The new Medicare prospective payment system for skilled nursing facilities: Impacts on payments, services and quality of care

In 1998 the Centers for Medicare and Medicaid Services (CMS) began phasing in a new prospective payment system (PPS) for Medicare payments to skilled nursing facilities (SNFs). In Chapter 1, I examine the effects of the new PPS on the level of rehabilitation therapy provided in SNFs. Among residents of freestanding SNFs, the fraction receiving extremely high levels of rehabilitation therapy dropped significantly, and the fraction receiving moderate levels of rehabilitation therapy increased. Freestanding SNFs, particularly for-profits, dramatically altered the services they provided in response to new financial incentives. This responsiveness underscores the importance of efforts currently underway to refine the SNF PPS. In Chapter 2, I measure SNF market entry and exit and days of care provided to assess the adequacy of Medicare payments under the new PPS. Following implementation of the PPS in 1998 there was substantial net exit by hospital-based SNFs, but net entry by freestanding SNFs. Among freestanding SNFs, for-profits were no more likely than non-profits to exit. Among hospital-based SNFs, for-profits and newly opened SNFs were most likely to exit. SNFs with a pharmaceutical-heavy casemix were more likely to exit. Payments to freestanding SNFs appear to be adequate, but payments may not be adequate for hospital-based SNFs. In Chapter 3, I examine two effects of the SNF PPS on payments: payment levels were changed (upward for most SNFs) and payments were no longer based on costs incurred. Using Medicare administrative data for all freestanding SNFs in the U.S., I measure facility-level changes in nurse staffing from 1997 (pre-PPS) to 2001 (post-PPS), and relate these staffing changes to payment changes. I find that the changes in payment levels are associated with positive but small changes in nurse staffing, whereas the elimination of cost reimbursement is associated with a large drop in nurse staffing. I also find that the elimination of cost reimbursement is associated with worsening of process-based quality of care measures, but I do not find associations between payment changes and outcome-based quality of care measures. I discuss alternative approaches to introducing elements of cost reimbursement into the existing SNF PPS.