Evaluation of Pediatric Heart Surgical Programs

Evaluation of heart surgical programs is hard due to unavailable influencing factors, volume-based performance of each hospital, case-mix pattern of each hospital. Here we consider a hiarchical model to address these concerns.

Causal Inference with Interference

In observation studies, interference may occur if our target population has a network structure. Basic idea is to extend traditional framework and assumptions to take neighbors into consideration. We reproduced simulations by Forastiere et al. (2016) to construct propensity score based estimator in a Facebook network.