University of South Carolina, Arnold School of Public Health, Dept. of Health Services Policy and Management, HSPM J712

Class 7 Notes -- Medicare Payment Reform

In preceding weeks, we've talked about the problems of a free-market health care system. The two big problems are:

  1. the lack of consumer sovereignty and
  2. the risk of expenses beyond the ability of most people to pay.
The solutions to these problems, respectively, were:
  1. professionalism in medicine, and
  2. health insurance, driven by providers
These solutions created their own problems:
  1. provider profiteering and administrative waste, at the expense of patients and payers
  2. incomplete coverage, leading to inadequate access for some people

Teaching economics and its critique simultaneously.

Does history matter?

Economists tend to ignore history. They imagine that this is a Newtonian world, a world of particles in which all you need to know is a particle's current make-up, its current position, and its velocity to tell what it's going to do. Market forces mold everything for the greatest possible efficiency.

The problem of hill climbing. If competitive assumptions don't hold, local optimization is not global optimization.

Why no turn in the US socialized medicine, unlike Europe and, later, Canada, and, recently, Taiwan?

Ideology shapes policy. Lack of strong labor movement, or farmer-labor movement? Presence in US of huge area of politically backwardness and racism? Canada: contrary to little house on the prairie, cooperation as well as individualism. A country that's New York + Minnesota.
General rallying cry in US: BIGGAS. An HMO to a tobacco activist: if I help you now, you'll be coming after me next.

A little more background on payment

Left out of Paying for Health Care article: charity

In Bodenheimer and Grumbach's old days scenario, care is a house call visit by a doctor

Hospitals -- where the poor went to die, like Mother Teresa's in India in recent times

Dispensaries in big cities, supported by philanthropists and religious organizations, did give primary care for the poor.

With the development of the germ theory of disease and aseptic surgery, and later antibiotics, the hospital became a place for progressively more expensive care, which they could no longer afford to give away.

Medicare Reform -- Government Sets Example for Reforming Incentives in Health Care

Medicare and Medicaid sought to solve the second of those problems, by extending insurance to the elderly and some of the poor. Medicare has been very successful at this. Medicaid, somewhat. These programs aggravated the provider profiteering problem, however. Both were designed to fit with existing professional relations and insurance systems.

This changed in the early 1980s, when Medicare took the lead in overthrowing the fee-for-service payment method, by introducing payment to hospitals by each patient's diagnosis-related group (D.R.G.). There were HMOs operating at the time, but, except in some special localities, they had not cracked the providers' power to set prices and terms of payment. Medicare had that power, and started a revolution in the industry by using it in this way.

Diagnosis-Related Groups for Hospital Payment

Medicare introduced DRGs in 1983, phasing it in through 1988. SC Medicaid adopted DRG-based payment in 1986. Modified to hybrid system in 1987.
DRG prospective payment system puts every patient into one of about 470 DRGs, according to the patient's diagnoses. The DRG determines the payment to the hospital (except for very long stay outliers).

Each DRG has a "weight" that represents the cost of treating such a patient relative to the average of all patients. A dollar figure is multiplied by the weight to give the payment level.

Here are some examples of weights for the SC Medicaid system:

(Payment is based on average unit payment for SC hospitals in May 1987, $1263.)
DRG Description Weight Avg. LOS Outlier LOS Payment
3 Craniotomy age<18 2.45 6.4 37.4 $3096
33 Concussion age<18 0.33 1.8 9.1 $417
106 Coronary bypass with
cardiac cath
8.92 10.9 25.8 $11272
371 C-section without
complicating condition
1.59 5.2 11.1 $2009
373 Vaginal delivery 0.69 2.3 5.8 $872

Some diagnoses and corresponding DRGs for full-term newborns:
391 Normal newborn 0.2883
390 Neonate with other significant problems 0.8347
389 Full term neonate with major problems 1.1672


 1 391 V3000
 2 391 V3001
 3 391 V310
 4 391 V3000 605
 5 391 V3000 7661
 6 391 V3000
 7 391 V3001
 8 391 V3001 7661
 9 391 V3000 605
10 391 V3000
11 391 V3001
21 391 V3000
22 391 V3000
23 391 V3000 7746
24 391 V3101
27 391 V3000 7746
28 391 V3000 7661
Typical 391 diagnoses:

 1 390 V3000 7525
 2 390 7786 7660
 3 390 V301 7526 V718
 4 390 V3000 7526
 5 390 V3101 7784
 6 390 V3000 4279
 7 390 V3000 37205
 8 390 V3000 76408
 9 390 V3000 7526
 10 390 V3000 7706
 11 390 V3001 71965 7706 7746
 12 390 V3000 7661 74910
 13 390 V3000 7793
 14 390 V3000 7706
 15 390 V3000 75501
 16 390 V3001 7526
 17 390 V3001 V718
Some 390 diagnoses:

 1 389 V3000 7701
 2 389 V3001 7731
 3 389 7701 7718 V3001 0389 7792
 4 389 7718
 5 389 V3001 7708 5531
 6 389 V3000 76408 7731
 7 389 V3000 7661 7731
 8 389 V3101 7731
 9 389 V3000 7731
10 389 V3000 7731
11 389 V3000 7454
12 389 V3000 7708
13 389 V3001 7731
14 389 7756 7824
15 389 V3001 7708 7718 75462 7746
16 389 V3001 7795 7863 7706
17 389 7732 7526
Some 389 diagnoses:

Incentives, desired and undesired, in DRG-based payment:

  1. Reduce costs, because hospital keeps difference as profit (or absorbs losses)
    1. more efficiency of providing services
    2. practice changes that lower costs: fewer tests, fewer procedures, cheaper drugs (Unless the provided service determines the DRG, as with C-sections.)
  2. Selection in favor of "product lines" that have high marginal revenue relative to marginal cost. Deselection of money-losing product lines. (Tempered by hospital's limited ability to control its own output level.) Results:
    1. Specialization of hospitals, if there are economies of scale.
    2. Oversupply (surplus) of facilities and docs in some product lines and undersupply (shortage) in others unless DRG weights can be adjusted.
Simborg, D.W., "DRG Creep: A New Hospital-Acquired Disease," N Engl J Med, June 25, 1981, 304(26), pp. 1602-1604.
DRG "creep" potential shown at the University of California at San Francisco Medical School hospital before DRGs were implemented. Error rates as high as 35% in identification of principal discharge diagnosis. Room for arbitrary classification. UCSF team found 159 patients in 1978 who had a major surgical procedure plus renal disease as second diagnosis. Switching priority of diagnoses would have raised DRG-based charge by over $5000 per case. 23% of all cases hospital saw could have been switched this way to increase reimbursement, had system been in effect. In the course of researching this, one doc developed a computer program to choose the highest-charge DRG (consistent with ethics, of course). That program, and others like it, became a hot seller.

Not on reading list. Hsia, D.C., Krushat, W.M., Fagan, A.B., Tebbutt, J.A., Kusserow, R.P., "Accuracy of Diagnostic Coding for Medicare Patients Under the Prospective-Payment System," N Engl J Med, February 11, 1988, 318(6), pp. 352-355.
Elaborate recoding effort on 7050 records from 239 hospitals. Federally funded and enforced. 21% error rate in DRG coding. 62% of errors favored hospitals.

Typology of coding changes:


Creep (This appears to be a more limited notion than Simborg's, though they cite only Simborg.)

Hsia et al concentrated on what they called "Creep." Errors raised case mix 1.9%.
Projecting to all Medicare gives extra creep cost of $300 million per year. Reabstracting all Medicare medical records would cost $200 million per year (8 million records at $25 each). Would that be cost-beneficial?

Baker, S.L., Kronenfeld, J.J., "Medicaid Prospective Payment: Case-Mix Increase," Health Care Financing Review, Fall 1990, 12(1), pp. 63-70.
Closer to home, this article looks at DRG creep when South Carolina Medicaid switched to DRGs in the mid-1980s. The newborn case mix value rose 67% in three months!

Not on reading list. Russell, L.B., Manning, C.L., "The Effect of Prospective Payment on Medicare Expenditures," N Engl J Med, February 16, 1989, 320(7):439-444.
Russell and Manning used Medicare's own expenditure data (actual and projected) to see if DRG payment saved money.
Hospital Trust Fund: 1988's projected 1990 expenditures are less than 1979's projected 1990 expenditures, adjusted for inflation.
Supplemental fund (pays for outpatient care) projected not to grow by enough to wipe out hospital cost savings.
Peer Review Organizations get much credit for decreasing admissions. No attempt to separate PRO from DRG contributions to projected expenditure decline.

Koescoff, J., et al, "Prospective Payment System and Impairment at Discharge," JAMA, October 7, 1990, 264(15), pp. 1980-1983.
Some, but not huge (less than I had expected based on stories circulating), increase in percent of patients discharged home in “unstable” condition.

Not on reading list:
Sager, M.A., Easterling, D.V., Kindig, D.A., Anderson, O.W., "Changes in the Location of Death After Passage of Medicare's Prospective Payment System," N Engl J Med, February 16, 1989, 320(7):433-439.
National mortality data from 1981-85.
More deaths in nursing homes (and relatively fewer in hospitals) after DRG-based payment than before. Little change in states exempted from DRG payment. Can't tell if transfers were medically appropriate.
Regression across states.
Reduction of hospital LOS big predictor of shift in deaths from hospital to nursing home. Dumping sicker patients on nursing homes that already have difficulty handling the patients they have.
States with high HMO enrollment had bigger increase in deaths in nursing homes (but not a bigger decline in hospital deaths).
Bigger change was in 1984, rather than 1985. Hospitals anticipated full implementation of PPS.

Not on reading list: Kane, N.M., Manoukian, P.D., "The Effect of the Medicare Prospective Payment System on the Adoption of New Technology," N Engl J Med, November 16, 1989, 321(20):1378-1383.
Cochlear implantation studied. DRG payment discourages new therapies that increase cost but improve outcomes, unless HHS Secretary changes payment or creates new DRG. Effort was made to get the change, but HCFA resisted. Result: Few implants done. Firms have stopped making the devices. Research and development stopped. When this article was assigned, students said that the demise of this technology was probably for the better.

Resource-Based Relative Value System for physician payment

In the late 1980s, Medicare led a direct attack on how physicians set their prices. This was the Resource-Based Relative Value System. It's now used, in various forms, by private as well as public payers.

RBRVS is not really "DRGs for docs." Can you say why? (Before I tell you, let's look at ...)

Roe, B.B., "The UCR Boondoggle: A Death Knell for Private Practice?" N Engl J Med, July 2, 1981, 305(1), pp. 41-45.
The background. Medicare used Usual and Customary Rates as a basis for pricing, which invites abuse. In 1981, a heart surgeon could do three 2-4 hour coronary bypass surgeries per week at $2500 each and make $350,000 annually.
Hsiao, W.C., Braun, P., Dunn, D., Becker, E.R., DeNicola, M., Ketcham, T.R., "Results and Policy Implications of the Resource-Based Relative-Value Study," N Engl J Med, September 29, 1988, 319(13), pp. 881-888.
Hsiao, W.C., Braun, P., Yntema, D., Becker, E.R., "Estimating Physicians' Work for a Resource-Based Relative Value Scale," N Engl J Med, September 29, 1988, 319(13), pp. 835-841.
RBRVS allows the government to set fees paid to physicians for specific services by simulating the fees the market would have set if the market functioned properly.
The first article gives the general idea. The second article (actually printed earlier in the same NEJM issue) looks specifically at how they measured the physician's work entailed by any particular procedure.
Hsiao was a major consultant to the Taiwan government for the reform of its health insurance system.
The Resource-Based Relative Value Scale (RBRVS) aims to make physician fees proportional to the cost of providing the services. A competitive market would do this by the invisible hand, but physician fees seem to be out of proportion to cost, with some surgical specialties much more handsomely reimbursed than primary care. RBRVS-based fees aim to encourage physicians to pursue careers in "primary care, rural practice, and out-of-hospital services" while discouraging specialization in expensive surgery.

RBRV = (TW)(1+RPC)(1+AST)
Resource-Based Relative Value = (Total Work)*(Specialty Practice Cost Index)*(Specialized Training Cost Index)

Specialized training cost is opportunity cost of spending time in residency.

Total work = Time*(Complexity Index) for Pre- + Intra- + Post-service work based on surveys of physicians

If Medicare fees were adjusted to the RBRVS but total spending unchanged ("budget-neutral"), thoracic surgery, opthamology fees would drop >40%. General surgery would drop about 15%.. Internal medicine would rise >30%. Family practice would rise >60%.
Ontario's negotiated fee schedule more uniform relative to RBRV than mean Medicare payment.
Some limitations of RBRVS, which Hsiao recognizes:

RBRV example
Naus, F. Medical Management Institute 1991
Nose fracture CPT 21325
RVU category US SC adj SC RVU
Work RVU 174.1 0.971 169.05
Overhead RVU 120.3 0.874 105.14
Malpractice RVU 20.0 0.457 9.14
Total 314.4 283.33
The views and opinions expressed in this page are strictly those of the page author. The contents of this page have not been reviewed or approved by the University of South Carolina.