Introduction to Treatment Effects for Observational Survival-Time Data

This guide provides an introduction to the treatment-effects estimators that utilize observational survival-time data, specifically through the stteffects command. It outlines the potential-outcomes framework and its relevance to survival-time data, detailing how average treatment effects (ATEs), average treatment effects on the treated (ATETs), and potential-outcome means (POMs) are estimated. The document discusses various estimators implemented in stteffects, including regression adjustment (RA), inverse-probability weights (IPW), inverse-probability-weighted regression adjustment (IPWRA), and weighted regression adjustment (WRA). It also presents the assumptions necessary for these estimators and the trade-offs associated with each method. Additionally, the guide addresses the missing-data problem inherent in treatment-effects estimation and provides examples of how to apply these estimators using fictional data on the time to a second heart attack among women aged 45–55. The document concludes with a technical note on limitations regarding left-truncation and time-varying covariates.