Standard versus adaptive monitoring procedures: a commentary

摘要: In the standard approach to designing definitive clinical trials, primary endpoint and test statistic be used for analysis are specified before trial initiation. The false positive error rate null hypothesis statistical power detect targeted size of treatment effect also specified. Standard monitoring procedures, such as group sequential guidelines, enable interim while maintaining integrity this approach. contrast, adaptive procedures seek provide flexibility modify these pre-specified design features during course trial. However, have several undesirable properties, including lesser efficiency, reduced interpretability outcome results, basing changes on unreliable estimates efficacy, risks credibility trial, loss use emerging results from external sources alter key features, overemphasis importance significance relative significance.