|
COST-EFFECTIVENESS OF STROKE REHABILITATION SETTINGS IN THE VA
Project Number B5-4346R funded
by RR&D.
March 1, 2007 -
February 28, 2009
Stroke is the leading cause of disability and the third leading cause of death in the U.S. In the Department of Veterans Affairs (VA), stroke related diseases consume at least $1 billion annually, or 5% of health care resources, with approximately 15,000 new stroke patients seen each year. A 1999 study suggested a much higher incidence of stroke than previously reported, with 259 new strokes per 100,000 population. The increasing number of older adults and the emergence of new therapies for acute stroke suggest that the number of survivors with significant residual physical, cognitive, and psychological disabilities will continue to increase, with more survivors living with the aftermath of stroke. Simultaneously, concerns over resource consumption in health care generally, and within the VA specifically, have focused increasing attention on finding efficient methods of achieving quality outcomes. Consequently, it is becoming increasingly important to examine the cost-effectiveness of stroke rehabilitation within the VA, and to understand how formal rehabilitation services across different settings contribute to achieving cost-effective improvements in functioning. The VA presently collects admission and discharge FIM scores on all stroke patients treated within the VA through its Functional Status Outcomes Database (FSOD). These data are incorporated into the Integrated Stroke Outcomes Database (ISOD) maintained by the Rehabilitation Outcomes Research Center (RORC), which will be the central data source for this proposed research. This availability of comprehensive functional recovery data across the continuum of care enables better evaluations of the effectiveness and efficiency of VA rehabilitation services.
The objective of this research is to compare the cost-effectiveness of stroke rehabilitation in the VA across different rehabilitation settings. We propose to examine the change in treatment cost per change in Functional Independence Measure (FIM) for stroke patients in four VA rehabilitation settings involving: (1) no post-acute rehabilitation care (2) post-acute rehabilitation in a non-specialized location (3) post-acute rehabilitation in a subacute rehabilitation unit (i.e., typically a nursing home) and (4) post-acute rehabilitation in a specialized acute rehabilitation unit (either alone or in any conjunction with any other combination of post-acute care).[Hoenig, 2000] In accomplishing this objective, we will also document variations in treatment costs and functional outcomes across these settings. Specific research questions include:
(1) What are the differences in treatment costs per patient across rehabilitation settings after controlling for the factors that influence costs?
(2) What are the differences in functional status outcomes across rehabilitation settings after controlling for factors that influence such outcomes?
(3) What are the differences in treatment cost per FIM point across rehabilitation settings after controlling for factors that influence both costs and outcomes
We will use data from the ISOD for our analyses. The ISOD is a census of stroke patients seen in the VA nationwide, so adequate power to detect statistically significant differences is not an issue. We will, however, address the issue of clinical significance of any differences as part of our research. We will use data from the 2004 National Patient Care Database (NPCD), consisting of data on inpatient, outpatient, and extended care VA patients as merged with the ISOD. In addition, we will use the VA Decision Support System (DSS) National Data Extracts (NDEs) for treatment cost data. We will validate these cost data with the VA Health Economics Resource Center’s (HERC) average cost data as recommended by HERC. We will use advanced linear model techniques to control for patient and facility characteristics that influence both costs and outcomes. While such techniques can control for observable, well-measured differences among patients and facilities, any observational design such as ours must also address the issue of selection bias. Selection bias in the present context arises from the decisions that clinicians and patients make in selecting stroke patients for rehabilitation services. To the extent that a facility or setting selects 'tougher' stroke patients for rehabilitation than other facilities or settings, that facility or setting will have comparatively poorer outcomes when factors influencing selection are unobserved or poorly measured. For this reason, we propose to explore the use of instrumental variables techniques along with the more traditional two-stage Heckman approaches to controlling for selection bias in our statistical analyses. In particular, we will explore the use of the differential distance instrumental variable used by McClellan and Newhouse in their seminal work in instrumental variables estimation applied to health care. We are specifically focusing on functional status as the key rehabilitation outcome in our analyses because of the continuing debate in the cost-effectiveness literature over the use of Quality Adjusted Life Years (QALYs) as a global measure of health outcomes. We believe that FIM scores are more appropriate for measuring rehabilitation outcomes, and will provide a more sensitive measure for detecting differences across rehabilitation venues.
No findings/results to report at this time.
Start-up activities.
The results of this research will provide critical guidance to VHA clinicians, planners, and administrators concerning how to achieve maximal functioning for VHA stroke patients within an increasingly constrained budget, thereby addressing the VHA mission to serve America's veterans and to ensure that quality medical care is provided on a timely basis.
The desired outcome of this research is to improve the efficiency and effectiveness of VHA rehabiliation resources by identifying differences in costs and outcomes achieved by different rehabiliation settings within the VHA, and thereby enabling the VHA to adjust its mix of rehabilitation settings in the most cost-effective manner.
|
