MIPE在临床试验转组研究中的应用

The Application of MIPE for Dealing with Treatment Switching in Randomized Controlled Trials

目的本文介绍一种新的处理临床试验转组数据的统计学方法——校正迭代参数法(MIPE),并通过模拟试验在不同情形下比较MIPE、IPE、意向性分析(ITT)三种方法对治疗效果估计的准确性。方法通过模拟数据探讨不同的治疗效应真值、潜在预后、转组率等24个假设场景中试验药疗效估计值的差异,并将MIPE、IPE、ITT方法应用于每个场景下进行效果比较。结果当试验药与安慰剂组疗效存在差异时,ITT和IPE方法得出的治疗效应估计均有偏,这种有偏估计与不同的治疗效应真值、潜在预后和转组率等因素相关,而MIPE方法不受上述因素影响,估计结果较为准确且稳定。结论 MIPE方法不受实际治疗效应、潜在预后、转组率等因素的影响,在各场景中治疗效应估计值的均值较为稳定,与其他方法相比,更接近真实值,在临床试验评价中使用更为可靠。

Objective To introduce a new statistical method for dealing with treatment switching in randomized controlled trials:the modified iterative parameter estimation(MIPE).To compare the MIPE with IPE and intention-to-treat(ITT)methods via extensive simulations to assess the performance in different scenarios.Methods A total of 24 different scenarios are identified which differ by the true treatment effect,proportion of patients switching and underlying prognosis of patients.In order to assess the performance of each method,MIPE,IPE,and ITT are performed in each scenario.Results The treatment effects estimated by the ITT and IPE methods were biased when true effects between new treatment and placebo group were different,and the biased estimation was related to the true treatment effect,the proportion of patients switching,and the underlying prognosis of patients,while MIPE method was not.Conclusion The method of MIPE is not affected by factors such as the true treatment effect,the underlying prognosis,the proportion of patients switching,and so on.The estimation of treatment effect in each scenario is stable and close to the true value so that it is convenient to use in clinical practice.