urban institute nonprofit social and economic policy research

Just Why Do We Adjust Medicare Physician Fees for Geographic Practice Cost Differences?

Testimony before the Committee on Ways and Means / Subcommittee on Health

Share on Facebook Share on Twitter Share on LinkedIn Share on Digg Share on Reddit
| Email this pageE-mail
Document date: July 23, 2002
Released online: July 23, 2002

The views expressed in this Testimony are those of the author and do not necessarily reflect the views of the Trustees and employees of The Urban Institute.

Chairman Johnson and members of the committee, I appreciate the opportunity to appear before you today to discuss the geographic practice costs adjustment in the Medicare Physician Fee Schedule. My name is Stephen Zuckerman and I am a Principal Research Associate at the Urban Institute, a non-profit, non-partisan research institute located in Washington, D.C. Along with Gregory Pope at the Center for Health Economics Research in Waltham, MA and W. Pete Welch, a former colleague of mine at the Urban Institute, I co-directed the development of the practice cost adjusters that were adopted for use in the Fee Schedule in 1992. I also worked on the first revision of the adjusters in 1995. The conceptual basis for the geographic cost adjusters has not changed in the intervening years.

There has been widespread agreement that, under the Medicare Fee Schedule, fees should be adjusted for geographic differences in costs incurred by physicians that are beyond their control. This suggests that a geographic practice cost index should reflect differences in wages for clinical and administrative staff, office rents and malpractice insurance premiums. However, the largest share of physician practice revenues represents the costs of compensating the physician for his or her own time, and there has been considerable debate over how geographic differences in these costs should be taken into account.

In my testimony today I will review the conceptual foundation for the geographic practice cost adjusters used in the Medicare Fee Schedule, emphasizing why we felt it was, and still is, appropriate to adjust for geographic differences in the costs of physicians' own time.1 I recognize that at the time the Fee Schedule was being developed some felt that physicians' work was the same in all areas of the country and, therefore, should be paid for at the same rate in all areas. This view still persists. However, I hope to show why, in the interest of creating an equitable compensation system, payments for physicians' own time should be allowed to vary in relation to costs of living and other factors.

Physicians' Own Time. The fundamental reason to allow for geographic variation in the costs of physicians' own time is to create fees that compensate physicians at the same real rate in all areas of the country. An area's real rate of compensation can be thought of as the ratio of the dollar payment to the area's costs. Although a cost-of-living adjuster is an intuitively appealing measure of an area's costs, that is not what we viewed as a desirable adjuster for the costs of physicians' own time. A cost-of-living adjuster would over-adjust fees by not taking into account the impact that an area's amenities might have on compensation. Amenities differ across areas due to professional factors such as access to quality colleagues and the presence of modern hospitals and medical technologies, and due to personal factors such as availability of good schools, proximity to cultural events, and clean air. Because of these differences in amenities, compensation will vary less across areas than costs of living.

Economics predicts that compensation would not fully reflect an area's high costs of living if the area had desirable amenities. Desirable amenities would be a type of compensation of their own and offset some of the high costs of living. For example, workers are willing to locate in Honolulu despite its high cost of living because of its attractive environment. Similarly, for low-cost areas with poor amenities to attract and retain physicians, compensation would have to exceed costs of living. For example, if physicians value urban amenities, they would need to be paid more relative to the cost of living to locate in rural areas. Over time, compensation differences across areas would adjust so that a physician who is deciding where to locate would not care in which area he or she locates. Properly adjusted, Medicare physician payments should tend to promote an adequate supply of physicians in both urban and rural areas.

If not costs of living, what can be used as a geographic adjuster of physician time costs to equalize real compensation? Data on geographic variation in physician earnings are available from the Census. However, it would have been inappropriate to use these data to adjust payments under the Medicare Fee Schedule because these earnings were, in part, determined by historical patterns of Medicare payment rates. Further, these data are hard to work with because they cannot be adjusted to control for specialty mix differences across areas and because they reflect the profitability of physicians' practices as well as earnings.

As an alternative, we used hourly earnings of workers in professional occupations with five or more years of college education to derive a proxy for the physician work component of the geographic practice cost adjuster. This group of highly educated workers can be viewed as being similar to physicians with respect to the types of goods and services they purchase and their preferences for area amenities. Therefore, they should have earnings that reflect the appropriate amount of geographic variation that should be captured in the Medicare Fee Schedule. In addition, this adjuster did not perpetuate distortions that may have been present in the geographic distribution of physician earnings. Essentially, we argued that the geographic variation in payments for physicians' own time should reflect the variation of earnings for other highly educated professionals.

The adjuster we developed based on professional earnings ranged from about 24 percent above the national average in Manhattan, New York to about 20 percent below the national average in rural areas of Missouri. After considerable sensitivity analysis, we concluded that this appeared to be the most defensible adjuster for physicians' own time costs. As the policy was implemented, a geographic adjuster was incorporated into the Medicare Fee Schedule that reflected only one-quarter of the geographic variation in professional earnings. Primarily as a result of this decision, the physician work value in Manhattan, New York is about 9 percent above average and about 5 percent below average in rural Missouri.

These examples show that there is considerably less variation in the actual physician work adjuster used in the Fee Schedule than in the one we had derived in our original research. The three legislatively required revisions to the index over the past decade have not resulted in movement away from the decision to reflect only one-quarter of the variation in professional earnings in the physicians' work adjuster. Although the one-quarter work approach is not consistent with our original conceptual and empirical work, it still may be a credible geographic adjuster. Research has shown that geographic variation in employee physician wages is very closely related to the variation in the one-quarter work adjuster.2 Therefore, retaining that adjuster throughout the three revisions may have been a good decision. However, this research also concluded that the one-quarter adjuster is superior to allowing for no geographic adjustment, suggesting that an adjuster with less geographic variation would not be advisable.

Because we were aware of concerns about the level of fees that could exist in rural areas under the Fee Schedule, we gave explicit consideration to this issue in our final report to the Health Care Financing Administration.3 Our analysis suggested that the indices did a reasonably good job of tracking actual practice expenses across rural and urban physicians. However, we pointed out that one way of raising fees in low-cost rural areas would be to set an arbitrary floor on the practice cost adjusters. Depending on how such a policy was implemented, this could lead to lower fees in high-cost areas. Moreover, an arbitrary change in the index would mean that it was no longer capturing practice costs differences.

If the policy goal is to raise Medicare fees in areas that have problems recruiting and retaining physicians, it is reasonable to build on a different mechanism that is already a part of the payment system. Currently, the Medicare Fee Schedule includes a 10 percent bonus payment that is added to fees in Health Professional Shortage Areas. The bonus could be increased and/or extended to other areas. This could achieve the desired objective explicitly as opposed to a less transparent change, such as putting an arbitrary floor in the practice cost adjuster.

Other Practice Expenses and Malpractice Insurance. Aside from physicians' own time, the largest component of physician practice expenses is employee wages. To calculate an employee price adjuster, we used median hourly earnings of administrative support occupations, Registered nurses, Licensed practical nurses, and Health technologists and technicians (excluding LPNs). To reflect the occupation mix in physicians' offices, each category of hourly earnings was weighted to reflect the occupation's share of physician expenditures for employees.

The next most important expense category is office rents, but there are no nationwide data on rental rates for physician office space. However, the U. S. Department of Housing and Urban Development annually derives a "fair market rent" for all areas with a Section 8 housing assistance program. These data represent the 45th percentile rent for various sized units in each geographic market and were used as a proxy for a geographic adjuster of physician office rents. A key advantage of this price information is that it is available for all metropolitan areas and rural counties. A weakness of this proxy is that physician offices are in commercial as well as in residential buildings. However, residential and commercial rents are likely to be highly correlated because the same factors—such as population density, construction costs, and area income—are likely to affect both. The limited evidence shows that residential and commercial rents do tend to track each other across areas.

Geographic differences in malpractice costs are measured by comparing premiums charged for a mature claims-made insurance policy with $1 million/ $3 million limits of coverage. Premiums are averaged across the top twenty Medicare specialties, according to their shares of Medicare physician spending, so as to represent the full range of malpractice risk classifications. The data are derived from periodic surveys of malpractice insurers in all states and, where necessary, reflect intrastate variation in premiums charged.

Our review of available data uncovered no information on geographic differences in the prices of medical supplies and equipment. Anecdotal evidence suggested that price variation in these inputs is minimal. In computing the geographic adjuster, we assumed that the costs of these inputs as well as prices for "other" expense items were the same in all areas. Since only about 14 percent of total practice revenues are accounted for by these inputs, our approach with respect to the other inputs still captured the bulk of the variation in practice input prices.

Geographic Areas. Prior to the implementation of the Fee Schedule, carriers administered physician payments within state boundaries and had a great deal of discretion as to how fees would vary across geographic areas within their jurisdictions. Although there were many statewide "payment localities," some states had highly disaggregated payment areas. For example, Texas was divided into 33 payment areas for some specialties. We developed the set of geographic adjusters based on a more consistent set of criteria to define areas in all states. We wanted to base the index on areas that (1) had reasonably consistent prices for practice inputs within their borders; (2) were large enough to be a fairly self-contained market for practice inputs; and (3) were compatible with Medicare's administrative practices. Our decision was to use metropolitan areas and state rural areas as were and are being used in the Medicare Prospective Payment System for hospitals. We viewed this as striking an acceptable balance across the three criteria we had established.

The original localities were retained during the initial stages of the Fee Schedule implementation and our index was adapted for use in this geographic configuration of payment areas. Subsequently, Medicare has changed to a greater reliance on statewide payment areas with exceptions that allow for intrastate variation in those states with substantial within state practice cost variation.

Conclusion. As the debate over a geographic adjustment in the Medicare Physician Fee Schedule continues, the adjuster for physicians' own time costs remains the most contentious issue. Some argue that the physician labor market is a national market and, as such, physicians should be paid the same in all areas. Even if physicians are recruited from all areas of the country, that does not mean that their nominal level of compensation needs to be the same everywhere. However, as implemented, 75 percent of the payment for the physician work component of a Medicare fee is the same in all areas. The remaining 25 percent is appropriately adjusted to reflect differences in earnings that capture differences in costs of living and area amenities. Although this partial adjustment may not achieve the original objective of the geographic adjuster that we proposed—an equalization of real compensation across areas, it moves fees in the desired direction in all areas and should be retained.


1. A great deal of this testimony is based on research summarized in Zuckerman, S., W.P. Welch and G. Pope, "A Geographic Index of Physician Practice Costs," Journal of Health Economics 90(1), June 1990, pages 39-69.

2. Gillis, K., R, Willke and R. Reynolds, "Assessing the Validity of the Geographic Practice Cost Indices," Inquiry 30(3), Fall 1993, pages 265-280.

3. Welch, W., S. Zuckerman and G. Pope, "The Geographic Medicare Economic Index: Alternative Approaches," Urban Institute Working Paper 3839-01-01, June 1989.

Topics/Tags: | Health/Healthcare | Retirement and Older Americans

Usage and reprints: Most publications may be downloaded free of charge from the web site and may be used and copies made for research, academic, policy or other non-commercial purposes. Proper attribution is required. Posting UI research papers on other websites is permitted subject to prior approval from the Urban Institute—contact publicaffairs@urban.org.

If you are unable to access or print the PDF document please contact us or call the Publications Office at (202) 261-5687.

Disclaimer: The nonpartisan Urban Institute publishes studies, reports, and books on timely topics worthy of public consideration. The views expressed are those of the authors and should not be attributed to the Urban Institute, its trustees, or its funders. Copyright of the written materials contained within the Urban Institute website is owned or controlled by the Urban Institute.

Email this Page