The project launched by the Health Care Financing Administration (HCFA) to convert the practice expense portion of the Medicare Fee Schedule (MFS) from a charge-based system to a resource-based methodology, has entered its second iteration.
First System Fails
The initial system proposed by HCFA in 1997 was a contrived collection of the estimated time and cost of non-physician labor, supplies and equipment consumed in the course of providing specific services and procedures. This methodology was supposed to take effect in January 1998, but collapsed when it became clear there would not be a large enough sample of respondents.
HCFA then focused on collecting data from panels of physicians, allied health professionals and administrators, in an effort to price the direct expenses of each procedure code at the CPT level. This data was used to match similar services provided by different specialties, and to determine the redistribution of practice expense payments to primary care providers. The calculated impact of this method on practice expense payments for neurosurgery would have caused a 25-30 percent decrease.
A coalition of specialties, including neurosurgery, similarly impacted by this proposed change was organized, and with substantial funding from the component specialty societies, began a campaign that eventually led to the Balanced Budget Act (BBA) of 1997. In the BBA, Congress required HCFA to use data on actual expenses rather than estimates; to use generally accepted accounting methods of allocating the expenses; and to consult with physician organizations in the development of a new methodology. To make sure that HCFA obeyed, the BBA required the general accounting office to oversee the process.
Second Effort Better
The second method for assigning practice expenses to the procedure codes was published in June 1998, and has adhered to the requirements of the BBA. It uses data on actual costs by specialty, and attempts to allocate those dollars to the procedure codes by using a composite allocation methodology that, although complex, is about as good as we could expect for the first round of this complicated project. There are a substantial number of soft spots in the data HCFA has used, but they have already agreed to an ongoing process of refinement, which will enable the specialties to contribute their input for improving the quality of data over the next several years.
The impact of this version is notably better than the first model — with a 10 percent reduction in expense payments for all neurosurgical services by the end of the transition period in 2002. However, a number of our high volume procedures, such as lumbar laminectomy codes, will suffer greater reductions due to the extra cut in their practice expense relative value units, combined with the change to a single conversion factor that were imposed for the 1998 fee schedule. When these are added to the cuts in the above method, the drop in payments will be about 25 percent for these codes.
Data Problems
Our comments and complaints about the proposed rule have recently been filed with HCFA. The problems with data quality and possible data manipulation have been detailed, and efforts to moderate the negative impact will continue.
For example, the data we have received from practices that have returned the survey of neurosurgery practice expenses have been extremely useful. They have demonstrated that our actual expenses exceed the amount HCFA used in calculating our total pool of practice expenses by more than $60,000 per neurosurgeon, per year. They also helped our argument that neurosurgeons are using their clinical staff outside of the office with increasing frequency, which demonstrates that the costs of such labor should be reimbursed rather than designated as a Part A Medicare expense paid by the hospital.
Survey will Continue
We plan to continue our survey because this is a data-driven game, and the players with good data generally prevail. At present, we have more than 100 neurosurgeons from nearly 30 practices in the database. Our goal is to have at least 250 neurosurgeons. This sampling should accurately reflect the spectrum of practice size, location and organization.
In this context, we will be able to use the data when we approach HCFA . In addition, this data will provide useful cost management information to our members as part of our Cost Containment Initiative.