Statistical Methods for Evaluating Clinical Therapeutic Effects
中医临床疗效评价的统计方法

The purpose of evaluating clinical therapeutic effects is to confirm the effectiveness of a therapeutic plan, method, and recipe in treating diseases, in the context of selected patients. In the course of clinical treatment, we may choose some patients groups, and control some influence factors using random grouping. However, the realization of anticipated results has to be determined by a range of mixed factors. As a result, it is difficult to make an evaluation merely using a statistical method with an identical time. Fortunately, we can treat them as decorative factors, i.e., explanation variables, independent variables, or covariates, before making the evaluation using statistical models. As index types that reflect the efficacy of clinical treatment are different, statistical models have to be different too. Logistic model, Cox proportional hazards regression model, longitudinal models, and multilevel linear model are useful statistical approaches for evaluating clinical therapeutic effects.