关于新药临床试验中控制混杂因素的几点意见

目的:通过探讨新药临床试验总结分析阶段几种控制混杂因素的统计方法,强调控制混杂因素对于新药临床试验的重要性.方法:通过实例说明几种方法的用途及注意事项,阐述各种方法控制中心效应、协变量等对主要结果指标的影响.结果与结论:针对资料的特点,选用适合的统计学方法控制混杂因素对主要结果指标分析的影响,尽可能使药物疗效真实的呈现.

Strategies to control confounding factors in clinical trials

Objective:To assess the feasibility of statistical methods in elimination of possible confounding factors in clinical trial settings.Methods:Several suitable statistical methods to control confounding factors,such as central effect,covariate,and so on,were exampled.Results:In general,these confounding factors were associated with the disease under study and with the exposure(of the subjects) to some agents or conditions.Comprehensive data collected by correct statistical strategies were essential to eliminate possible confounding fallaciousness.This includes not only prospectively collected data on therapy use,but also information on access,healthcare utilization and practices,and adherence.Conclusion:A more elaborate statistical analysis rendered better agreement in fully mimicking what outcomes of study drugs accomplished in a clinical trial setting.