RAND Statistics Seminar Series

Causal Inference for Randomized Trials of Two Active Treatments Subject to Non-compliance

Presented by Jason Roy, PhD, Geisinger Center for Health Research
Thursday April 17, 2008
1:30 P.M. ET
Conference Room 6202
RAND Corporation, Pittsburgh, PA
Please contact Denise Miller if you would like to attend this seminar.

Abstract

In behavioral medicine trials, such as smoking cessation trials, 2 or more active treatments are often compared. Noncompliance by some subjects with their assigned treatment poses a challenge to the data analyst. In particular, the joint distribution of the observed and counterfactual compliance variables cannot be identified, without imposing strong assumptions. However, due to randomization, each marginal compliance distribution can be identified. These marginal distributions impose bounds on the joint distribution. Our approach is to use a copula model to link the two marginal distributions, up to a sensitivity parameter. We then take a principal stratification approach to estimate causal effects. We develop this approach when compliance is either binary (yes/no) or continuous (dose). We apply the methods to data from a smoking cessation trial. This is joint work with Joe Hogan, Yan Ma and Bess Marcus.

Speaker Bio

Jason Roy, PhD, is a Research Investigator at Geisinger Center for Health Research. His primary research interests are in developing statistical methods for estimating intervention effects in studies subject to missing data and selection bias (e.g., confounding). In particular, a focus is on developing methods that will enable the use of longitudinal electronic health record data as a clinical decision-making tool. Dr. Roy has developed statistical methods for multiple outcome data, longitudinal studies with informative dropout, observational studies and clinical trials with non-compliance. His collaborative research has primarily been in health services research and behavioral intervention trials (obesity and smoking cessation research).



Attending a Seminar

Other Locations/Times:
Washington, D.C. Conf. Rm. 4132: 1:30pm ET
Santa Monica, Conf. Rm. 1226/1228: 10:30pm PT

RAND visitors are welcome to attend and must RSVP at least one day prior to the seminar. To ensure your attendance please contact Denise Miller at dmiller@rand.org with your name, company (or university) affiliation, and national citizenship (for security purposes).

For parking and directions to RAND's Santa Monica office, please see: http://www.rand.org/about/locations/santa-monica.html.

For parking and directions to RAND's Pittsburgh office, please see: http://www.rand.org/about/locations/pittsburgh.html.

For further information and to be added to the mailing list contact Denise Miller at dmiller@rand.org.