Statement of Glenn M. Hackbarth, J.D., Chairman,
Medicare Payment Advisory Commission

Testimony Before the Subcommittee on Health
of the House Committee on Ways and Means

Hearing on Physician Payments

February 28, 2002

The current formula for updating physician payments, known as the sustainable growth rate system (SGR), should be repealed. We made this recommendation last year in our March report to Congress because the conceptual basis of the system is flawed and in June we warned of potentially large negative updates for 2002 and in the future. That future has arrived, and CMS now projects four years of negative updates. The basic problem is that in seeking to control spending, the SGR causes large swings in updates from year to year that are unrelated to changes in the cost of providing physician services. Although input price increases for physician services have been in the 2-3 percent range for the last few years, the SGR has produced payment updates of +5.4, +4.5, and -5.4 percent over the 2000 - 2002 period.

We recommend treating payment updates for physician services as we do payment updates for other services. Accordingly, we recommend that the Congress repeal the SGR and instead require that the Secretary update payments for physician services based on the estimated change in input prices for the coming year, less an adjustment for growth in multifactor productivity. We also recommend that the Secretary revise the productivity adjustment currently used for physician services to make it a multifactor instead of a labor-only adjustment. Taking into account current estimates for input prices and productivity, we thus recommend that the Congress update payments for physician services by 2.5 percent for 2003.

Current estimates of updates using the SGR formula show three more years of negative updates for a total decrease of about 17 percent over the 2002-2005 period, a situation that is unsustainable (Figure 1). In addition, updates under the SGR will remain below estimated increases in the cost of providing physician services thereafter. Because the SGR is the current law, our recommendation--or any other action that corrects this problem--will show major budgetary costs. Nonetheless, maintaining access for Medicare beneficiaries and keeping physicians participating in the program and accepting new patients, will require that action be taken.

The problem with the current update system

Setting prices correctly in Medicare's payment systems is essential to maintain access to services for Medicare beneficiaries. The underlying problem with the current SGR system is that it attempts both to set individual prices accurately and to control total spending on physician services delivered to Medicare beneficiaries. These two goals can seldom be achieved simultaneously. The SGR attempted to achieve both goals and failed, as did the Volume Performance Standard system before it.

The SGR system causes payments to diverge from costs because, although the system accounts for inflation in input prices, productivity growth, and other factors affecting costs, it overrides these factors to achieve an expenditure target based on growth in real gross domestic product (GDP) per capita. If actual spending for physician services differs from the expenditure target, updates under the SGR system will diverge from costs. When this occurs, payments will be either too low, potentially jeopardizing beneficiaries' access to care, or too high, making spending higher than necessary. This is a particular concern given that the SGR system only applies to services paid for under the physician fee schedule. Services provided in physicians' offices are paid for entirely under the fee schedule, whereas when services are provided in other settings such as hospitals or ambulatory surgical centers, part of the payment is outside the fee schedule. Updates based on an expenditure target that fully applies to only one setting could create financial incentives that inappropriately influence clinical decisions about where services are provided.

An expenditure target approach, such as the SGR, assumes that increasing updates if overall volume is controlled, and decreasing updates if overall volume is not controlled, provides physicians a collective incentive to control the volume of services. However, this assumption is incorrect because people do not respond to collective incentives but individual incentives. An individual physician reducing volume does not realize a proportional increase in payments. Instead, the increase in payments is distributed among all physicians providing services to Medicare beneficiaries. If anything, in the short run an individual physician has an incentive to increase volume under such a system and the sum of those individual incentives will result in an increase in volume overall. In fact, CMS makes exactly that assumption when it estimates the so-called behavioral response of physicians to lower payments--which is an increase in volume of services provided.

Over a longer period, if payments were clearly less than physicians' marginal cost of providing a service, we might see physicians cut back their Medicare practice and concentrate on other patients, devote more time to other professional or leisure activities, or leave practice altogether. Ultimately, we could see fewer applicants to medical school or a shift in residency preferences away from those specialties most heavily dependent on Medicare. The result eventually would be decreased access for Medicare beneficiaries which would be very difficult to reverse.

Compounding the problem with the conceptual basis of the system, the SGR system produces volatile updates. Updates went from large increases in 2000 and 2001 of 5.4 percent and 4.5 percent, respectively, to an unexpected large reduction in 2002 of 5.4 percent. The recent volatility illustrates the problem of trying to control spending with an update formula. To control spending the update formula compares actual spending to an expenditure target. That target changed abruptly last year because of two corrections. First, the Department of Commerce re-estimated historical GDP, which made prior physician spending growth too high by lowering the historical spending targets. Second, both actual and projected GDP went down since last spring, bringing down estimates of allowed future spending growth. In addition to corrections in the expenditure target, CMS found it had not counted some physician spending, correcting for that error increased actual expenditures and thus further increased the difference between actual and target expenditures. As a result of these corrections, CMS's estimates for the update in 2002 changed from -0.1 last March to -5.4 percent in November.

To address such problems, in our March 2002 report we recommend that the Congress replace the SGR system with an annual update based on factors influencing the unit costs of efficiently providing physician services (MedPAC 2002). The Commission's recommendation is based on a belief that getting the price right is important when making update decisions. If total spending for physician services needs to be controlled, it may be better to look outside the payment update mechanism, achieving appropriate use of services through outcomes and effectiveness research for example, as we suggested in our March 2001 report to the Congress (MedPAC 2001). If controlling total Medicare spending is the goal, then an approach that targets all of Medicare spending, not just physician spending, would be more appropriate. Below we describe how the Congress should replace the SGR system.

What should be done?

Replacing the SGR system would solve the fundamental problems of the current system and would allow updates to account more fully for factors affecting costs. The change also would uncouple payment updates from spending control and make updates for physician services similar to the updates for other services. This change would promote the goal of achieving consistent payment policies across ambulatory care settings, including physician offices, hospital outpatient departments, and ambulatory surgical centers. Accordingly, the Commission recommends that:

The Congress should repeal the sustainable growth rate system and instead require that the Secretary update payments for physician services based on the estimated change in input prices for the coming year, less an adjustment for growth in multifactor productivity.

To replace the SGR system, the Congress could repeal provisions in current law and replace them with language similar to that for other services. For example, the Social Security Act requires updates for inpatient hospital care to equal the increase in the hospital market basket index, except when the Congress chooses to make the update smaller or larger. The Congress generally makes these choices after considering advice from MedPAC and the Secretary. With a similar update method for physician services, the Commission intends to base its advice to the Congress on assessments of payment adequacy such as the one discussed later in this testimony.

Payment updates should take into account productivity improvements that enable physicians to provide care more efficiently. Revising the productivity adjustment to account for labor and nonlabor factors is consistent with the way physician services are produced. Labor accounts for most of the cost of providing physician services, but capital inputs are also important, including office space, medical materials and supplies, and equipment. The production of physician services, like the production of most other goods and services, is a joint effort that requires both labor and non-labor inputs. ⁽¹⁾ Therefore the Commission recommends that:

The Secretary should revise the productivity adjustment for physician services and make it a multifactor instead of labor-only adjustment.

Productivity gains are certainly possible in physician services and should be taken into account. For example, research suggests that doubling the size of a physician practice (from the current average of about 2.5 physicians to 5 physicians) increases productivity by 9 percent with no increase in practice expense per physician (Pope and Burge 1996). Physicians apparently perceive the advantages of group practice: in 1990, 52 percent of self employed physicians were in solo practice, but by 1998, that percentage had dropped to 42 percent. Other gains might come from new technology, economies of scale, managerial skill, and changes in how production is organized.

In other health care delivery settings such as hospitals, MedPAC assumes that cost savings from improved productivity are usually offset by cost increasing factors such as scientific and technological advances or complexity changes within service categories. However, Medicare's payment system for physician services accounts for those cost increasing factors by either, creating new billing codes or revising existing codes in the physician fee schedule, or by recalibrating the fee schedule's relative weights every five years. Thus, those cost increases do not offset cost decreases from productivity and productivity must be accounted for separately.

Productivity growth is the ratio of growth in outputs to growth in inputs. Measuring productivity growth requires detailed information on the personnel, facilities, and other inputs used and on the quantity, quality, and mix of services (outputs) produced. Because such data are generally not available, MedPAC has adopted a policy standard, 0.5 percent, for achievable productivity growth that is based on growth in multifactor productivity in the national economy. Such a measure should be used in the physician update as well.

In making its update for physician services in 2003 MedPAC considered three things: the adequacy of Medicare physician payment in 2002, the inflation in input prices projected for 2003, and an adjustment for multifactor productivity. Although payments for physician services have not kept pace with the change in input prices since 1999, MedPAC recommends no adjustment for payment adequacy at this time, pending collection of further data. The other components of the update are the estimate of the change in input prices for 2003, which is 3.0 percent, and MedPAC's adjustment for growth in multifactor productivity, which is 0.5 percent. Therefore:

The Congress should update payments for physician services by 2.5 percent for 2003.

Our assessment of the first two components of our update, payment adequacy and inflation in input prices, is discussed briefly in the following sections.

Assessing payment adequacy

Is the current level of Medicare's payments for physician services adequate? The information available to answer this question is limited and better measures of payment adequacy are needed. We lack information on the cost of physician services, so we cannot compare Medicare's payments and costs the way we can for other services, such as hospital care. However, we do have information about several other factors that allow us to judge the adequacy of payments. This information includes data on the number of physicians furnishing services to Medicare beneficiaries, physicians' perceptions of the Medicare program and their willingness to furnish services to beneficiaries, and information from surveys of beneficiaries on their ability to obtain care and their satisfaction with the care received. However, because it takes some time for providers to respond to changes in payment, these indicators may lag behind payment changes and must be interpreted carefully. Additional measures of payment adequacy are needed that are sensitive to possible short-term effects of inadequate payments, such as the duration of office visits and changes in the volume of services.

Available information suggests that, as of 1999, payments were not too low. From 1999 onward, we have very limited data; we do know, however, that payments did not keep up with increases in input prices. This suggests that payments for 2002 could be too low, raising concerns about beneficiaries' access to care. We will not know if payments are too low until we have further information on payment adequacy. One source of that information will be MedPAC's newest survey of physicians which will be fielded this spring.

Entry and exit of providers

Data on provider entry and exit yield information regarding the adequacy of current payments. Rapid growth in the number of providers furnishing services to beneficiaries may indicate that Medicare's payment rates are too high. Conversely, widespread provider withdrawals from Medicare could suggest that the rates are too low.

Counts of physicians billing Medicare show that the number of physicians furnishing services to beneficiaries has kept pace with growth in the number of beneficiaries. From 1995 to 1999, the number of physicians per 1,000 beneficiaries grew slightly, from 12.9 to 13.1.

Physician willingness and ability to serve Medicare beneficiaries

MedPAC's 1999 survey of physicians suggests that physicians were willing and able to serve beneficiaries.

• Only about 10 percent of physicians reported any change between 1997 and 1999 in the priority given to Medicare patients seeking an appointment. Of those changing their appointment priorities, the percentage that reported giving Medicare patients a higher priority was almost the same as the percentage that assigned Medicare patients a lower priority.
• Only 4 percent of physicians said that it was very difficult to find suitable referrals for their fee-for-service Medicare patients, a finding comparable to the percent who reported problems referring their privately insured fee-for-service patients.

One of the most important findings of the survey was that, among physicians accepting all or some new patients, more than 95 percent said they were accepting new Medicare fee-for-service patients--a finding consistent with the results of another recent survey.

While these findings are positive, many doctors participating in MedPAC's survey expressed concerns about payment levels. About 45 percent said that reimbursement levels for their Medicare fee-for-service patients were a very serious problem although, even more, about 66 percent, said that HMO reimbursements were a very serious problem.

Beneficiaries' access to care

 Another way to assess the adequacy of payment rates is to evaluate beneficiaries' access to and quality of care. Evidence of widespread access or quality problems for beneficiaries may indicate that Medicare's payment rates are too low. Access and quality measures are often difficult to interpret, however, because they are influenced by many factors. Access to care for specific services, for example, may be influenced by beneficiaries' incomes, secondary (medigap) insurance coverage, preferences, local population changes, or transportation barriers, all of which are unrelated to Medicare's payment policies.

Access to care was not a problem in 1999, according to data from the Medicare Current Beneficiary Survey. The percentage of beneficiaries reporting trouble getting care (4 percent) was low and essentially unchanged from previous years. The data also show that beneficiaries were overwhelmingly satisfied with the care they received. We will continue to track these indicators as newer data become available.

Accounting for cost changes in the coming year

Given the information about the adequacy of the current level of payments, the next step in determining payment updates is to ask how much costs will change in the coming year. Several

factors will affect the cost of physician services, but the most important one is inflation in input prices. The available measure--the MEI--has two problems, but the Secretary can correct them. Other factors that may increase costs include scientific and technological advances and the regulatory burden of the Medicare program, including the burden of compliance with requirements of the Health Insurance Portability and Accountability Act of 1996. These other factors are likely to have small or unmeasurable effects on costs. The remaining factor--productivity growth--will reduce costs. Using appropriate measures of inflation and productivity growth, it appears that the cost of physician services will increase by 2.5 percent during the coming year.

Measuring inflation in input prices The MEI is the SGR system's measure of input price inflation. It is calculated by CMS as a weighted average of price changes for inputs used to provide physician services. Those inputs include physician time and effort, or work, and practice expense. Physician work, which accounts for the time, effort, skill and stress associated with providing the service, has a weight of 54.5 percent; the remaining 45.5 percent is allocated among categories of practice expense. Practice expense includes support staff wages and benefits, office expense, medical materials and supplies, professional liability insurance, medical equipment, and other professional expenses, such as private transportation.

Although the MEI is analogous to the market basket indexes used to update payments for inpatient hospital care, it currently differs from those indexes in that it includes an adjustment for productivity growth. Productivity growth is an important factor and MedPAC believes that it should be considered separately in update decisions. This would allow input price indexes to account only for changes in prices, not other changes in cost.

As used in the SGR system, the MEI also differs from the market basket indexes in that it is not a forecast of the change in input prices for a given year, but a measure of input price inflation for the previous year. To allow payment updates to anticipate changes in costs during the coming year CMS should use a forecast of the MEI when making payment updates for physician services.

By removing the productivity adjustment and making it a forecast, the MEI would become a better measure of input price inflation. So modified, the index shows that input prices for physician services are expected to increase by 3.0 percent in 2003.

Other cost-increasing factors

The cost of physician services may increase because of factors other than changes in input prices. The overall effect of these factors is likely to be small, however. As noted the costs of scientific and technological advances are already accounted for in the physician fee schedule when new billing codes are created or existing codes are revised.

Other factors increasing costs are difficult to measure. For example, the regulatory burden of the Medicare program is an important concern of physicians. Nevertheless, estimates of the cost of this burden are not available. One way to account for any measurable increases in cost due to these factors is to assess payment adequacy, as described earlier, and adjust payments accordingly in future updates.

1. The labor-only adjustment may simply be an artifact. It has been part of the MEI since the index was first used in paying for physician services in 1975, which was before the Bureau of Labor Statistics (BLS) began publishing measures of multifactor productivity in 1983.