|
Activity Number:
|
417
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 9, 2006 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Section on Health Policy Statistics
|
| Abstract - #305819 |
|
Title:
|
Combining Information from Various Data Sources To Improve Analyses of Adjuvant Cancer Therapies
|
|
Author(s):
|
Yulei He*+ and Alan M. Zaslavsky
|
|
Companies:
|
Harvard Medical School and Harvard Medical School
|
|
Address:
|
Department of Health Care Policy, Boston, MA, 02115,
|
|
Keywords:
|
administrative records ; hierarchical Bayesian model ; measurement error ; missing data ; multiple imputation
|
|
Abstract:
|
Cancer registry records and administrative systems contain valuable data on provision of adjuvant therapies for cancer patients. Previous studies, however, have shown that these therapies are underreported in those systems. The Cancer Care Outcomes Research and Surveillance Consortium, a multi-center study on treatment and outcomes of colorectal and lung cancer patients, collects treatment data from various sources, e.g. a patient survey and medical records. We propose statistical strategies to combine information from the registry data and patient survey (both subject to undercoverage) and medical records data. Our hierarchical Bayesian models jointly model provision of multiple cancer therapies (e.g. adjuvant chemotherapy and radiation therapy) and reporting in various data sources. Multiple imputations for the true therapy status are hence created to facilitate improved analyses.
|
