|
Activity Number:
|
313
|
|
Type:
|
Invited
|
|
Date/Time:
|
Wednesday, August 6, 2003 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Academy for Health Services Research and Health Policy
|
| Abstract - #301852 |
|
Title:
|
The Effect of Hospice Care on Costs at the End of Life
|
|
Author(s):
|
Pedro Gozalo*+ and Orna Intrator and Jason A. Roy and Susan Miller
|
|
Companies:
|
Brown University and Brown University and Brown University and Brown University
|
|
Address:
|
Center for Gerontology & Health Care Research, Providence, RI, 02912,
|
|
Keywords:
|
causal inference ; time-varying treatment effects ; instrumental variables ; hospice benefit
|
|
Abstract:
|
Our study considers the evaluation of the effect of the Medicare hospice benefit on health care costs for nursing home patients at the end of life. Hospice is an elective Medicare benefit designed for terminal patients (expected to live six months or less at the time of the election) that prefer palliative rather than curative treatments at the end of life. For frail patients in nursing homes, access to hospice treatment is complicated by the willingness of the nursing home to offer that benefit to their residents. Previous hospice cost studies have primarily utilized retrospective cost comparisons between decedents that chose hospice and those that did not, and fail to correct for potential self-selection of low health care users into hospice. We consider a six-month follow-up prospective analysis and focus on total costs to both Medicare and Medicaid insurance programs. The analysis takes into account the nonrandom time-varying treatment induced by the elective nature of hospice treatment and by the preferences of nursing homes for offering hospice. Inverse probability weighting and instrumental variables techniques are combined to correct for selection bias of our estimates.
|
