Abstract
ABSTRACTWhen testing for superiority in a parallel-group setting with a continuous outcome, adjusting for covariates is usually recommended. For this purpose, the analysis of covariance is frequently used, and recently several exact and approximate sample size calculation procedures have been proposed. However, in case of multiple covariates, the planning might pose some practical challenges and pitfalls. Therefore, we propose a method, which allows for blinded re-estimation of the sample size during the course of the trial. Simulations confirm that the proposed method provides reliable results in many practically relevant situations, and applicability is illustrated by a real-life data example.
Highlights
In clinical trials with a parallel-group design, the main interest is often focused on the question whether a particular intervention is more efficacious than a competitor or placebo
In a recently published clinical trial from stroke research, the National Institutes of Health Stroke Scale (NIHSS) score at 24 hours after the intervention was considered as the primary outcome, and group means were adjusted for NIHSS at baseline (Schönenberger et al 2016)
We propose a blinded sample size recalculation procedure for an ANCOVA model with multiple random covariates, extending the methods that have been examined by Friede and Kieser for the case of a single random covariate (Friede and Kieser 2011)
Summary
In clinical trials with a parallel-group design, the main interest is often focused on the question whether a particular intervention is more efficacious than a competitor or placebo. Sample size calculation in the classical parallel two-group setting was discussed in a very concise and readily understandable way by Borm et al (2007) They exploited some analogies between the t-test and the ANCOVA test statistics, in order to get two approximate formulas, which were based on the classical normal approximation and the Guenther-Schouten adjustment, respectively (Guenther 1981; Schouten 1999). This approach has been criticized by Shieh, who derived an exact method for calculating power and sample sizes for an ANCOVA model with multiple random covariates (Shieh 2017). Sample size formulas, blinded sample size recalculation procedure, and practical considerations
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have