|
Activity Number:
|
373
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Wednesday, August 9, 2006 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Section on Survey Research Methods
|
| Abstract - #306239 |
|
Title:
|
Approximation of Skewed Health Care Expenditure Distribution Using a Mixture Model
|
|
Author(s):
|
William Yu*+
|
|
Companies:
|
Agency for Healthcare Research and Quality
|
|
Address:
|
540 Gaither Road, Rockville, MD, 20850-6649,
|
|
Keywords:
|
healthcare expenditures ; skewness ; mixture model ; MEPS
|
|
Abstract:
|
The expenditure data from the Medical Expenditure Panel Survey (MEPS) have been shown to exhibit a marked positive skewness, characterized with a few high-expenditure respondents and many zero-expenditure respondents. Any approximation of the distribution of MEPS expenditure will need to capture the bimodality feature of the MEPS expenditure data. A mixture model with a point mass at zero and the positive half of the real line is proposed. This mixture model captures the bimodality feature of the MEPS expenditure distribution and allows for evaluation of the probability that a MEPS respondent has zero expenditure varies significantly with a selected set of variables. The positive values of MEPS expenditure may be assumed to follow one of the Weibull, Gamma, or Log-normal distributions. The mixture model will be estimated and evaluated for goodness-of-fit with 2003 MEPS expenditure data.
|
