Background: In the standard approach to designing clinical trials in nursing research, the primary endpoint and test statistic to be used for the primary analysis are specified before trial initiation. The false positive error rate for the null hypothesis and statistical power to detect the targeted size of treatment effect are also specified. The main purpose of this paper is to introduce the usage of monitoring procedures on emerging data during the course of the trial, such as the group sequential and adaptive designs, to the nursing science study setting.
Method: The article briefly describes the bases which group sequential and adaptive designs rely on as well as their statistical properties and give some practical considerations to ease orientation in both respects. The different approaches to sample size estimation are applied to a non-inferiority trial on the effect of Skin-to-Skin contact after caesarean birth on infant temperature.
Results: Group sequential designs enable interim monitoring while maintaining the integrity of the standard approach. By contrast, adaptive monitoring procedures, seek to provide the flexibility to modify the prespecified design features during the course of the trial at the expense of statistical efficiency. Clinically meaningful effect size different than originally thought and initial budget limitations are the situations where the flexibility of adaptive designs may offset their loss of efficiency.
Discussion: Interim analyses in nursing clinical trials are an important tool developed to balance ethical and financial advantages of coming to conclusions of efficacy and safety before the planned end of the study against the risk of an incorrect conclusion.
© 2011 Blackwell Publishing Ltd.