Falissard B, Lellouch J
INSERM Unité 169, Villejuif.
Rev Epidemiol Sante Publique. 1991;39(4):365-72.
It is often interesting to introduce, in the plan of a clinical trial, the possibility to make interim analyses. However, this induces an increase of type one error and it is thus necessary to use appropriate statistical methods. Usual group sequential procedures are mentioned, just as more recent approaches like flexible methods with free number and time of analyses, or like the succession procedure.
在临床试验方案中引入进行期中分析的可能性通常很有意思。然而，这会导致一类错误增加，因此有必要使用适当的统计方法。文中提到了常用的序贯分组程序，以及更新的方法，如具有自由分析次数和时间的灵活方法，或连续程序。
