Unlike TEFRA, no automatic exception is granted to small rural hospitals. Be cause of the new reporting requirements and needed access to computer services, these hospitals should be allowed additional lead time to adjust to the new plan, if not exempted entirely. Payment of average rates may also be especially harmful to small hospitals. The reason lies partly in the fact that most DRGs often contain only a handful of patients. In New Jersey, where all patients are reimbursed DRG rates, and most hospitals have over 250 beds, approximately 100 of the 393 DRGs with acceptably defined clinical characteristics had around 500 patients statewide, or about 6 patients per hospital. About 70 percent had no more than 3,000 patients, or 30 patients per hospital.

Another area of concern relates to the constraint that the plan proposes to place on DRG prices. Page 45 of the proposal states: "the actual level of prices initially will be determined by the constraint that the prospective payment system not increase Medicare outlays over the amount that would be spent were the present TEFRA system of limits continued." The Tax Equity and Fiscal Responsibility Act tightened Medicare reimbursement significantly. To ratchet down from this reduced level at the same time that an entirely new payment plan is inaugurated introduces too many shocks into hospitals over a short time. This is also a dangerous approach because DRGs have not been tested on a national scale. Experiences with essentially 100 New Jersey hospitals suggest that many problems will arise during the startup

year.

ERROR RATES

The proposed plan and an article in a recent HCFA journal (Health Care Financing Review, December 1982) report that the sample of bills in the MEDPAR file are replete with errors. Perhaps 20 to 30 percent of the bills in the MEDPAR file contain errors that may affect DRG assignment.

The Department makes the assumption that errors will to a large extent be self cancelling. In any event, adverse consequences of data errors can be minimized if the rates are initially based on a hospital's actual costs and payment is based on case mix in the rate year rather than case mix in the base year. Use of historical case mix will not allow for case-mix changes occurring between the base and rate years.

Furthermore, the Department's observation that "there is no evidence from New Jersey" that false coding of patients' medical records is a problem needs to be emphasized. Hospitals have been criticized unjustly for taking meticulous steps to code properly. Under per diem reimbursement systems the diagnosis codes did not determine reimbursement. As soon as they do-as can be expected-every legitimate effort will be taken to scrutinize this area of a hospital's operation. The fact should be emphasized that there is no indication in New Jersey of the phenomenon labelled "DRG creep" or coding up to enhance reimbursement.

IMPLEMENTATION TIME

The Department of Health and Human Services has indicated that it could begin implementing its plan on October 1, 1983. Hospitals would be brought into the program based on their cost reporting period.

Because of problems inherent in a uniform startup date, we agree that hospitals should be phased into the program based on their cost reporting periods. However, it is absolutely essential for hospitals to have sufficient lead time, at least 6 months, to learn about the program, adopt the requisite changes necessary to meet the regulations, and to educate staff about what is expected from them under this entirely new system of reimbursement. Widespread confusion will result if the plan is imple mented too rapidly. The consequences are apt to spill into the future insofar as hospitals are unable to collect information needed for reimbursement.

REPORTING REQUIREMENTS

One of the Department's objectives is to reduce Medicare reporting requirements. New Jersey experiences suggest that the paperwork hospitals face will increase considerably under DRG-based reimbursement. In short, any reduction in Medicare reporting requirements may easily be overpowered by information hospitals must collect and have to operate under the program.

DETAILS OF THE PLAN

There is merit in many of the components in the proposed plan. We are concerned, however, that important details of the Department's proposal are missing.

This limits the ability of hospitals to comment on the proposal and to gauge its financial impact. The following are indicative of some of the missing details:

This plan does not describe the adjustment methodology that will be used to pay teaching hospitals the same DRG rates as other hospitals.

The plan does not indicate whether mean or median costs per Medicare discharge will be used to establish DRG prices. As I stated earlier, median rates would be about 7 percent lower than mean rates.

The plan does not explain how the rates will be adjusted for technologic developments occurring between the base and rate years.

The plan does not indicate how or how often the DRG prices will be calculated after the first year.

The plan does not explain how base year costs will be rolled forward to establish future rates.

The plan does not explain how outliers will be identified and reimbursed, an omission that is especially important for small hospitals.

The plan does not explain how hospitals will be paid for patients who were admitted in one year and discharged the following year.

The plan does not indicate what type of exceptions and adjustments would be granted to sole community providers.

The plan does not indicate whether an annual technology allowance will be added to the rates.

Apart from missing details, in at least two places the plan seems contradictory. These problems may have important financial implications. First, page IV states that "all patients can be categorized into one of 467 different groups," while page 43 indicates that "the category definitions cover virtually the entire patient population." The latter statement is correct. As presently constructed, patients with an operating room procedure unrelated to the principal diagnosis are assigned to DRG 468, one of three catchall DRGs. About 5 percent of the MEDPAR file falls into DRG 468. The proposal must address how these patients will be reimbursed.

Second, rate calculation schemata on page 81 indicate that 1981 costs adjusted for inflation and other factors will be used to set 1984 rates. Yet, according to page 45 medicare payments will be limited to "the amount that would be spent were the present TEFRA system of limits continued." These two approaches may yield significantly different payments.

In conclusion, Mr. Chairman, I hope these comments will prove useful to this committee as it considers this proposal and others such as the American Hospital Association's prospective plan. We commend these efforts and these hearings as an attempt to develop an equitable, effective reimbursement system. We are proud of our accomplishments in New Jersey in serving as a "laboratory" for the nation in testing new systems. We reiterate our sincere offer to share what we have learned and stand ready in this capacity to assist this committee in any way we can.

I would be pleased to answer any questions you may have

STATEMENT OF STANLEY BREZENOFF, President of the NEW YORK CITY HEALTH AND HOSPITALS CORP.

We are pleased to have this opportunity to comment on the Administration's proposed system for prospective payment of hospitals under Medicare. As public providers, we view this as a particularly critical time in the evoluation of our national system of health care financing. The implementation of a well-developed prospective system of payment under Medicare would be a significant step forward in the effort to restrain health care costs while maintaining access to health care services.

In my statement for the record, I will describe how the Administration's proposed payment system as currently drafted would affect the New York City Health and Hospitals Corporation, the largest municipal hospital system in the country. I will also outline several measures which we believe must be included to insure the viability of our hospitals and our ability to provide for our patients, should a national prospective payment system be enacted. They are measures which, to varying de grees, underly the New York State system of payment which began evolving toward its current form in 1969. As you know, the State is currently operating under a three-year waiver from the Health Care Financing Administration which permits it to operate a prospective system of reimbursement under Medicare. In addition, the State reimburses prospectively under Medicaid and Blue Cross, thus including all major payors in our system.

Let me begin by briefly outlining for the record the exact nature of the Corporation that I represent as President. As I noted above, we operate the largest municipal hospital system in the country. It is comprised of 12 acute and four long-term care facilities, 36 community clinics and neighborhood family care centers and the emergency medical services system for the City of New York. We operate on a budget of $1.6 billion. Nearly one third of a billion dollars ($329 million) is funded by the New York City tax levy, $283 million by Medicare, and $702 million by Medicaid which in addition requires a 25 percent contribution by the City.

We provided over 3.2 million inpatient days of care in our facilities last year, and 4.1 million outpatient visits. HHC provides care to all patients regardless of their ability to pay; the overwhelming majority come from poor and low income areas of the city. The poor socio-economic status of our patients has a profound impact on their health status, which in turn influences the services we are required to provide.

EFFECT OF THE PROPOSED PAYMENT SYSTEM

It is this latter issue-the special needs of poor and low income patients for carewhich is of particular concern to us, as we evaluate the potential effect of the Administration's proposal on our Corporation.

Specifically, we are concerned with the method of classifying patients into diag. nostic related groups (DRGs) as a basis for determining rates of reimbursement. Research undertaken at John Hopkins University and elsewhere on this issue indicates that DRGs do not, contrary to the intent of that approach, produce homogeneous groupings of patients with respect to severity of illness. As a result, under a DRG-based reimbursement system, inner city hospitals serving more seriously ill patients within individual DRGs would be inadequately reimbursed for the care they provide.

Recently, our Corporation commissioned two separate studies* to determine how the population we serve influences the amount and type of services we deliver. The first study was specifically conducted to determine the potential impact of a DRG system on our hospitals vis-a-vis private facilities. The second study evaluated the medical needs of patients who were found by our State Department of Social Services to have excessive lengths of stay.

We are submitting copies of our studies for your staff to review. In addition we are providing a summary of these documents and related correspondence to HCFA which we hope may be included with our statement as part of the record.

The findings of both studies are extremely persuasive in documenting:

First, that in general, a greater intensity of service is in fact required by a disproportionate number of our patients, in comparison with patients served by private facilities; and

Second, that the DRG system is seriously deficient in taking into account the factors which are measures of the higher levels of care provided. In particular, it was

"The Impact of Case Mix Measures on HHC Hospitals," by Jeffrey Merrill and Michael Schwartz, and Bellevue Hospital Center-1982 Length of Stay Appeal.