Intention to treat analysis in randomized clinical trials: A review

Background & Aim: Randomized controlled trials often suffer from two major problems, i.e., noncompliance and missing outcomes. One potential solution to this problem is using the intention-to-treat (ITT) analysis approach. Therefore, the aim of this study was to review the concept of ITT and the...

Full description

Bibliographic Details
Main Authors: Mina Mohammady, Masoumeh Sadeghi, Leila Janani
Format: Article
Language:fas
Published: Tehran University of Medical Sciences 2017-07-01
Series:حیات
Subjects:
Online Access:http://hayat.tums.ac.ir/article-1-1802-en.html
Description
Summary:Background & Aim: Randomized controlled trials often suffer from two major problems, i.e., noncompliance and missing outcomes. One potential solution to this problem is using the intention-to-treat (ITT) analysis approach. Therefore, the aim of this study was to review the concept of ITT and the most important issues related to it in practice since RCT researchers utilize it as a guide in order to improve the quality of RCT studies. Methods & Materials: A review study was performed using available resources and comprehensive analysis. For this purpose, we searched the relevant articles in databases including Ovid/Medline, SCOPUS, Web of Science, Google scholar and Magiran. The key words that were used included randomized clinical trials, randomized controlled trials, intention-to-treat analysis, per-protocol analysis, ITT, as-treated. Results: The advantages of ITT, the critique of ITT, the alternatives of ITT and their limitations, missing data and their management in clinical trial studies were discussed in this paper. Conclusion: ITT approach, due to its adherence to the principles of randomization, protects clinical trials from confounding and bias and therefore leads to the generation of the highest quality scientific evidence in the clinical research field. ITT should be considered a strategy at all stages of research, including the design of study, implementation and data analysis, and it should not be considered only a statistical approach.
ISSN:1735-2215
2008-188X