Despite the rapid growth of hospice in nursing homes (NHs) and its benefits, NH residents constitute a vulnerable population facing access barriers to hospice beyond those experienced by persons in the community. Research shows hospice enrollment is associated with a greater probability of residents having their pain assessed and managed and with a lesser probability of being hospitalized at the end-of-life. However, enrollment in Medicare hospice is only available to NH residents if a NH has a contract with a hospice provider, and in practice, not all NHs have such contracts. Furthermore, low Medicaid reimbursements to NHs may have a negative effect on their decision to offer hospice. In this proposed project we use national data to examine the process by which the Medicare hospice benefit becomes available to terminally ill NH residents. We take advantage of linked Medicare claims, nursing home and hospice provider data and Minimum Data Set (MDS) clinical information. We will describe and characterize at the local, state and national level the penetration of Medicare hospice in nursing homes. Using multilevel modeling techniques, we will examine the effect of facility, market (including providers and population), and state characteristics on the penetration of hospice in nursing homes. Using detailed hospice provider data, we will evaluate the effects of hospice revenue and referral sources on the penetration of hospice in nursing homes in North and South Carolina, adjusting for facility, market, and state characteristics. Finally, we will develop an understanding of the non-market factors associated with hospice penetration in nursing homes when the observed practice is discordant with the penetration as predicted by our multilevel model. We will identify hospice providers in each of the four major U.S. geographic regions (i.e., West, Midwest, South and Northeast) who are providing hospice care in nursing homes when there is a low prediction of such provision (""""""""false negatives"""""""") and, identify hospice providers not providing care in nursing homes when there is a high prediction of such provision (""""""""false positives""""""""). We will conduct case studies of one """"""""false positive"""""""" and one """"""""false negative"""""""" hospice provider in each of the four geographic regions by conducting site visits and interviewing hospice and nursing home staff (and their geographic proximal nursing homes).
