Causal Inference Methods in Pharmacoepidemiology

Abstract

This chapter provides an introduction to pharmacoepidemiology, with a special focus on some of the key concepts related to the application of causal inference methods in pharmacoepidemiology, particularly in the time-varying setting. The chapter first describes frequently used study designs in pharmacoepidemiologic studies of treatment effects. The chapter then follows with drug exposures and confounding in time-fixed and time-varying settings. The chapter continues with a range of causal inference methods developed to evaluate treatment effects while appropriately accounting for time-fixed and time-varying confounding, which is followed by select examples of published studies that have applied these methods to evaluate effects of drug treatments in the field of Neuro-Psycho-Pharmacoepidemiology. Finally, the chapter concludes with a discussion on the considerations needed in the applications of these causal inference methods in studies with observational data.