Skip to content

goseberg/samplesizr_app

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

18 Commits
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Hello! Welcome to the samplesizr web application.


Introduction

While planning a medical experiment or a clinical trial, the calculation of the sample size is essential. This web application gives you a fast first access to calculate the required sample size for your project. On the panel above, you can choose the statistical test you plan to perform. Choose the input parameters and press 'Calculate!' to perform the sample size calculation.

For an introduction to sample size calcuation for clinical trials and to understand the theory behind this application the following book may be helpful:

[1] M.Kieser: Fallzahlberechnung in der medizinischen Forschung [2018], 1th Edition, Springer Verlag.


Example

Imagine a clinical trial with an intervention group and a control group. The primary endpoint is the quality of life for cancer patients. The QLQ-C30 questionnaire is used to operationalize this endpoint. For a new intervention you expect an improval in quality of life as compared to the control. A difference of 10 points on the scale is needed to consider the new intervention clinically relevant 'better'. To achieve a decision in favour of the new intervention, given the intervention is really 'better' by this amount, a probability of at least 0.9 (power) is desired. You want to tolerate a maximal type I error rate of 0.05 (two-sided) / 0.025 (one-sided). You know that the standard deviation will be 20 points.

The Results:

First the Input parameters will be reported:

Significance level : 0.050
Desired power : 90.00 %
Effect size : 10.00
Standard deviation : 20.00
Allocation : 1.00.

Allocation is defined as the ratio (n intervention group) : (n control group). The resulting sample size and the actual power using this sample size is reported:

n control group : 85
n intervention group : 85
n total : 170
Actual power : 90.31373 %.

R package

This application developed on basis of the calculations presented by M. Kieser [1] and uses the functions provided by the R package samplesizr. To install the R package follow these steps.

Install and load the R package devtools:

install.packages(devtools)
library(devtools)

Install the samplesizr package from github.com:

install_github('imbi-heidelberg/samplesizr')

Load the package and look up the documentation for an overview:

library(samplesizr)
?samplesizr

The R package includes the functions needed to calculate power and sample-size for each test. The power function for ANCOVA is not included.


About samplesizr

samplesizr is a web application based on the R package samplesizr. Special thanks to Kevin Kunzmann for his supervision, patience and help whilst teaching me the tools which made our work on this project very efficient and productive. Special thanks to Professor Meinhard Kieser for his inspiration to this project and teaching of the theory behind sample size calculation.

A click on the logo will lead you to the webpage of our Institute:

IMBI Heidelberg

Daniel Goseberg,
IMBI Heidelberg,
March 16th in Heidelberg

About

Sample size calculation web app

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages