|
Activity Number:
|
426
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 1, 2007 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Section on Statistics in Epidemiology
|
| Abstract - #308345 |
|
Title:
|
Analysis of Recurrent Event Data in Case-Crossover Studies
|
|
Author(s):
|
Xianghua Luo*+ and Chiung-Yu Huang and Gary S. Sorock
|
|
Companies:
|
The University of Minnesota and National Institutes of Health and Johns Hopkins Bloomberg School of Public Health
|
|
Address:
|
420 Delaware St SE, Minneapolis, MN, 55455,
|
|
Keywords:
|
conditional logistic regression ; incidence rate ratio ; Mantel-Haenszel ; within-subject averaging
|
|
Abstract:
|
The case-crossover design is useful for studying the transient effects of exposures on risk of events when only data on cases are available. While the original design was proposed for univariate failure time data, in many applications recurrent events (e.g., elderly falls and school absences) are often encountered. In such situations, the within-subject dependence among recurrent events needs to be taken into account. Existing approaches rely heavily on independence assumptions for the recurrent events, either through a non-testable Poisson process assumption or a hypothetically complete set of observed covariates. We propose to adjust the Mantel-Haenszel estimator of incidence rate ratio and the conditional logistic regression, using a within-in cluster averaging technique. The proposed methods are applied to a study of the effect of medication changes on falls among the elderly.
|
