Conference Program

Missing Data Introduced by Applying Hypothetical Strategies: Are They Really Missing? (303723)

View Presentation

*Woojung Bae, University of Florida 
Michael Joseph Daniels, University of Florida 

Keywords: Estimand; Hypothetical Strategies; Intercurrent events; Missing data; Sensitivity Analysis

The recently published ICH E9(R1) guidelines distinguish intercurrent events (ICEs) and missing data when defining estimands. Furthermore, the guideline proposed five strategies for addressing ICEs. The hypothetical strategies (HSs) envisage a scenario in which the ICE would not occur. Subsequently, data observed after the ICE are excluded from the analysis when applying an HS and become de facto "missing." In this paper, we conduct simulation studies to explore whether missing data mechanism assumptions (e.g., MAR) can be readily made for the "missing" data introduced by applying HSs for addressing ICEs, and propose a new missing data analysis method for handling HS-introduced missing data. We provide real applications to demonstrate the proposed method.