The Intention-to-Treat (ITI) Project: A Systematic Review of the Methodological Literature

Abstract

<p>OBJECTIVES: The primary objective of this systematic review was to illustrate the various definitions authors of methodology articles have offered for intention to treat (ITT) in relation to loss to follow up (LTFU).</p> <p> METHODS: We searched MEDLINE® for publications appearing between 1950 to 2008. Eligible papers devoted at least one paragraph to ITT and two other paragraphs to either ITT or LTFU. Investigators independently extracted relevant information from each eligible article. Discrepancies between data extractors were adjudicated. Data was extracted and variables used during data extraction were analyzed using the Landis and Koch guidelines kappa values.</p> <p> RESULTS: The MEDLINE® search yielded 1007 articles. One-hundred and ten articles underwent full text screening yielding 66 articles. All kappa's were substantial to near perfect agreement (>0.74). Of the 66 articles, five (8%) did not define ITT, 25 (41%) mentioned LTFU but did not discuss its relationship with ITT, 36 (59%) commented on LTFU in the context of ITT. These 36 articles segregated into three distinctive definitions for ITT: "full-follow-up required" (mentioned 58%), "ITT and LTFU are separate issues" (mentioned 17%), and "ITI involves specific strategy for LTFU" (mentioned 78%). Of the 36 articles, 17 (47%) had multiple definitions for ITT. The most frequent strategies mentioned for handling LTFU were last outcome carried forward (50%), sensitivity analysis (50%), and use of available data to impute (46%). Most articles (81%) specifically excluded complete case analysis under ITI.</p> <p> CONCLUSION : The most striking finding of our systematic review is that there is no meaningful consensus on the definition of ITI. This review, considered alongside previous reviews of RCTs, demonstrates that simply stating a study employed ITT is useless at best and misleading and dangerous at worst. It is the recommendation of this thesis that trialists replace the term ITI with a clear statements about analytic strategies applied to participants who were followed and with those not followed.</p>