B1590
Title: Identification and estimation of treatment effects on long-term outcomes in clinical trials with an external data
Authors: Peng Wu - Beijing Technology and Business University (China) [presenting]
Abstract: Motivated by a kidney disease study, the drug effects are investigated on long-term outcomes by combining an RCT without long-term outcomes and an observational study in which the long-term outcome is observed, but unmeasured confounding may exist. Under a mean exchangeability assumption weaker than the previous literature, the average treatment effects are identified in the RCT and derive the associated efficient influence function and semiparametric efficiency bound. Furthermore, a locally efficient doubly robust estimator and an inverse probability weighted (IPW) estimator are proposed. The former attains the semiparametric efficiency bound if all the working models are correctly specified, which may be hard to achieve due to the intertwined working models. While the latter has a simpler form and requires much fewer model specifications. The IPW estimator using estimated propensity scores is more efficient than true propensity scores and achieves the semiparametric efficient bound in the case of discrete covariates and surrogates with finite support. Both estimators are shown to be consistent and asymptotically normally distributed.
