Application of Propensity Score in Estimating Dose Response: Maternal Antidepressant Use and Adverse Pregnancy Outcomes

• Activity Number: 594
• Type: Invited
• Date/Time: Thursday, August 2, 2012 : 8:30 AM to 10:20 AM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): Pingsheng Wu*+ and Chang Yu and William Dupont and Tina Hartert and Richard Shelton and Ed Mitchel and Jeffrey Horner and Patrick Arbogast
• Companies: Vanderbilt University School of Medicine and Vanderbilt University School of Medicine and Vanderbilt University School of Medicine and Vanderbilt University School of Medicine and The University of Alabama at Birmingham and Vanderbilt University School of Medicine and Vanderbilt University School of Medicine and Kaiser Permanente
• Keywords: propensity score ; multi-level treatment

Abstract

Propensity score methods with a dichotomous treatment have been extensively evaluated and proven to be an effective approach for reducing bias in observational studies, and extension of the propensity score methodology with multi-valued treatments has been developed. However, experience in evaluating propensity score methods with non-dichotomous exposure variables is limited. Medication therapy for depression during pregnancy has been utilized for decades, still, the safety of antidepressant use during pregnancy remains unclear. As a result, physicians and patients make best decisions based on the information available. We propose to apply propensity score methods to estimate the independent effects of antidepressant medications on pregnancy outcomes. Simulation as well as application to address the potential dose response and timing of antidepressant medication use during pregnancy are conducted and reported.