|
Activity Number:
|
202
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, August 7, 2006 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Biometrics Section
|
| Abstract - #305370 |
|
Title:
|
A Random-Effects Four-Part Model for Longitudinal Medical Costs
|
|
Author(s):
|
Lei Liu*+ and Mark R. Conaway
|
|
Companies:
|
University of Virginia and University of Virginia
|
|
Address:
|
Division of Biostatistics, 3181 Hospital, W., Charlottesville, VA, 22908,
|
|
Keywords:
|
generalized linear mixed model ; hierarchical linear model ; four-part model ; Laplace approximation ; random effects ; longitudinal
|
|
Abstract:
|
We extend the four-part model (Duan et al. 1983) in the cross-sectional medical cost setting to longitudinal medical cost setting (e.g., monthly medical costs) with correlated random effects. Our model characterizes four major features of longitudinal medical cost data: a fair amount of monthly medical costs are zero; there is a remarkable difference between inpatient and outpatient monthly costs; monthly medical costs are correlated for the same subject; and the responses in the four-part model also are correlated. We introduce the random-effects four-part model and propose an estimation method by the high-order Laplace approximation technique in Raudenbush et al. (1998) and Olsen and Schafer (2001). Our method is applied to monthly medical costs for heart failure patients from clinical data repository at the University of Virginia Health System.
|
