An EM Algorithm for Mixed-Type, Multiple Outcome Regressions with Applications to a Prostate Cancer Study

JoAnn Alvarez, M.A.

Biostatistician II, Department of Biostatistics, Vanderbilt University School of Medicine

Wednesday, October 7, 1:30-2:30pm, MRBIII Conference Room 1220

Intended Audience: Persons interested in applied statistics, statistical theory, epidemiology, health services research, clinical trials methodology, statistical computing, statistical graphics, R users or potential users

Biomedical studies typically have several endpoints which are sometimes correlated. If the endpoints are a mixture of continuous and binary, they can be modeled separately using linear regression for the continuous variables and probit or logistic regression for the binary ones. However, if the outcomes are moderately or highly correlated, this would not be the most efficient way to model them because it does not take the correlation into account. The EM (expectation-maximization) algorithm can be used to link the two models, thereby simultaneously estimating the regression parameters and the correlation between the responses. I develop an EM algorithm specifically for this problem using a latent normal variable model for the binary variable. In this talk I will describe the model and present the algorithm for estimating the maximum likelihood estimates along with the results of a simulation to examine the performance of our algorithm. I will also show an application to a prostate cancer study.

Topic revision: r2 - 06 Oct 2009, EveAnderson

Main

Department Home Page

Biostatistics Graduate Program

Vanderbilt University Medical Center

Biostatistics Webs
- Archive
- Main
- Sandbox
- System

Copyright &© 2013-2022 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding Vanderbilt Biostatistics Wiki? Send feedback