Intention to Treat / IIT Analysis

RCTs >

Intention to treat analysis (ITT analysis) is a method of statistical analysis often used in medical research. In ITT analysis, a study participant is analyzed as belonging to whatever treatment group he/she was randomized into, whether or not the treatment course was completed as intended.

itt analysis intention to treat
Randomized experiments are commonly used to assign participants to treatment and control groups.

Reasons Behind Using Intention To Treat Analysis

Intention to treat analysis is typically used to cut down on bias that might occur if drop outs or switch overs in a treatment study are non random. For instance, in the testing of a drug those patients who feel worse might be tempted to either switch treatment groups or drop out of the study. Eliminating these participants, as in traditional analysis, would lead to misleading results.

In general, though, rather than considering ITT as describing the outcomes if a specific treatment is given as planned, it should be considered as giving an estimate of the benefit a change in treatment policy might give a patient.

ITT Analysis vs. PP analysis

ITT analysis is the gold standard for randomized clinical trials where data from all subjects initially enrolled in a clinical trial are included in the analysis. When subjects do not complete the trial, or complete it imperfectly (i.e., they deviated from the protocol), a PP analysis can be performed instead. ITT and PP analysis can each introduce different biases into study conclusions regarding safety and efficacy. It is for this reason that researcher usually perform both types of types of analysis (Brody, 2006).

Weak Points of ITT Analysis

However well intentioned they might be, intention to treat analyses tend to be lacking in both application and write-up. Often researchers claim to use intention to treat analysis without specifying what is done in case of drop outs with missing data or other deviations.

ITT analysis can only be carried out fully if there is a complete and full set of data. Since the core of the method is that drop outs or treatment switches are ignored, outcome data for every study participant who was present at the initial randomization must be included in the randomization. This means that, though it is logistically difficult, even drop outs must be followed up on.

References

Brody, T. Chapter 8 – Intent-to-Treat Analysis Versus Per Protocol Analysis, Editor(s): Tom Brody, Clinical Trials (Second Edition), Academic Press, 2016,
Pages 173-201, ISBN 9780128042175,

Detry MA, Lewis RJ. The Intention-to-Treat Principle: How to Assess the True Effect of Choosing a Medical Treatment. JAMA. 2014;312(1):85–86. doi:10.1001/jama.2014.7523. Retrieved from https://jamanetwork.com/journals/jama/article-abstract/1884555 on August 26, 2018.

Science Direct Topic: Intention to Treat. Retrieved from https://www.sciencedirect.com/topics/medicine-and-dentistry/intention-to-treat-analysis on August 26, 2018.

Hollis Sally, Campbell Fiona. What is meant by intention to treat analysis? Survey of published randomised controlled trials BMJ 1999; 319 :670. Retrieved from https://www.bmj.com/content/319/7211/670 on August 26, 2018.


Comments? Need to post a correction? Please Contact Us.