does not affect nor is it driven by the volume and intensity of spending of individual physicians. In fact, there is some concern that expenditure targets may actually exacerbate the incentives for individual physicians to increase the volume and intensity of services they provide. The reason is that nothing that they do as individuals is likely to affect to affect the overall spending level for physician services.

As Paul Ginsburg noted, at least some advisers to the Congress recognized the lack of a relationship between what individual physicians do and physician spending in the aggregate early on, but seemed to believe that somehow specialty societies would be able to pressure the behavior of its individual members. While it is difficult for me to believe anyone seriously could have thought that specialty societies would be able to have that kind of influence over their individual members, the last 15 years has more than demonstrated the falsity of such a presumption. 

Most of the previous postings on this issue have advocated alternative ways to define the relevant groups and/or services to which the SGR should apply. Some have been rather simple modifications such as using two separate SGRs: one for evaluation and management (E&M) services, primarily the office visits used by primary care physicians, and the second for all other services; or separating out multispecialty group practices; or differentiating according to the value of services provided to beneficiaries or Medicare; or the use of six expenditure targets as defined in the Children’s Health and Medicare Protection Act (CHAMP), which not only distinguished between various types of service categories but also allowed the spending target for primary care and preventive services to be substantially greater than the spending target increases for the other categories of service. 

What Next? 

In looking forward, the Congress urgently needs to decide on the basic direction of a future reimbursement system for physicians. Then it can consider what it can or should do in the short term while Medicare gets ready for the next generation of physician payments. I believe that developing a more aggregative payment strategy for physicians is the key to resolving both the frustrations and the perverse incentives associated with the current combination of a disaggregated RBRVS with a spending target. 

As long as Medicare uses a disaggregated fee schedule like the current one, with its more than 6,000 codes for reimbursement, the use of an expenditure target or targets is inevitable. To do otherwise is to invite rapid rates of spending increase, as was seen during the various previous periods when Medicare used disaggregated fee schedules without spending limits. This occurred during the first 20 years of Medicare, when rapid increases occurred in both charges and volumes of services, and also during Medicare’s second period, from 1984 to 1991, when Medicare used the Medicare Economic Index to control fee increases but without a spending limit. 

While it is true, as Bob Berenson has argued, that payment strategies using different approaches to a disaggregated fee-for-service system are not ready for current adoption, the decision about the direction of the next evolution of payment strategies needs to occur now (or at least as soon as possible after the election) so that work can be begin on its adoption. Otherwise, we are likely to be in the same position another decade from now as we are now, bemoaning the perverse incentives of the current system without an ability to substitute a better one in its place. 

As we all know too well, moderating spending is not the only goal for Medicare, important as that is, not only for fiscal sustainability of the program but also for the seniors who fund 25 percent of the Part B premium and also face a 20 percent copayment for all Part B services. Developing a program that encourages the provision of high-quality, efficiently produced care and rewards the clinicians that can produce such care also remain important goals of the Medicare program and need to be reflected in the physician reimbursement system.

The current disaggregated payment strategy used to reimburse physicians is fundamentally incompatible with achieving these combined goals. Aggregating across chronic care treatments so that physicians can receive a bundled payment for single or multiple chronic diseases; combining Part A and B payments for high-cost, high-volume acute procedures, particularly those that show substantial variation in expenditures such as for coronary artery bypass surgery; and use of partial or full capitation payments are all ideas that have been raised periodically in the past and that represent a far superior reimbursement system to the RBRVS combined with an SGR. While replacing the current SGR with different conversion factors for specialty-specific targets and service category targets, such as occurs in the CHAMP bill, makes the spending targets less pernicious, it still keeps payments at the micro level and inhibits the development of more accountable units. 

Ultimately, we need to develop ways to have physicians more associated with the full spectrum of professionals that provide care for an individual during an episode of illness or for the treatment of chronic care as well as with the institutions that provide care during acute interventions. But even if that goal remains substantially in our future, figuring out how to bundle payments for physicians need not be. It will however, take several years to get there. As soon as there is some agreement on the basic design of a future reimbursement system for physicians, it may be desirable for the Congress to set a date when the new system will be implemented. After all, the RBRVS was passed into law in November 1989 with an implementation date of 1 January 1992, even though the RBRVS was still in the process of being developed at the time the statute was enacted. 

In the interim, parsing the SGR into smaller components has some advantages compared to the current single target, but it is worth the effort only if it can be done quickly. Even so, it will not provide a mechanism for rewarding or penalizing physicians for behavior that is under their control. For those of us who believe having the right incentives in place is crucial to building a better Medicare program, multiple SGRs with a disaggregated RBRVS just aren’t going to get the job done.