The Medicare Prospective Payment System instituted for hospitals in 1983, excluded facilities, such as rehabilitation hospitals, whose patients could not be classified reliably using DRGs. These facilities are paid under an alternative system which has become unsatisfactory. Congress mandated HCFA to develop an alternative payment option. Our pilot project demonstrates the feasibility of creating a patient classification system based on Function Related Groups (FRGs) for medical rehabilitation which is analogous to the DRG system for acute-care patients. The FRGs in their current form use only three categories of explanatory variables: the Rehabilitation Impairment Category, the admission functional status score, and age. The predictive power of our FRG system for lengths of stay in rehabilitation is comparable to that of the DRG system for acute- care patients. The source of data used was the Uniform Data System for Medical Rehabilitation (UDS), which represents the largest information system utilizing standardized functional status scores. UDS contains a nationally representative sample of distinct-part rehabilitation units and rehabilitation hospitals. The principal analytic tool was CART, a technique for recursive-partitioning. We propose to complete work on the FRG patient classification system by examining alternative specifications of explanatory variables, and by measuring the effects of medical comorbidities on length of stay. We also propose to begin work on a payment system based on our patient classification system. That work will measure the effects of atypical patients excluded from the prototype FRGs. It will further determine whether the discharge functional status score might be useful to provide incentives for high-quality care. Successful completion of this project will provide the foundation needed to create a more suitable payment system for facilities that provide care to rehabilitation patients. It also will facilitate research in rehabilitation medicine by providing a patient classification system that can adjust outcome data for admission case mix and, thus, make it possible to compare quality and cost across facilities.
Showing the most recent 10 out of 14 publications
