Causal Inference of Maternal Antidepressant Use and Pregnancy Outcomes Using Propensity Scores for Multi-Level Treatment

• Activity Number: 113
• Type: Topic Contributed
• Date/Time: Monday, August 5, 2013 : 8:30 AM to 10:20 AM
• Sponsor: Mental Health Statistics Section
• Author(s): Chang Yu*+ and Huiyun Wu and William Dupont and Richard C Shelton and Tina Hartert and Edward Mitchel and Jeffrey Horner and Pingsheng Wu and Hui Nian
• Companies: Vanderbilt University and Vanderbilt University and Vanderbilt University and The University of Alabama at Birmingham and Vanderbilt University and Vanderbilt University and Vanderbilt University and Vanderbilt University and Vanderbilt University School of Medicine
• Keywords: causal inference ; propensity score ; multi-level treatment

Abstract

Drug therapy for depression during pregnancy has been of great concern to both pregnant women and their physicians. Hayes and Wu et al (2012) reported associations between antidepressant medications and pregnancy outcomes in a large cohort of 228,876 pregnancies. We re-analyze the data using propensity score methods of Imai and van Dyk (2004), Imbens (2000), and Joffe and Rosenbaum (1999). The focus of our work is to make causal inference in observational studies with multi-level treatment or continuous treatment using these propensity scores. We also develop new propensity score models. We examine the performance of these methods in terms of bias reduction using simulation studies.