Related Experiment Video
Updated: May 23, 2026
10:46
A Method of Trigonometric Modelling of Seasonal Variation Demonstrated with Multiple Sclerosis Relapse Data
Published on: December 9, 2015
MRI-based clinical trials in relapsing-remitting MS: new sample size calculations based on a longitudinal model.
R M Altman1, A J Petkau, D Vrecko
1Department of Statistics and Actuarial Science, Simon Fraser University, Burnaby, BC, Canada. rachelm@sfu.ca
Multiple Sclerosis (Houndmills, Basingstoke, England)
|April 13, 2012
Summary
Magnetic resonance imaging (MRI) trials for multiple sclerosis (MS) often use the negative binomial (NB) model. This study shows a more realistic longitudinal model provides better sample size estimates for MS clinical trials.
Area of Science:
- Neurology
- Biostatistics
- Clinical Trials
Background:
- Magnetic resonance imaging (MRI) is crucial for multiple sclerosis (MS) clinical trials.
- Current trial sample size calculations often rely on the negative binomial (NB) model for lesion counts.
Purpose of the Study:
- To evaluate the suitability of the NB model for MS lesion count data.
- To propose accurate sample size calculations for relapsing-remitting MS trials using a novel longitudinal model.
Main Methods:
- Assessed NB model fit in five MS clinical trials using Pearson's chi-squared statistic.
- Simulated data from a new longitudinal model to estimate sample sizes for treatment effect tests.
Main Results:
- The NB model showed poor fit (p < 0.05) in at least one arm of four out of five trials.
- Trials designed with the NB model may be under-powered if data deviate from its assumptions.
Conclusions:
- A longitudinal model offers more realistic sample size estimations for MS clinical trials.
- Sample sizes derived from the longitudinal model are frequently smaller than those from the NB model.
