Statement of Robert A. Berenson, M.D., Senior Fellow, the Urban Institute

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

May 10, 2007

Chairman Stark, Mr. Camp, and members of the Committee:

I appreciate the opportunity to provide testimony to the Health Subcommittee on a subject I have been deeply involved with through most of my professional career. I practiced internal medicine for over twenty years, twelve of which were in a group practice just a few blocks from here.  I was the first representative of the American College of Physicians to the American Medical Association’s Resource-Based Relative Value Scale (RBRVS) Update Committee (RUC). In the last part of the Clinton Administration, I had operational responsibility for the Medicare Physician Fee Schedule at the Centers for Medicare and Medicaid Services (CMS). Finally, in recent years as a Senior Fellow at the Urban Institute, I have had a chance to study how well the Medicare Physician Fee Schedule has worked and what might be done to improve it. 

I believe that this is an important hearing -- because the focus of the hearing is not on how to use marginal dollars – 1-2% -- to try to influence physician performance or on paying third-party disease management organizations that are separated from the physicians actually providing the medical care to beneficiaries with chronic conditions -- but rather explores how the program might better spend the 100% base of physician spending, which is now approaching $60 billion. It is important to explore the likely effects of these newer approaches to improving quality and efficiency on beneficiaries, physicians, and the Medicare program overall.

The hearing is also important because it signifies that the budgetary pressure of finding a solution to the shortfall created by the cumulative deficit produced by the sustainable growth rate (SGR) formula should not occupy all of the time and attention of health policy makers. Indeed, as I will try to make clear, I believe that greater attention to how we spend the base of $60 billion can provide both short-term and long-term improvement to the financial bottom-line and ease off some of the SGR pressure that currently exists. In recent months, very constructive ideas, including some presented at today’s hearing, have been raised. I hope to contribute to that discussion in my remarks today.

Many policy makers use the term “fee-for-service Medicare” to designate the original Medicare program and to distinguish it from the various kinds of Medicare Advantage products. However, this convenient short-hand actually mischaracterizes how the traditional Medicare program pays providers. Indeed, in a book on Medicare prospective payment that I co-authored with Rick Mayes last year, I emphasize that the Medicare Fee Schedule (MFS) is one the last payment approaches in Medicare that remains truly fee-for-service (FFS).[1]  Initially, with the Hospital Inpatient Prospective Payment System and then subsequently with a series of prospective payment systems created in the Balanced Budget Act of 1997 and later legislation, providers typically receive bundled payments for an episode of care, appropriately case-mix adjusted to take into account patient severity. Under these bundled payment approaches, providers have an incentive to provide services more efficiently, for less than the average costs on which payment amounts are based. Accumulated evidence documents that prospective payments based on episodes of care have moderated cost increases in the traditional Medicare program. 

In contrast, the physician payment system remains FFS, although even in the fee schedule there are significant examples of bundled or packaged payments, most notably the 90-day global fees for surgical procedures under which routine pre- and post- operative services are included into the global payment amount, and the monthly payment to renal physicians overseeing renal dialysis for patients with End Stage Renal Disease.  These long-standing approaches to bundling can be looked to for guidance on how to expand episode-based payments to physicians. 

Because the physician payment system is almost purely FFS, it was understandable that Congress, in OBRA 1989, placed a volume expenditure target – then called the Volume Performance Standard – as an admittedly crude approach to containing spending growth under the MFS that began in 1992. It is interesting to note that the 1989 Physician Payment Review Commission Report thought that the expenditure target mechanism could work only for a few years and that organized medicine needed to actively develop clinical practice guidelines, with accompanying physician education efforts, as a needed long-term solution to constrain volume growth. Unfortunately, efforts to find alternatives to the top-down expenditure target approach were not sustained. And the program is now experiencing an explosion of volume and intensity growth in some clinical areas.

Yet, for the first decade or so of the MFS, the evolving expenditure target approaches actually worked reasonably well to constrain spending growth. The situation has clearly changed in the past 6 years, and Congress, with the exception of 2002, has acted to override the across-the-board fee reductions called for under the SGR mechanism.  In the absence of broad-based clinical practice guidelines and because of the volume growth of services that are inherently discretionary in nature and, increasingly, under physicians direct control, in my opinion there is little question that bundling payments for episodes of care needs to be a primary objective of physician payment reform, just as it has been successful when applied to other providers in Medicare.

Examples of Bundled Services

I will provide one important example of why moving to bundled payments for physicians, in contrast to fee-for-service, makes good policy sense. The work of Dr. Edward Wagner, at the MacColl Institute for Healthcare Innovation in Seattle, Washington, on what he calls the Chronic Care Model makes clear that the proper management of patients with one or more severe chronic conditions, such as diabetes and congestive heart failure, involves lots of communication with patients outside of standard office visits by phone and, possibly, email; care by multi-disciplinary professional teams; active use of patient registries; and enhanced coordination among professionals and providers practicing in many locations. In my view, it would be foolhardy to try to pay for most of these additional services on an a la carte basis, as FFS does.

Consider, as an example, phone calls. The transaction costs of billing and collecting would be more than the reimbursement for most of the individual services; program integrity concerns would abound; and the inevitable explosion of volume on easily provided and well-appreciated phone calls would become financially prohibitive. The alternative that MedPAC and others have discussed is a chronic care management fee for primary and principal care physicians who would agree to be accountable for providing the array of services in the Chronic Care Model, much as the American Academy of Family Practice, the American College of Physicians and others have envisioned in the patient-centered medical home.  My own preference would be to provide a “per beneficiary per month” fee not only for care coordination but also for some or all of the actual medical services provided by the same practice.[2] The right approach, which should be tested in multi-payer demonstrations, might actually be a mixture of reduced fee-for-services combined with monthly fees for specified bundles of services.

The medical home concept presents a number of specific operational challenges, which I am prepared to discuss, but the main point to make is that it is the conceptually right thing to do. The approach not only should improve the care provided to beneficiaries with chronic health problems, but importantly, would provided involved practices with improved incentives to avoid unnecessary downstream utilization by other providers. In this context, pay-for-performance to reward efficiency and to protect against under-provision of important primary and secondary preventive services might play a useful, supportive role.

Episode-based payment not only for primary care physicians but also for specialists caring for a variety of acute and chronic health care medical problems has inherent appeal. There, are, however, important implementation issues regarding specialist bundling as well. In particular, given the documented problem of inappropriate procedures producing unjustifiable and costly practice variations, any episode-based payment system should not incorporate an inherent bias for performance of procedures, as already exists in the RBRVS-based fee schedule. Although the costs of an episode need to recognize that there are direct physician expenses associated with the procedure provision itself, the valuation of condition-specific episodes should minimize payment differentials that reward clinical decisions to provide the procedural intervention.

Further, as with all episode or period of time based payment approaches, clinically sophisticated case-mix adjustment is needed to prevent perverse effects, such as physicians giving preference to less severe patients within a cohort with a particular condition or over-diagnosing relatively minor complaints to generate compensable episodes.  All payment systems offer “gaming” opportunities. The work on developing payment bundles and episodes needs to protect against such behavior.  Fortuitously, in recent years, we now have much more sophisticated approaches to case-mix adjustment such that payment approaches, such as capitation, that often foundered when used by private health plans in the past, now might be much more successful.  

 One size no longer fits all

It is time to recognize that a “one size fits all” physician payment system may no longer work properly to support the increasing diversity of physician activity that has resulted from sub-specialization. Primary care physicians and particular sub-specialists typically care for patients over many years, and much of their value derives from continuity and consistency. As already noted, an immediate Medicare challenge is to develop a payment approach to support robust chronic care coordination and management. At the other end of the physician spectrum, some physicians, including radiologists, pathologists, anesthesiologists, and emergency room physicians, mostly provide one-time, discrete services and typically do not have ongoing responsibilities regarding individual patients. For these physicians, FFS would seem to be an appropriate reimbursement mechanism for a third-party payer, such as Medicare, which does not employ physicians and thus are unable to pay a salary. In the middle of the spectrum, many physicians provide both discrete, one-time services and have ongoing care responsibilities.

Ideally, all specialties would work together, either in real multi-specialty group practices or in virtual multi-specialty collaborations, with payment made to the organization on a per beneficiary per month basis for “medical home” services, with payment adjustments for episodes of illness that require highly specialized services. The current physician group practice demonstration is a very important one in recognizing the opportunity to compensate large real and virtual groups differently from the payment approaches that apply to individual physicians or single specialty groups. Further, physician pay-for-performance, generally should attempt to measure group-level, rather than individual, physician performance.

In sum, Medicare should develop and maintain different payment approaches for multi-specialty groups and collaboratives able and willing to be accountable for costs and quality, rather than pay them on the lowest common denominator approach that would apply to a solo practitioner. At the same time, FFS will be with us for a long time – for those physicians unable or unwilling to accept bundled payments that places them at significant financial risk and for physicians outside of large groups who provide specialized, one-time services. 

I have recently co-authored medical journal articles critiquing recent implementation of the MFS, especially the RBRVS component.[3] But I do not want these published comments and concerns to be misunderstood. The RBRVS approach, first implemented in 1992 and still a work in progress, was a marked improvement over the charge-based fee schedule that preceded it in Medicare. And for all of RBRVS’s complexity, the right institutions are in place to make important and overdue improvements to the fee schedule refinement process. Unfortunately, the MFS, I believe, has suffered from a relative lack of attention in recent years by Federal policy makers – at CMS, at MedPAC, and in Congress, as policy interest has focused elsewhere. As a result, the program has spent unnecessarily because of a failure to anticipate and guard against highly inflationary increases in the volume and intensity of many physician services.

To use a sports metaphor, attempting to get the prices right is the blocking and tackling of a fee schedule. Yet, in recent years, fee schedule prices have become distorted, but without much notice. These pricing distortions have occurred in Medicare but even more so in most commercial health plan fee schedules, which are based on Medicare’s. Prices have been allowed, increasingly, to deviate from the underlying costs of production, producing unfortunate behavior responses by physician, which I will detail in a moment. Yet, in my view, it would be relatively straight-forward technically to correct the distorted prices, if there were the political will and support to do so.

In the recent articles, colleagues and I have attempted to explain some of the technical reasons why the prices became distorted. I will emphasize two issues here. Keeping the relative values accurate requires an effective process that reflects changes in medical practice and trends in physician productivity. But, for the most part, relative values have defied gravity – going up or staying the same but rarely coming down.[4]  Because physician time spent is a crucial element in estimating both the work and practice expense components that make up the RBRVS approach, it is time to base time elements for high frequency services on objective time data, rather than on surveys of self-interested specialty groups. In that way, time estimates can be kept more current and accurate than under the five-year review process that is now used.

Second, problems with accurate estimation of relative values for practice expenses have worsened as physicians in some specialties have billed for more ancillary services associated with high equipment expenses. CMS has used unrealistically low assumptions about rates of use if equipment and unrealistically high assumptions about amortization rates for large equipment purchases. Furthermore, the payment of average costs for services whose variable costs are low encourages physicians to order more services and to view the services as profit centers. These services include imaging and clinical tests, which are among the fastest growing services in Medicare. In short, because of the failure to consider that the cost of providing a service such as an MRI scan is reduced with every scan performed, Medicare’s reimbursements overpay and create an incentive for ordering and providing too many such scans.

 During site visits to twelve nationally representative metropolitan areas through work conducted by the Center for Studying Health System Change, my colleagues and I have observed increasing numbers of physicians building capacity to compete with hospital outpatient departments by offering these lucrative services.[5] Indeed, such market-based developments provide a direct signal to policy-makers of distorted payment levels, pointing to priority targets for price error corrections.

It is not by simple chance that CMS and MedPAC find the volume and intensity of imaging, tests, and minor procedures – all discretionary services which ostensibly produce little or no patient harm – are growing much faster than the categories of major surgical procedures and evaluation and management services. The latter services provide much less opportunity for physician-induced demand.

We now see that single specialty groups are merging to have the size and scope to purchase or lease imaging equipment, such as MRI and PET scans. This behavior suggests that the prices for advanced imaging services, such as MRI and PET scans are too high and can be safely reduced without compromising patient access to these important services. (Conversely, other imaging services, such as screening mammograms and DEXA scans for osteoporosis, where access problems appear to exist are likely under-priced.)

There are many technical reasons for why the RBRVS system has gotten off track. The Congress can play an important role in assuring that the technical experts within organized medicine, at MedPAC, and at CMS make the needed corrections to currently distorted prices. And while work proceeds to adopt bundled-based payments for physician services, in my opinion there remains a strong policy rationale for expenditure targets, but specifically targeted to discretionary services that are growing rapidly. In sum, in the long-term we need fundamental reform of how physicians are paid in traditional Medicare. In the short-term, greater attention to correcting incorrect prices and more carefully targeting expenditure targets can produce savings and produce the climate needed to accomplish the needed fundamental reforms that the witnesses have discussed at this hearing.

[1] Rick Mayes and Robert A. Berenson, Prospective Payment and the Shaping of U.S. Health Care, (Baltimore: Johns Hopkins University Press, 2006)

[2] Goroll, HA, Berenson RA, Schoenbaum SC, Gardner, LB. Fundamental Reform of Payment for Adult Primary Care: Comprehensive Payment for Comprehensive Care. Journal of General Internal Medicine, 22(3):410-415, 2007.

[3] Bodenheimer T, Berenson RA, and Rudolf P. The  Primary Care-Specialty Income Gap: Why It Matters, Annals of Internal Medicine 146(4):301-306, 2007; Ginsburg PB and Berenson RA. Revising Medicare’s Physician Fess Schedule – Much Activity, Little Change. New England Journal of Medicine 356(12):1201-1203, 2007; Maxwell S, Zuckerman, S, and Berenson RA. Use of Physicians’ Services Under Medicare’s Resource-Based Payments, New England Journal of Medicine 356(18):1853-1861, 2007. 

[4] Ginsburg and Berenson

[5] Berenson RA, Bodenheimer T and Pham, HH. Specialty-Service Lines: Salvos in the New Medical Arms Race, Health Affairs 25:w337-w343, 2006.