Jarosek, Stephanie L.2015-04-132015-04-132013-12https://hdl.handle.net/11299/171230University of Minnesota Ph.D. dissertation. December 2013. Major: Health Services Research, Policy and Administration. Advisor: Dr. Beth A. Virnig. 1 computer file (PDF); viii, 98 pages, appendices A-C.The Medicare Hospice Benefit (MHB) provides terminally ill Medicare beneficiaries with a program of care specifically targeted to shifting the goals of medical care from curative to palliative care. The American Society of Clinical Oncologists (ASCO) considers hospice to be the optimal system of care for patients with cancer who are dying.This research evaluates use of the MHB among patients with cancer using the Surveillance, Epidemiology and End Results (SEER)-Medicare linked database. We use novel methodologies and data fields which have only recently become available to understand whether these patients are using the MHB as expected given its design. The Peters-Belson method is used to understand the disparity in hospice use between blacks and whites over time in patients with terminal cancer diagnoses (pancreatic, lung and advanced gastric cancers). We find that while use of hospice has grown for both groups, the observed disparity has increased from 1992 to 2008 while the difference between the observed and expected disparity (based on differences in population factors between the groups) has increased over time from <-1% to 33%. We also find that patients with lung cancer who use no cancer directed treatment use hospice at the same rate as patients who use three or more cancer directed treatments (adjusted odds ratio (OR) of 0.892 (95% CI: 0.775-1.026)). Patients who use two or more treatments are significantly less likely than patients with three or more treatments to use hospice (OR=0.838, 95% CI: 0.728-0.965). Finally, we find that 30% of hospice patients with cancer do not die at home, despite the home-based design of the benefit. Almost half of hospice users who died in the hospital had a total length of stay of 3 days or fewer, and two-thirds had a length of stay less than 1 week, suggesting a shift to the hospice payment system without time to benefit from the system of care.Our results demonstrate the continuing importance of monitoring and investigating disparities in hospice use, expanding access to palliative care for patients who use cancer directed treatment, and reforming payment incentives for hospice care provided in institutions.enCancerEnd of life careHospiceHealth services research, policy and administrationThe medicare hospice benefit: peering into the black boxThesis or Dissertation