Artikel
Implementing blinded sample size re-estimation for clinical trials with longitudinal negative binomial counts
Suche in Medline nach
Autoren
-
Thomas Asendorf
- Department of Medical Statistics, University Medical Center Göttingen, Göttingen, Germany
-
Robin Henderson
- Mathematics, Statistics and Physics, Newcastle University, Newcastle upon Tyne, United Kingdom
-
Heinz Schmidli
- Novartis Pharma AG, Basel, Switzerland
-
Tim Friede
- Department of Medical Statistics, University Medical Center Göttingen, Göttingen, Germany
| Veröffentlicht: | 26. Februar 2021 |
|---|
Gliederung
Text
Background: A justification of the sample size required within a clinical trial is a mandatory step in planning. However, the calculated sample size is always subject to a certain degree of uncertainty, as, among other factors, it may depend one or more nuisance parameters, which can be difficult to specify at the planning stage. This is also true for longitudinal negative binomial outcomes, e.g. lesion counts obtained from magnetic resonance imaging in multiple sclerosis [1], in which overall rate and shape parameter can substantially influence the required sample size. To cope with this uncertainty in the case of longitudinal negative binomial counts, blinded sample size re-estimation techniques have been proposed, which recalculate the required sample size based on data gathered up to an interim time point [2], [3].
Methods: Based on a gamma-frailty model by Fiocco et al. [4], we present sample size estimation and re-estimation techniques. Unblinded data is modelled as a mixture distribution of negative binomial counts at interim, thereby maintaining the trials integrity. As the used statistical framework is not standard in common statistical software, these methods are made available within the R-package spass and their usage demonstrated.
Results: The presented methods based on the gamma frailty model from Fiocco et al. [3] are shown to maintain the desired study power without inflation of the type I error rate. The implementation within the R-package spass is explained in detail and demonstrated on an example with longitudinal negative binomial counts.
Conclusion: Blinded sample size re-estimation is a powerful and at the same time easy to implement adaptive procedure in clinical trials. The presented R-package spass allows for fast and efficient blinded sample size re-estimation in clinical trials with longitudinal negative binomial outcomes, even in the case of temporal trends.
The authors declare that they have no competing interests.
The authors declare that an ethics committee vote is not required.
References
- 1.
- European Medicines Agency (EMEA). Guideline on clinical investigation of medicinal products for the treatment of Multiple Sclerosis. 2015.
- 2.
- Asendorf T, Henderson R, Schmidli H, Friede T. Modelling and sample size reestimation for longitudinal count data with incomplete follow up. Statistical Methods in Medical Research. 2019; 28(1): 117-133. DOI: 10.1177/0962280217715664
- 3.
- Asendorf T, Henderson R, Schmidli H, Friede T. Sample size re-estimation for clinical trials with longitudinal negative binomial counts including time trends. Statistics in Medicine. 2019; 38: 1503-1528. DOI: 10.1002/sim.8061
- 4.
- Fiocco M, Putter H, Van Houwelingen JC. A new serially correlated gamma-frailty process for longitudinal count data. Biostatistics. 2009;10(2):245-57.
