|
Activity Number:
|
234
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Tuesday, July 31, 2007 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Biometrics Section
|
| Abstract - #309523 |
|
Title:
|
Estimation and Inference for the Causal Effect of Receiving Treatment on a Multinomial Outcome
|
|
Author(s):
|
Jing Cheng*+
|
|
Companies:
|
University of Florida
|
|
Address:
|
1329 SW 16th Street, Gainesville, FL, 32610,
|
|
Keywords:
|
Randomized trials ; noncompliance ; multinomial outcomes ; causal effect ; bootstrap
|
|
Abstract:
|
This paper considers the analysis of two-arm randomized trials with noncompliance which have multinomial outcomes. We define and estimate the compliers average causal effect (CACE) in these trials with the likelihood method. A simulation study shows that the likelihood ratio statistic for testing the CACE does not follow a chi-squared distribution. We propose a bootstrap version of a likelihood ratio test for the CACE in these trials. The methods are illustrated by an analysis of data from a randomized trial of an encouragement intervention to improve adherence to prescribed depression treatments among depressed elderly patients in primary care practices.
|
