Hospice provider characteristics and the volume and intensity of nursing home hospice use

Objective: Given much variation in nursing home (NH) hospice use and concerns about growth in use, this study aimed to understand the association between hospice provider characteristics and NH hospice use.

Methods: We merged resident assessment data (MDS) with Medicare claims to identify the proportion of a NH's year 2000 decedents using hospice and hospice care at higher (and costlier) levels (i.e., continuous home or general inpatient care). US NHs with any hospice were included (N= 11,709). We linked NH-level data with OSCAR data, hospice provider of service data and census data. Multivariate linear regression with clustering on counties and state fixed effects examined NH hospice use rates and a multinominal logit model examined whether the proportion of hospice days at higher levels of care was above the median of 0, or above the top decile of 10%.

Results: In NHs contracting with 2-3 or 4+ hospices (versus 1), hospice use was significantly more (coefficients of 6.48 [95% CI 5.57, 7.39] and 13.13 [95% CI 11.07, 15.20] respectively); and, these NHs also had significantly greater odds of using costlier hospice care. Additionally, when a NH's most used hospice was for-profit (versus non-profit) or free-standing (versus other) hospice use was significantly greater as were the odds of receiving costlier hospice care.

Implications: While contracting with more hospice providers may increase appropriate access, use was differential by profit status. These and longitudinal findings (to be presented) on provider characteristics and hospice use inform the policy debate on needed Medicare hospice payment reform.