Causal Analyses of Electronic Health Record Data for Assessing the Comparative Effectiveness of Treatment Regimens

This study aims to strengthen causal inferences based on Marginal Structural Modeling (MSM) and large healthcare databases for Comparative Effectiveness Research in Type 2 Diabetes Mellitus (T2DM), specifically, it aims to: 1) Develop a consensus for the types and quantity of MSM parameters relevant to T2DM patients and other stakeholders; 2) Evaluate the feasibility and performance of MSM approaches based on Targeted Learning to estimate these MSM parameters more efficiently and with more robustness to modeling assumptions than the current Inverse Probability Weighting approaches; 3) Explore the biases that may arise from current MSM applications with data where patient monitoring is not fixed but random or where the outcome is not right-censored but interval-censored and evaluate innovative MSM methods that aim to address these biases.

0