Intention to Treat and Alternative Approaches

Abstract

“Intention to treat” or “intent to treat” (ITT) is the principal approach for the evaluation of the treatment or intervention effect in a randomized clinical trial (RCT). In an RCT, patients or subjects are randomized to one or more study interventions according to a formal protocol that describes the entry criteria, study treatments, follow-up plans, and statistical analysis approaches. In an ideal trial, all randomized patients or subjects have the correct diagnosis, are randomized correctly, comply with the treatment, and are evaluated according to the study plan. These patients would have complete data and follow-up. In this case, the ITT analysis that respects the randomization principle provides unbiased tests of the null hypothesis that there is no treatment or intervention effect. The goal in many cases is to establish the efficacy of a treatment or intervention: does the planned treatment work? In practice, however, because of the many ways in which the ideal is not the reality, an ITT analysis provides a comparative evaluation of the effectiveness of the randomized intervention strategy (does the strategy work), rather than of the efficacy of the planned intervention itself. Examples of blinded, unblinded, screening, and drug clinical trials are provided. Approaches to handling deviations from ideal are described.