Changyu Shen and Xiaochun Li propose new method for sensitivity analysis
New method proposed for sensitivity analysis to assess biases due to unmeasured confounding
In the analysis of observational data, biases often arise due to lack of accommodation of unmeasured confounders. To assess the impact of these hidden biases, investigators often conduct sensitivity analysis. Changyu Shen and Xiaochun Li, in collaboration with Martin Were at Regenstrief Institute, and Lingling Li at Harvard University proposed a new propensity score based sensitivity function to quantify the hidden biases. They published their work in the American Journal of Epidemiology and the Biometrical Journal. Related work was also presented at the Midwest Biopharmaceutical Statistics Workshop in Muncie Indiana, at the invitation of the meeting organizer.
Li L, Shen C, Wu AC and Li X. Propensity Score-based Sensitivity Analyses for Unmeasured Confounding. American Journal of Epidemiology. 2011. 174(3), 345-353.
Shen C, Li X, Li L, and Were M. Sensitivity analysis for causal inference using inverse probability weighting. Biometrical Journal. 2011. 53(5), 822-837.