|
Activity Number:
|
518
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 5, 2009 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Statistics in Epidemiology
|
| Abstract - #304289 |
|
Title:
|
Causal Inference in Trials with Partial Compliance by Means of Continuous Principal Strata: A Likelihood Approach Based on Copulas
|
|
Author(s):
|
Francesco Bartolucci and Leonardo Grilli*+
|
|
Companies:
|
University of Perugia and University of Florence
|
|
Address:
|
Viale Morgagni 59, Firenze, International, 51011, Italy
|
|
Keywords:
|
EM algorithm ; principal causal effects ; profile likelihood ; randomized clinical trial
|
|
Abstract:
|
We propose a flexible approach based on copulas for drawing causal inferences in the principal stratification framework when the strata are continuous. Indeed, partial compliance implies continuous principal strata defined by the levels of placebo and drug compliances. In order to reduce the modeling assumptions, we link the observed marginal distributions of drug and placebo compliances via a copula, so their association is modeled in a flexible way through a single parameter without imposing any restriction on the marginal distributions. The association parameter is then studied via profile likelihood. We apply this method to data previously analyzed by Efron and Feldman (1991) and Jin and Rubin (2008), comparing the assumptions and the results.
|
