Pricing
Free Trial

Sample Size Determination Example

AREA: Regression | Neurology

nQuery In Use Background Web Opacity 35

Sanofi Multiple Sclerosis Research

Title: Oral teriflunomide for patients with relapsing multiple sclerosis (TOWER):

A randomized, double-blind, placebo-controlled, phase 3 trial.

Objective: Teriflunomide is an oral disease-modifying therapy approved for treatment of relapsing or relapsing–remitting multiple sclerosis. We aimed to provide further evidence for the safety and efficacy of teriflunomide in patients with relapsing multiple sclerosis.

Year: 2014

Source: The Lancet Neurology

Link: http://www.sciencedirect.com/science/article/pii/S1474442213703089

Clinical Area: Neurology

Sample Size Section in Paper/Protocol:

“It was estimated that 370 patients randomly assigned to each treatment group would provide 94% power to detect a 25% relative risk reduction in annualized relapse rate, with the following assumptions: an annualized relapse rate of 0·74 for placebo; the number of relapses follows a Poisson distribution with an over- dispersion parameter of 1·3; a 1·5-year recruitment period with a linear recruitment rate (average exposure duration 1·75 years); and a 20% discontinuation rate.“


Summary of Necessary Parameter Estimates for Sample Size Calculation:

 Parameter Value
 Significance Level (2-Sided) 0.05
 Expected Placebo Rate 0.74
 Expected Rate Ratio 0.75
 Exposure Time 1.75
Overdispersion Parameter 1.3
Covariate Distribution Binomial
Proportion of Covariate 0.5
Total Sample Size (adjusted for 20% dropout) 592


Note that a binomial covariate proportion of 0.5 is equivalent to equal sample size in both groups.

Email me this example in PDF format

Step 1: 
Select the “Poisson Regression” table from the Select Test Design & Goal window.

This can be done using the radio buttons or alternatively, you can use the search bar at the end of the Select Test Design & Goal window.

Sample Size Calculator Example- nQuery- Example 19- Img 01- Poisson Regression

Step 2:
Enter the parameter values for the power calculation taken from the study protocol

Remember to enter “0” in the R-squared field.

Sample Size Calculator Example- nQuery- Example 19- Img 02- Poisson Regression

Step 3:
Enter 0.5 for “Proportion” in the binomial side table.

Click “Compute” to calculate the coefficient variances.

Then click “Transfer” to transfer them main table

Sample Size Calculator Example- nQuery- Example 19- Img 03- Poisson Regression

Step 4:
Beneath the table, select “Calculate attainable power with the given sample size” from the dropdown menu and click Run.

Sample Size Calculator Example- nQuery- Example 19- Img 04- Poisson Regression

This gives a power of 92.2%. This is different from the stated power of 94%.

However, based on the study protocol of the related TEMSO study and FDA Reviews, the power was originally calculated using a Mann-Whitney U test rather than using the Poisson Regression assumptions appropriate for the methodology used in this study. Using a sample size calculation designed for the actual statistical methodology is usually preferred.

 

Email me this example in PDF format

Free 14 Day Trial
Demo Icon white
No download required
 
LOCATION
GraphPad Software DBA Statistical Solutions, 225 Franklin St FL 26 Boston, MA, 02110-2853, United States

Copyright © Statsols. All Rights Reserved. Do Not Sell or Share My Personal Information. Privacy Policy .