OE Journal

ORIGINAL ANALYSIS

Why Do Researchers Fuss about “Intention-to-treat” Analysis?

Randomized trials rely on balanced groups, and that balance hinges on randomization. When participants cross over, deviate from treatment, or drop out, that balance becomes vulnerable, making results harder to interpret. Two analytic approaches often appear in trial reports: intention-to-treat, which keeps participants in their original groups regardless of adherence, and per-protocol, which analyzes only those who followed the plan. Per-protocol can appear appealing but risks bias by excluding those who deviated for reasons tied to prognosis or adverse effects. Intention-to-treat preserves the benefits of randomization and is essential for unbiased estimates, though it cannot solve the separate challenge of missing outcomes when participants are lost to follow-up.