Correspondence
Physician-Payment Reform
N Engl J Med 1993; 329:808-810September 9, 1993
- Article
To the Editor:
Hsiao et al. (April 1 issue) assess the implementation of physician-payment reform and conclude that the “misallocation of practice expenses in the Medicare fee schedule results in serious underpayment for medical services”1. As a general internist practicing primary care internal medicine, I have two specific concerns about their data.
Table 2 of the article suggests that overhead for internal medicine should be approximately $129,800 a year on the average, and Table 5 that the estimated actual (nonphysician) practice costs should be approximately $139,000 a year. My office practice costs and overhead are closer to $190,000 to $200,000 a year. I conducted a brief informal survey of my colleagues and found that their overheads ranged from $140,000 to $200,000 a year. I considered the possibility that my overhead might be relatively high, given that I dictate all my records and employ a full-time transcriber. However, in reviewing the Medical Group Management Association cost survey,2 I found that the median nonphysician expenses of internists in single-specialty groups were $171,884. The reimbursement system should not be based on data that suggest that an internist's expenses are substantially lower than they really are. The administrative burden associated with outpatient care is likely to increase as governmental scrutiny and insurance policies become more complex.
I am also concerned about the usual workweek of 60 hours described for full-time physicians. I agree that this figure is approximately average, yet many physicians work more hours. Given the reimbursement system, some physicians cannot pay the overhead unless they work 60 hours or more. That currently is the reality for me. I love medicine, but I don't love it that much. At a time when state legislatures are establishing upper limits on the number of hours that residents may work, I am disturbed by reimbursement policies that dictate a lower limit on the number of hours that physicians in practice must work. Although medicine has a history of long working hours, they take a toll on physicians' well-being, family life, and relationships with patients.
2 ReferencesJames J. Miller, M.D.
1322 E. Michigan, Lansing, MI 489121
Hsiao WC ,Dunn DL ,Verrilli DK . Assessing the implementation of physician-payment reform. N Engl J Med 1993;328:928-933
Full Text | Web of Science | Medline2
1992 Cost survey report based on 1991 data. Englewood, Colo.: Medical Group Management Association, 1992.
To the Editor:
Hsiao et al. ignore the problem of the income physicians lose when patients fail to pay the Medicare deductible or coinsurance. The U.S. Health Care Financing Administration sets a maximum fee but reimburses only 80 percent. In my experience, no practitioner has been 100 percent successful in collecting the remaining 20 percent. Because the explanation-of-benefit forms from Medicare are often difficult to understand, many patients have adopted the attitude that they are not legally responsible for the copayment. Depending on the demographics of one's community, a net collection of 85 to 90 percent from Medicare patients is considered good. The remaining 10 to 15 percent of the fee is an income write-off attributable to practice expense. Pursuing this money creates an overhead cost for billing services, and many practitioners find it cheaper to write off small coinsurance amounts than to collect them. The public benefits, but the physician's bottom line suffers. . . .
Jeffrey L. Kaufman, M.D.
Baystate Medical Center
Springfield, MA 01199To the Editor:
Hsiao et al. argue two points that have merit -- that the relative values for practice expenses in the Medicare fee schedule are distorted, and that the Medicare conversion factor is probably not an appropriate basis for an all-payer system. Their empirical analysis, however, has little to do with these points and has severe problems with face validity. The authors imply that implementation of the fee schedule had effects on physicians' incomes that were the opposite of the effects described in their seminal study1. More careful empirical work indicates that a universal application of this fee schedule would reduce, not exaggerate, differences in income among specialties.
The problem with face validity is seen in the simulations using data on submitted charges. The simulation suggests that “potential incomes” of physicians vary 13-fold across specialties, with family physicians earning $70,000. But survey data show actual incomes vary threefold across specialties, with family physicians' income exceeding $100,000. The simulations of the Medicare fee schedule are similarly distorted. Although surgical fees were reduced and visit fees increased, the potential income of thoracic surgeons is simulated as being only slightly lower than the current income. By comparison, the potential income of family physicians is less than half of the current level.
There are pitfalls in basing this analysis on a single service per specialty. Most specialties provide a range of services. Measurement of the amount of work per unit of time for a single service may be imprecise. Differences between that single service and the range of services in the specialty may create large errors. It is unnecessary to extrapolate from data on a single service, since claims data from Medicare and private insurers can be combined.
Using extensive claims files from Medicare and private insurers, we simulated physicians' net incomes, assuming that all payers used the Medicare fee schedule. Our simulation reflected both the restructuring of payment from the fee schedule and the fact that Medicare payment rates are approximately 35 percent less than the average rates of private insurers. Under a worst-case assumption that balance billing, practice expense, and service mix did not change in response to these payment changes, we estimated that family physicians and orthopedic surgeons would earn $90,000 and $140,000, respectively. Under more realistic assumptions, we estimated incomes of $100,000 and $190,000, respectively.
The fact that Medicare payment rates are much lower than those of private insurers has little to do with the fee schedule. The pattern of lower rates for Medicare reflects numerous policy decisions made since the program's inception in 1966. Congress called for the Medicare fee schedule to be implemented on a “budget neutral basis.” According to the calculations of the Physician Payment Review Commission, the conversion factor was set too low by 4 percent. This miscalibration may be resolved automatically through the mechanism of Medicare volume-performance standards2.
We agree that the practice-expense component of the Medicare fee schedule should be based on estimates of resources rather than on historical charges. The Commission has developed such a method and has recommended it to the Congress2.
2 ReferencesPaul B. Ginsburg, Ph.D.
Christopher Hogan, Ph.D.
Physician Payment Review Commission, Washington, DC 200371
Hsiao WC ,Braun P ,Yntema D ,Becker ER . Estimating physicians' work for a resource-based relative-value scale. N Engl J Med 1988;319:835-841
Full Text | Web of Science | Medline2
Annual report to Congress, 1993. Washington, D.C.: Physician Payment Review Commission, 1993.
To the Editor:
In their editorial on the specialty distribution of physicians in the United States, Schroeder and Sandy (April 1 issue)1 fail to mention another important explanation for the oversupply of specialists -- the large number of training programs available, especially in the subspecialties of internal medicine. One reason for this is that the training programs, which are mostly at universities, need cheap labor. Sooner or later, the country is flooded with subspecialists as a result.
1 ReferencesDavid Glaser, M.D.
Fountain Valley Regional Hospital, Fountain Valley, CA 927081
Schroeder SA ,Sandy LG . Specialty distribution of U.S. physicians -- the invisible driver of health care costs. N Engl J Med 1993;328:961-963
Full Text | Web of Science | Medline
To the Editor:
Schroeder and Sandy fail to discern that the practice patterns of primary care physicians in the United States, carefully nurtured in our academic centers, are as much a factor in driving up health care costs as the practice patterns of specialists. The best medical resident has the broadest differential diagnosis and the most detailed problem list. The cascade toward further testing and interventions, often performed by specialists, is as certain as that night follows day. . . .
Charles F. Von Gunten, M.D., Ph.D.
Northwestern University Medical School, Chicago, IL 60611Author/Editor Response
The authors reply:
To the Editor: Drs. Ginsburg and Hogan state that our paper implies that universal application of the Medicare fee schedule would exaggerate the differences in income among specialties. We did not imply this anywhere in our paper. In fact, we showed in our Table 4 that the Medicare fee schedule, based on the resource-based relative value scale, narrows the differences in income among specialties.
Our central focus was to develop a method that could objectively evaluate what physicians would earn if they performed the services that they were trained to perform in their specialties. We argued that the current practice profiles do not represent the services that each specialist should be performing, because specialists' practices are distorted by the existing overpayment or underpayment for various services and by the oversupply of physicians in some specialties. For example, most family physicians perform more radiography, electrocardiography, diagnostic tests, and minor procedures because these services generate more net income than office visits. Many surgeons spend a substantial portion of their time providing nonsurgical services because of the oversupply of surgeons. The average income of surgeons is depressed because nonprocedural work is reimbursed at a lower rate than procedures. Therefore, basing a simulation on current practice profiles would incorporate the distorted practice patterns produced by present market conditions. Our method would reduce these distortions. We made it clear that our method would yield net income that differed from actual income, and explained the reasons.
Dr. Kaufman highlights the need for a uniform and objective method to evaluate fee schedules. We believe that our standard-service method provides such an approach to analysis of monetary conversion factors and the reimbursement of practice expenses across specialties. It can easily be modified to include losses due to bad debts.
Two issues are central to Medicare's reimbursement of practice costs. Current law on the allocation of costs should be changed. As for measuring practice costs, Dr. Miller believes that for primary care physicians these costs are higher than the median amount that we stated. This question can be settled only by more complete and up-to-date cost studies.
William Hsiao, Ph.D.
Daniel Dunn, Ph.D.
Harvard University School of Public Health, Cambridge, MA 02138Diana Verrilli, M.S.
Urban Institute, Washington, DC 20037Author/Editor Response
Dr. Glaser is correct in highlighting the “supply side” forces that foster growth in specialty training positions. The total number of residents in allopathy has grown 24 percent since 1981, in spite of a relatively stable output of graduates by U.S. medical schools. In internal medicine, with a 5 percent increase in total number of residents, the number of residents in cardiology, gastroenterology, and pulmonary disease has more than doubled over the past decade1. This growth has been fueled by an increase in international medical graduates training in internal medicine. In the 1993 National Residency Matching Program, these graduates accounted for 27 percent of all residents in internal medicine who were matched2. Training programs have been extraordinarily myopic in failing to see that there is insufficient disease prevalence to support the current output of subspecialists. Perhaps more important, academic medical centers are now beginning to see the effects of training their own competition, as their trainees duplicate in community hospitals the tertiary and quaternary care previously available only in university hospitals. As the largest potential generalist discipline, internal medicine has to take the lead in providing the nation with an adequate supply of generalist physicians.
Dr. Von Gunten makes the case that residency programs at academic health centers train physicians, even generalists, for an elaborate, expensive style of practice. Indeed, evidence suggests that graduating residents do not receive training in basic areas necessary for practice3. Moreover, family physicians, with more extensive training in non-hospital primary care centers, appear to be more parsimonious with resources than internists4. Clearly, residency programs need to establish more appropriate training sites for primary care beyond present-day teaching hospitals and to ensure that trainees are well versed in disciplines such as clinical epidemiology, ethics, cost-effectiveness analysis, and psychosocial dimensions of care. Knowledge and skills in these areas are critical to effective practice.
Academic health centers, as public trusts, have a responsibility to produce physicians who meet the nation's need for high-quality, affordable, and humane care.
4 ReferencesSteven A. Schroeder, M.D.
Lewis G. Sandy, M.D.
Robert Wood Johnson Foundation, Princeton, NJ 08543-23161
Annual report to Congress, 1993, Washington, D.C.: Physician Payment Review Commission, 1993:58.
2
Page L. Family practice gains big numbers in resident match. American Medical News. April 5, 1993:5.
3
Martin GJ ,Curry RH ,Yarnold PR . The content of internal medicine residency training and its relevance to the practice of medicine: implications for primary care curricula. J Gen Intern Med 1989;4:304-308
CrossRef | Web of Science | Medline4
Greenfield S ,Nelson EC ,Zubkoff M , et al. Variations in resource utilization among medical specialties and systems of care: results from the Medical Outcomes Study. JAMA 1992;267:1624-1630
CrossRef | Web of Science | Medline
