Which of the following is the PRIMARY reason for using intention-to-treat (ITT) analysis in randomized controlled trials?

Explanation

## Intention-to-Treat (ITT) Analysis in RCTs **Key Point:** ITT analysis is a fundamental principle that maintains the integrity of randomization by analyzing all enrolled participants in their originally assigned groups, regardless of whether they received, completed, or adhered to the assigned intervention. ### Why ITT Preserves Randomization Benefits ```mermaid flowchart TD A[Randomization creates balanced groups]:::outcome --> B[Participants assigned to treatment] B --> C{What happens next?}:::decision C -->|Per-Protocol Analysis| D[Exclude non-compliant participants]:::action C -->|ITT Analysis| E[Analyze ALL in assigned groups]:::action D --> F[Selection bias introduced]:::urgent D --> G[Breaks randomization balance]:::urgent E --> H[Randomization balance preserved]:::outcome E --> I[Real-world effectiveness estimate]:::outcome ``` **High-Yield:** ITT analysis answers the question: "What is the **effect of assignment** to treatment?" (pragmatic question), whereas per-protocol analysis answers: "What is the **effect of receiving** treatment?" (explanatory question). ITT is the gold standard for efficacy/effectiveness trials. ### ITT vs. Per-Protocol Analysis | Aspect | ITT Analysis | Per-Protocol Analysis | |--------|--------------|----------------------| | **Participants analyzed** | All enrolled, in assigned groups | Only those who adhered to protocol | | **Bias risk** | Lower (preserves randomization) | Higher (introduces selection bias) | | **Effect measured** | Assignment effect (real-world) | True treatment effect (explanatory) | | **Sample size** | Larger, includes dropouts | Smaller, excludes non-compliant | | **Use case** | Regulatory approval, pragmatic trials | Mechanistic studies | **Clinical Pearl:** ITT typically shows smaller treatment effects than per-protocol analysis because it includes non-compliant participants and dropouts. This is a **feature, not a bug**—it reflects real-world effectiveness. **Warning:** Excluding non-compliant participants (per-protocol analysis) introduces selection bias and breaks the balance created by randomization. This is a common mistake in trial design.