JMP as a tool for exploring and interpreting simulation results
Mitsubishi Tanabe Pharma Corporation, Data Science Department, Drug Development Division, Manager, Naoya Niimura Mitsubishi Tanabe Pharma Corporation, Data Science Department, Drug Development Division, Yuichiro Kato, Nobuhiro Nakanishi
In the role of a trial statistician in drug development, the planning of trial design, including sample size design, is very important. However, due to the increase in information to be considered, such as various sensitivity analyses, and the diversification of trial designs, such as interim analysis/adaptive design, it is becoming more difficult to plan trial designs using only numerical calculations. For example, in case size design, the sample size design function implemented in JMP alone is not easily sufficient when you want to know the simultaneous detection power of multiple evaluation indices. In addition to theoretical approaches based on numerical calculations, approaches that combine simulation are often used, but the difficulty is that these results are sometimes difficult to understand for trial implementation personnel who are not experts in statistics. When showing simulation results to the above personnel or decision makers at the affiliated institution, the ease of use and intuitive visualization of JMP are very useful. In this presentation, we examined how the visualization function in JMP can be used to help with decision-making in trial design considerations. In this presentation, we will introduce the results of simulation and the process of considering those results, using sample size design as an example, with demonstrations.
