Abstract：Causal inference using observational data is vulnerable to the hidden bias. The impact of unmeasured covariates on the intervention effect can be assessed by conducting a sensitivity analysis. A comprehensive framework of sensitivity analyses has been developed for matching designs. Sensitivity parameters are introduced to capture the association between the missing confounder and the exposure or the outcome. Fixing sensitivity parameter values, it is possible to compute the bounds of the p-value of a randomization test on causal effects. We propose a model assisted sensitivity analysis with binary outcomes for the general 1:k matching design, which provides results equivalent to the conventional nonparametric approach in large sample. Our method substantially simplifies the implementation and interpretation of the sensitivity analysis. More importantly, we are able to provide a closed form representation for the set of sensitivity parameters for which the maximum p-values are non-significant. This methodology can be easily extended to matching designs with multilevel treatments. We illustrate our method using a U.S. trauma care database to examine mortality difference between different trauma care centers.

About the speaker：Dr. Bo Lu is professor of Biostatistics at the Ohio State University. He received his Bachelor degree from the Department of Probability and Statistics at Peking University, and PhD degree in Statistics from the University of Pennsylvania. His primary research area is statistical methods for causal inference with real world data, with applications in health services and health outcomes research. Dr. Lu has developed propensity score matching and weighting methods in complex observational studies, including multiple treatment groups, interventions initiated as different time points, and adjustment for both propensity score and survey weights. He is the associate editors for Journal of Statistical Computation and Simulation, and Observational Studies. He has been the senior statistician for the Ohio Medicaid Assessment Survey series since 2008, which is the largest state-level health survey in US.