11 What is intention-to-treat analysis?

Our expert at JBI, Dr Tim Barker, outlines the importance of intention-to-treat analysis, the reporting of primary research data from randomised controlled trials, and the considerations relevant to authors of systematic reviews of effectiveness.
Chapters:
00:00 Introduction
00:22 What is intention-to-treat analysis?
00:57 What are the benefits of intention-to-treat analysis?
02:30 What are common problems of intention-to-treat analysis?
03:31 What resources are available for researchers?
Resources:
synthesismanual.jbi.global (Chapter 3)
jbi.global/education/systemat...
jbi.global/
JBI is an international research organisation based in the Faculty of Health and Medical Sciences at the University of Adelaide, South Australia. JBI develops and delivers unique evidence-based information, software, education and training designed to improve healthcare practice and health outcomes.
#jbiebhc

[@GeoffGraham8002]

Thanks Dr. Tim

[@richardfeinman6581]

Essence of intention-to-treat: If you assign a group to take a pill, and you know that nobody took the pill, you can then say that their outcome tells you about the effect of the pill. If it sounds crazy, it is. Can craziness be defended. I recommend the video.

[@thinhpham49]

You try to give the video more brightness it will be great if you do

[@ProfFeinman]

“You want an even distribution between groups,” and doctors want what they want whether the data is their or not. In short, they analyze the data in terms of their intention rather than what happened, because what happened contains two variables, the intervention under study and the adherence of the subject. The first is about the biology and if the experiment is done right can be expected to be reproducible, the second depends on the patient, the experimenter, the weather or whatever and is not predictive. Of course, if you don’t know which patient did what, you have no choice but to do ITT which is what we always did but didn’t give it a pretentious name. If you know who adhered then you have an additional piece of information. You can’t “break the randomization.” You are randomized before the experiment and your performance can’t change that. In the end, ITT has to admit that if the patients all had spontaneous remission of symptoms and none of them took the pill, then it was a very good medicine. If that makes sense to you, you should do an ITT analysis.

[@ayyo4173]

Thanks for this info. Could you guide me in which direction I can get more info regarding your last two statements? A systematic review I'm reading that is cited by my local health unit, includes mostly studies done using ITT regarding efficacy of ma$k w3arlng. Do you think this type of study is appropriate for that topic?