Replacing randomisation lists mid-trial
✋ This feature is available only for Red Pill version 28 and later.
Adaptive and master protocol type trials (such as basket trials, umbrella trials, platform trials, multi-arm, multi-stage trials) have designs that allow the treatment arms and allocation ratios to change during the trial. The uploads page now allows the randomisation list to be replaced even when the trial has started and some rows have been used. This makes it easier to use adaptive designs in Red Pill randomisation systems.
Imagine a trial is using the blocked randomisation list shown in Figure 1. There are two treatment arms (Active and Placebo) in a 1:1 ratio. The list is stratified by Gender and site.
Suppose the trial has started and a few randomisations have taken place, so some rows in this list have been used. We now wish to add a new treatment group at double the active dose, so the statistician prepares a new list as shown in Figure 2. The existing Active treatment group is now called Active 10mg and the new group is called Active 20mg.
The new list is uploaded with the option Delete unused rows in existing list selected. A reason must be given for replacing the existing list.
The preview page shows the new list and a summary of changes and any warnings. In this example we expected the treatment group Active to be missing from the new list and the new Active 10mg and Active 20mg groups to be added.
Once the upload is confirmed all unused rows in the original randomisation list will be deleted and the new list will be appended. This means the new list will be used for the next and subsequent randomisations. Note that if this randomisation system was blinded, you would need to make sure that the code list contained kit codes for the new Active 10mg and Active 20mg groups. It may be necessary to upload a new code list to achieve this.
It's not possible to change the stratification variables by uploading a new list. The list you upload must have the same stratification factors as the original list.
Behind the scenes
The current randomisation list can be downloaded as part of the archive. Inspecting the randomisation list CSV file in the archive for the example above (Figure 5), shows that the first four randomisations used the rows from the original list (uploadId = 1). These used rows have been retained and the new list has been appended (uploadId = 2). The next randomisation to be carried out will use this new part of the list.
The original full randomisation lists can be viewed or downloaded from the 'Past uploads' section of the uploads page.