Randomize Participants for Multi-Arm Studies

Randomized Controlled Trials (RCTs) are crucial in clinical research. This guide shows two ways to randomly assign participants to different study arms or groups within Avicenna, even though it's not a built-in feature yet.

The Situation

You are running an RCT and need to randomize participants into different arms (e.g., control vs. intervention). Avicenna currently lacks a dedicated randomization feature. You need a workaround to achieve this randomization directly within the platform or by integrating an external process.

Alternative Use Case

The methods described here can also generate a random value as a survey response for other purposes.

What to Do?

Here are two approaches to randomize participants:

Approach 1: Using a Single Study

This approach uses features within one Avicenna study to assign groups. It supports Simple randomization but not complex techniques like Block or Stratified. Learn more about randomization types here.

You can implement this using either survey questions or participant variables.

Option 1.1: Using Survey Questions

1. Create a baseline survey: This survey will capture the participant's assigned group. See How to Create a Baseline Survey.

2. Add a single-answer question: Use a single-answer question to represent the group assignment. Avoid number questions as some numbers might be chosen more often than others intuitively.

3. Configure for two arms:

   • Set question content: Write Choose one.
   • Define answer options: Enter identical text (e.g., A or spaces) for both options to avoid bias. Note which answer ID corresponds to which arm. You can also use image answers.
   • Enable randomization: Turn on Randomize Answer Order for the question.

   

4. Configure for more than two arms:

   • Set question content: Write Choose the option below.
   • Define answer options: Use identical text (e.g., A or spaces) for all options. Remember the ID-to-arm mapping. Image answers are also possible.
   • Enable randomization: Turn on Answer Randomization. Select all answer IDs for Included Answer ID(s) and set Selection Count to 1.

   

5. Label the question: Use a clear label, like Arm.

6. Make it mandatory: Mark the question as Mandatory.

7. (Optional) Add more questions: Include other necessary baseline questions in the survey.

8. Use the group in criteria: Apply the Arm question's response in criteria for activities, triggering logics, or notification templates to customize the experience per arm. Using this question as a placeholder is not recommended as it shows identical text.

9. Participant experience: When participants join, they complete the baseline survey. They will see either two identical options (for two arms) or one randomly selected option (for more than two arms) to determine their group.

   

Option 1.2: Using Participant Variables

1. Define a participant variable: Treat the participant's arm/group as a variable. Define and set this variable when a participant joins.
2. Assign the variable value: You can either:
   • Randomize externally: Perform randomization outside Avicenna (e.g., using a spreadsheet) and set the variable for each participant accordingly.
   • Randomize internally: Use the survey method described in Option 1.1 to determine the group and set the variable based on the response. This has the same limitations regarding randomization techniques.

Limitation of This Approach

The first approach (using one study) works best if your study arms differ only in their activities and notification templates, not in aspects like data sources.

Approach 2: Using Multiple Studies

This approach involves creating a separate study for each arm.

1. Create studies per arm: Set up distinct studies for each arm (e.g., "Control Study", "Intervention Study"). See Create a New Study.
2. Use identical consent forms: Ensure all studies use the same Consent Materials. This prevents participants from noticing differences if they have study IDs or registration URLs.
3. Randomize externally: Assign participants to arms outside Avicenna (e.g., using spreadsheet randomization).
4. Enroll participants: Direct participants to join the study corresponding to their assigned arm, or invite them via email.

Maintainability

Using multiple studies means duplicating the study setup. Common protocols require repetitive implementation. Mid-study changes must be applied across all study copies, though some tools might help streamline this.

Pro Tips

Blinding Considerations:

• Single-blinded studies: Where participants are unaware of their group, the approaches described work well.
• Double-blinded studies: Where researchers also shouldn't know group assignments, use these approaches but restrict access to data revealing group allocation (like baseline survey responses). Manage this using Roles and Permissions.

Choose the approach that best fits your study's complexity and operational workflow.