Skip to Main content Skip to Navigation
Master Thesis

Calcul du nombre de sujets nécessaires dans les essais cliniques randomisés contrôlés de phase III et IV étudiant les traitements de fond de la sclérose en plaques de forme récurrente-rémittente : une revue systématique

Abstract : Background : In randomized controlled trials (RCTs), the sample size estimation (SSE) is an important step of the planning of statistical analyses. The calculation method must be adapted to the distribution of the studied variable. In multiple sclerosis (MS) studies, the Annualised Relapse Rate (ARR) is a marker of the disease activity, which distribution follow a negative binomial (NB) distribution. Our objectives were to determine whether the SSE and the analysis of ARR in phase 3 and 4 RCTs of drug therapies for relapsing remitting MS were consistent with the distribution of the ARR, and to compare the SSE across several statistical distributions. Methods : A systematic review of the literature included phase 3 and 4 RCTs which primary endpoint was ARR in relapsing remitting MS, published since 2008 in scientific journals with an Impact Factor ≥ 4. We checked whether the SSE and the ARR analyses were congruent with a NB model. We also performed SSE based on normal and NB distribution, using the data collected from the studies. Results : Nineteen articles corresponding to 20 studies were selected. Nine of them indicated the method used for SSE. Of these, only 4 used a model derived from a Poisson or NB distribution. In contrast, the ARR analyses were based on these two models in 19 out of 20 studies. SSE using identical parameters showed highly variable results, depending on the ARR and dispersion used. Conclusions : Our study pointed out that the SSE in MS RCTs studying ARR is rarely performed with the accurate model. The complexity of the NB model creates variable difference compared to the gaussian model. As it is difficult to predict these differences, the use of a NB model seems mandatory to optimize the number of included patients.
Document type :
Master Thesis
Complete list of metadatas

Cited literature [46 references]  Display  Hide  Download

https://dumas.ccsd.cnrs.fr/dumas-02942764
Contributor : Santé Uca <>
Submitted on : Friday, September 18, 2020 - 11:38:51 AM
Last modification on : Wednesday, October 14, 2020 - 3:57:45 AM

File

Thèse PONCET MEGEMONT Louis.p...
Files produced by the author(s)

Identifiers

  • HAL Id : dumas-02942764, version 1

Collections

Citation

Louis Poncet-Megemont. Calcul du nombre de sujets nécessaires dans les essais cliniques randomisés contrôlés de phase III et IV étudiant les traitements de fond de la sclérose en plaques de forme récurrente-rémittente : une revue systématique. Médecine humaine et pathologie. 2020. ⟨dumas-02942764⟩

Share

Metrics

Record views

28

Files downloads

18