JMP as a tool for exploring and analyzing simulation results
Naoya Niimura, Chief, Data Science Department, Drug Development Division, Tanabe Mitsubishi Pharma Corporation; Yuichiro Kato, Nobuhiro Nakanishi, Data Science Department, Drug Development Division, Tanabe Mitsubishi Pharma Corporation
In drug development, the role of trial statisticians is crucial, particularly in designing trial designs, including sample size determination. However, with the increasing amount of information to consider, such as various sensitivity analyses, and the diversification of trial designs, exemplified by interim analysis/adaptive design, designing trial designs solely through numerical calculations has become increasingly difficult. For example, in sample size determination, the sample size determination function implemented in JMP is insufficient when it is necessary to determine the simultaneous detection power of multiple evaluation indices. While approaches combining simulations with theoretical approaches based on numerical calculations are frequently employed, the results of these simulations can sometimes be difficult for trial administrators who are not statistical specialists to understand. When presenting simulation results to the aforementioned administrators or decision-makers at their institutions, JMP's ease of use and intuitive visualization are extremely useful. This paper explores how JMP's visualization function can be utilized to aid in decision-making during trial design considerations. This presentation will demonstrate the simulation results and the process of analyzing those results, using sample size determination as an example.
