대응 짝 비열등성 임상시험에서 모의실험을 통한 ITT군과 PP군의 비교

Abstract

Objectives: This research tries to evaluate proper analysis sets for matched-pair noninferiority clinical trials based on Sanchez and Chen’s work (2006). Methods: We evaluate intention-to-treat (ITT) with last observation carried forward (LOCF), ITT with mixed model for repeated measures (MMRM) and per-protocol (PP) using simulation study wit various scenarios. Results: From the simulation study, it is found that LOCF is simple and convenient to handle missing data, but it should be used with caution because it can result in biased effect estimation and lack of control on Type I error. On the other hand, ITT-MMRM has more control over Type I error and bias compared with ITT-LOCF and PP. Conclusions: In matched-pair noninferiority clinical trials, ITT with consideration of MMRM can be recommended for the primary analysis.