Intention to Treat

Intention to treat (ITT) involves the statistical analysis of data from randomized clinical trials. At the beginning of their participation, patients are randomized to one of the trial interventions. Over the course of their participation, they may leave the group to which they were assigned – they may stop treatment altogether, start a treatment not tested in the trial, or switch to one of the other trial interventions. This can happen for various reasons, such as experiencing an adverse effect from the treatment to which they were originally assigned or not seeing enough benefit from that treatment.

When it’s time to analyze the data, patients who switched away from their randomly assigned treatment will still be included in their originally assigned study group when calculating treatment effects and performing statistical tests, even though they didn’t get the full intervention – this is called “intention to treat.” The reason is by maintaining the original randomized groups, they will be equivalent on all factors (known and unknown) that affect the outcome. The downside is that those patients who don’t receive the full study intervention may not accurately reflect its effect, but this is outweighed by the need to have the study groups remain comparable, which occurred through randomization. It is also thought to reflect more of the “real world” situation where patients come off treatment for a variety of reasons.

A properly conducted randomized trial will include an intention to treat analysis as the primary way to evaluate study results. The best situation is to have very few patients switch groups.