|
Activity Number:
|
238
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Tuesday, August 8, 2006 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Section on Health Policy Statistics
|
| Abstract - #305698 |
|
Title:
|
Methods for Profiling the Value of Hospital Care Following Acute Myocardial Infarction
|
|
Author(s):
|
Justin Timbie*+
|
|
Companies:
|
Harvard University
|
|
Address:
|
180 Longwood Ave., Suite 301, Boston, MA, 02115,
|
|
Keywords:
|
profiling ; Bayesian methods ; cost-effectiveness ; hierarchical models
|
|
Abstract:
|
This paper compares three methods to assess the relationship between in-hospital survival and costs for hospital care following acute myocardial infarction (AMI). The first approach models the two outcomes assuming independence. The second method uses a Bayesian hierarchical cost-effectiveness model in which survival and spending for each hospital are modeled jointly. Incremental costs and survival probabilities are estimated by comparing the performance of each hospital to a standard---a hospital with matching case mix. In the third approach, survival is modeled as a function of spending per episode using a hierarchical linear model with random intercepts and slopes. The top decile of hospitals according to each method are selected for comparison. The advantages and disadvantages of each modeling approach are discussed in an attempt to identify a preferred method.
|
