Medical Applications
This application has been written to overcome a fundamental problem in statistical analysis.
When starting a randomised trial, the number of subjects needed to achieve a desired power is first decided. Once that is done, it is necessary to decide the optimal way to randomise these subjects.
There are a number of approaches that can be taken:
1. SIMPLE RANDOMISATION
This is the simplest form of treatment allocation. In this case, the system randomly selects an intervention or treatment for each subject. This is equivalent to tossing a coin to decide which arm or treatment regime into which the subject should be placed.This can result in a different number of subjects in each arm. To overcome this problem, the following approaches have been devised:
2. PERMUTED-BLOCK RANDOMISATION
In this approach, the subjects are allocated to small blocks with equal numbers in each treatment arm. Blocks of different sizes are then combined to make up the randomisation list.
3. STRATIFIED RANDOMISATION
This uses permuted block randomisation to create treatment arms with equal numbers of subgroups of interest. An example of this would be to generate different lists for each site by characteristics of interest, say the difference in response to a treatment by gender.
4. MINIMISATION RANDOMISATION
Used to overcome the deficiencies of stratified randomisation. It is designed to minimise the imbalance between treatments, taking into account the variables defined as being of interest.
Companies providing these services can charge thousands of dollars for this.
A much simpler and cheaper approach is to use this application.
This application guarantees that equal numbers of subjects will be in each treatment arm. This means that there is no need for permuted block randomisation. If the researcher wishes to stratify the subjects by a characteristic they wish to study, they can simply enter the characteristics of interest into the subgroups.
Subgroups can be features of interest, such as the difference in response by gender, or different locations. If more than five subgroups are needed, the application can be run a number of times.
The output from the application can be viewed on screen and, if found satisfactory, can be printed or saved to a .CSV file for later analysis.
Privacy Policy
The developer does not collect any data from this application.