Methods consideration in the analysis of observational data: a causal inference perspective
View Presentation View Presentation
*Eric Tchetgen Tchetgen, HSPH 


Challenges in the analysis of modern observational studies will be discussed from a causal inference perspective, including confounding bias and selection bias due to loss of follow-up or death. Possible solutions will be briefly discussed including confounding adjustment techniques in complex longitudinal contexts, instrumental variable design and negative control design to account for unobserved confounding and selection bias due to selective mortality.