Definition—Using ITT, study participants are analyzed as part of the group to which they were originally randomized, regardless of whether they actually received or completed the intervention. For example, a participant randomized to Treatment A is counted in this group for statistical analysis even if they died, dropped out, crossed over to Treatment B, or didn’t adhere to Treatment A. “Once randomized, always analyzed”
Why use ITT? This method avoids overestimation of treatment effects that might occur if non-adherent participants or drop outs were eliminated from the statistical analysis. In the real world, patients choose not to follow through or may not take a medicine as directed and ITT allows a closer or more realistic estimate of the effect of choosing a particular treatment. ITT allows for preserved statistical power because sample sizes are not reduced as participants drop out. It also can make results more generalizable to a real patient population.
Are there any downsides to ITT? Because ITT includes all participants, if dropouts or protocol deviations are numerous, outcomes may be diluted, conservative estimates of the treatment effects. Essentially, a participant who didn’t take their study medication is not a good representative of what the treatment really does (you might hear a pharmaceutical rep citing “per protocol analysis” statistics instead of ITT because “it gives a better example of the drug’s true efficacy”). ITT minimizes statistical differences between groups but in doing so reduces the chance of type I error.
Reference: Gupta SK. Intention-to-treat concept: a review. Perspect Clin Res. 2011 Jul-Sep; 2(3): 109–112.