Intention to Treat
Introduction to Intention to Treat
Host: Anthony Croco of Sketchy EBM
Topic: Intention to treat in randomized control trials (RCTs)
Personal anecdote: Bought candy for Halloween but did not give it out—illustrates having intention but not following through.
Overview of Randomized Control Trials (RCTs)
Definition: RCTs involve a randomization process to divide participants into two groups:
Experimental/Intervention Group: Receives the treatment being tested.
Control Group: Receives a placebo or standard treatment.
Objective: To observe the effects of the treatment without bias from group assignment.
Compliance Issues in RCTs
Participants may deviate from study protocols:
Random reasons (e.g., receiving the wrong intervention).
Specific reasons (e.g., participants stop adhering due to side effects).
Challenges for researchers on how to deal with non-compliant participants' data.
Data Handling Options
Per Protocol: Only includes data from participants compliant with the study protocol.
As Treated: Analyzes data based on the group participants actually received, irrespective of randomization.
Intention to Treat (ITT): Maintains participants in their originally assigned groups, regardless of compliance or treatment received.
Example Analysis
Example with two truths:
No difference exists between groups; 75% of participants have good outcomes.
Scenario with four participants each in control and experimental groups:
Control Group: 3 out of 4 have a good outcome (75%).
Experimental Group: 3 out of 4 have a good outcome, with one non-compliant participant.
Analysis Outcomes
Per Protocol:
Control Group: 75% (3/4 good outcomes).
Experimental Group: 100% (3/3 good outcomes).
As Treated:
Control Group now has 5 participants (including one non-compliant)—3 good outcomes (60%); experimental remains at 100%.
Intention to Treat:
Both groups remain 75% (3/4 good outcomes).
Another Scenario
If the non-compliance participant had a good outcome:
Per Protocol: Control remains 75%; Experimental group now 67%.
As Treated: Control Group: 80% (4/5 good outcomes); Experimental Group: 67%.
Intention to Treat: Maintains both groups at 75% good outcomes.
Importance of Intention to Treat Analysis
Minimizes bias from non-compliance and helps better reflect real-world scenarios.
CONSORT Diagram: Visual representation of participant data throughout the study.
Key for understanding how data was handled in RCTs.
Conclusion
Importance of evaluating randomized control trials for intention to treat analysis.
Encouragement to draw personal conclusions from the analysis presented.