Critical assessment of intent-to-treat analyses

Abstract

SUMMARYRandomised clinical trial (RCT) is considered by many to be the gold standard in cost-effectiveness studies, as one is able to balance confounding factors by the procedure of randomisation. In many cases, intent-to-treat analysis (ITT) is preferred over other approaches in handling RCT data, for its extensive inclusion of all study subjects and complete collection of cost data. However, in situations where there are a large amount of drop-out subjects, it is necessary to consider other analytical methods together with ITT. Per-protocol analysis (PP) and as-treated analysis (AT) are two frequent options. PP only includes subjects with sufficient compliance to treatment, while AT categorises patients not only by treatment but also by compliance status. In real practice, scientists should balance the pros and cons of these methods and appropriately apply them to achieve maximum study validity.