Intention to Treat Analysis: Basic †

In a randomized clinical trial design for the comparison of treatments, subjects are randomly assigned to different treatments. The intention-to-treat principle requires that any comparison of the treatments be based upon comparison of all patients in the treatment groups to which they were randomly assigned. This approach maintains the benefits of randomization, one of which is that the characteristics of the group of patients receiving the different treatments should be roughly equivalent at the onset of the trial. Several conditions can influence the actual versus intended protocol under which subjects are studied including: nonadherence to the assigned treatment regimen, or termination of participation prior to outcome measurement. Investigators may attempt to compare groups defined by factors that might be influenced by the treatments under test, for example, the comparison of only those patients in the trial that actually complied with the prescribed treatments. Any observed differences among treatment groups constructed in this manner may be due not to treatment, but to factors associated with compliance. Keywords: clinical trial; intention-to-treat analysis; noncompliance; randomization; selection bias; missing data; as-treated analysis