目的 探讨混合线性模型在带有时依协变量的重复测量资料分析中的应用.方法 以治疗轻、中度原发性高血压病临床试验资料为例,考虑到给药方案在各个时间点随病情而变化,利用SAS中的MIXED过程,选择合适的协方差结构来实现带有时依协变量的重复测量资料的统计分析.结果 时依协变量(给药方案)对治疗轻、中度原发性高血压病有统计学意义(P＜0.05)；时间因素有统计学意义(P＜0.05)；给药方案与时间因素之间有交互效应(P＜0.05)、给药方案与处理因素之间有交互效应(P＜0.05).结论 采用混合线性模型对带有时依协变量的临床试验重复测量资料进行统计分析,可以更客观地进行药物疗效评价.%Objective To apply mixed linear model to analyze the repeated measures data with time varying covariate. Methods An example of light,medium primary hypertension clinical trial was given to show the application of mixed linear model considering the dosing at all time points changing with illness. Using the MIXED procedure of the SAS system, appropriate covariance structure was choosed to achieve analysis of repeated measurement data with time varying covariate. Results At 5% level of significance,the interaction effects of covariate * time,covariate * group and the effects of time and covariate were statistically significant to treat light, medium primary hypertension. Conclusion It can get more objective result of medicine effect by using mixed linear model in the repeated measures data of clinical trials with time varying covariate.
展开▼
