ITT analysis denotes the analysis of outcomes with this rule:-

All randomized patients in the groups to which they were randomly assigned, regardless of their adherence with the entry criteria, regardless of the treatment they actually received.

Regardless of subsequent withdrawal from treatment or deviation from the protocol. It ignores noncompliance, protocol deviations, withdrawal, and anything that happens after randomization. ITT analysis is usually described as “once randomized, always analyzed”

what authors should report ? they should say they used ITT in the statistical analysis that's all but you should examine this by examine the patients flow chart from selection to randomization then follow up to the final outcome and if there is withdrawals or change in treatment they will report it in the result section you have to well appraise Methods and Results. 

finally if they said they used ITT and you didn't anything suspetious in the results so they managed the defect properly with statistical analysis 

Hope these words helped you 

Hey Nancy!

I saw this article which is short and very clear.  http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3159210/

I generally am not a fan of ITT.  The drawbacks are just too significant.  I'd prefer to use all available data for patients over time either through an approach like last observation carried forward or by using a multilevel modeling approach.

Hope this helps and that you are well!

Loren

Thanks! Interesting, I found that exact article which prompted my question.  If I understand correctly, ITT does not compensate for differential attrition - it simply serves to reduce potentially spurious effect sizes by including the initial data of people who dropped out and did not complete outcome measures - am I correct?

Hope you are very well, also!

Nancy

Yes.  I believe you have it. A crude approach to account for attrition or noncompliance.