多组比较样本含量及检验效能的线性算图估计

目的 为临床试验中多处理组比较时样本例数估计提供一种简单、实用的方法。方法 利用线性算图方法,通过计算“差别参数”,得出在一定检验水准和效能下,多处理组比较时每组所需样本含量。结果 实例分析表明,该法所得结果与其他方法结果相近,但计算过程简单且可用于其他类型资料分析的样本例数估计。结论 线性算图法是确定多处理样本含量的一种有效、可行的方法。

The Linear Nomogram Method for Sample Size and Power Estimation of Comparing Two or More Treatment Groups

Objective To provide a simple practical method for determining sample size when comparing two or more treatments in clinical trials.Methods Using the Linear nomogram method,through calculating the "difference parameter" get the needed sample size per group in ANOVA analysis,while the power and significance level of be specified before.Results It is proved that it's results are close to those of other methods with the simple calculation and wide usage.Conclusion The linear nomogram method is a valid and practical technique to determine the sample size when there are more treatment groups.