This tutorial outlines how to set up the randomization for your study. In our blog we provide a description of randomization in clinical trials and the method used by Castor. Always consult a statistician before setting up your randomization, to ensure proper selection of the block sizes and strata.
Note: To define the randomization settings for the study you need Manage Settings rights.
To enable randomization for your study, set "Randomization" to "Yes" using the drop-down in the Settings tab:
After saving the changes in the settings tab go to the randomization tab to define the randomization settings:
- Define randomization groups. To define groups, use the 'Add group' button. Define a group name and if applicable, choose a weight. The weight indicates the relative proportion of records placed in each group. The standard randomization is 1:1 (weight = 1 for each group) and this should only be changed if you have a specific reason. Save the group once you have defined it. Your defined groups will appear in the 'Randomization groups panel', as shown above. Please note: There is a max of 9 randomization groups.
- Define randomization strata (in case of stratified randomization). To define strata, use the 'Add stratum' button. You will be prompted to locate the field you want to use for stratification. Strata in Castor EDC are based on option group fields (dropdown/radio button), so to add a stratum, you must first create the field you want to stratify based on, that will have the option values that are relevant to your stratification. Only use stratification if you are sure that this is required by your study design. When you select a stratum, it will appear in the 'Stratify by' panel, as shown above for the selected stratum gender.
- Select whether to stratify per institute. This is only applicable for multi-center studies. We recommend setting this to yes, as stratification per institute is usually required for multi-center studies. However, if you are unsure, please discuss it with your project group.
- The only current randomization algorithm is block randomization, so this will be chosen by default.
- Define your block divisions.You can choose different block sizes for your variable block randomization. The options in the dropdown menu will change depending on the amount of randomization groups and the weight of the groups. Read our blog to find out more about randomization blocks.
Save the randomization settings. Once you have saved them, you can test your randomization with test records.
After going live it is no longer possible to change the Randomization settings. If your study requires a change in the randomization settings after going live, please contact us.
Note: The users in your study who will be doing the randomization need to have 'Randomize' (Rand.) rights. If you want them to be able to see the randomization allocation, you need to also assign them 'View randomization' (or View rand.) rights. Thus, if your study is a double-blinded study, the users in your study should have only 'Randomize' rights and thus be able to randomize, but be blinded to the actual allocation. Find out more about defining user rights here.
Randomization when going 'Live'
When you decide to start collecting real data, you will be able to reset your randomization. When you set your study to live for the first time, you will be asked whether you want to reset the randomization. That means that the test records that were randomized will not influence the future randomization of the real records. This will not remove the randomization allocation of the test records, but will clear the randomization blocks, meaning that you can start including and randomizing records with clean randomization blocks. Before you click ‘yes’ please make sure all the data in the study is test data. If you are unsure about this choice or if you have any questions please contact us. When a record has been randomized, it is no longer possible to change the data for a field or its parent if it has been used for stratification.