1999; 319 doi: (Published 11 September 1999) cite this as: 1999;319:670


Objectives: To evaluate the methodological high quality of intention to treat analysis as reported in randomised controlled trials in four huge medical journals.

You are watching: What is meant by "solar radiation equilibrium" in the related ecology principle?

Design: inspection of every reports the randomised regulated trials released in 1997 in the, Lancet, JAMA, and New England newspaper of Medicine.

Main outcome measures: approaches of dealing with deviations from arbitrarily allocation and lacking data.

Results: 119 (48%) the the reports pointed out intention come treat analysis. That these, 12 excluded any patients that did not start the allocated intervention and also three did no analyse every randomised subjects as allocated. Five reports explicitly proclaimed that there to be no deviations from random allocation. The continuing to be 99 reports seemed to analyse follow to arbitrarily allocation, but only 34 of this explicitly proclaimed this. 89 (75%) trials had some lacking data top top the major outcome variable. The techniques used to address this were normally inadequate, potentially leading to a biased treatment effect. 29 (24%) trials had an ext than 10% the responses absent for the primary outcome, the methodsof managing the missing responses were comparable in this subset.

Conclusions: The intention come treat approach is regularly inadequately described and inadequately applied. Authors need to explicitly define the taking care of of deviations from randomised assignment and absent responses and also discuss the potential impact of any lacking response. Readers need to critically evaluate the validity of report intention to treat analyses.

Key messagesIntention to treat gives a pragmatic estimate of the advantage of a change in treatment plan rather than of potential benefit in patients who get treatment specifically as planned

Full applications of intention to treat is possible only when finish outcome data are obtainable for every randomised subjects

About fifty percent of all published reports the randomised regulated trials declared that intention come treat was used, but handling the deviations indigenous randomised allocation differed widely

Many trials had some absent data ~ above the primary outcome variable, and methods offered to deal with this were normally inadequate, potentially resulting in bias

Intention come treat analyses are regularly inadequately described and also inadequately applied


“Intention to treat” is a strategy because that the analysis of randomised managed trials that compares patient in the teams to which castle were initially randomly assigned. This is generally taken as including all patients, regardless of even if it is they actually satisfied the entrance criteria, the therapy actually received, and subsequent tap the money or deviation native the protocol. However there is a debate around the validity of excluding specific cases within every of these categories indigenous an intention to treat analysis.1 Clinical performance may it is in overestimated if an intention come treat evaluation is no done.2

The intention to treat method has two main purposes. Firstly, the technique maintains treatment teams that are similar apart from random variation. This is the factor for randomisation, and also the attribute may be lost if analysis is no performed ~ above the groups developed by the randomisation process. For example, in a trial comparing medical and surgical treatment for stable angina pectoris, some patients allocated come surgical intervention died before being operated on.3 If this deaths are not attributed to surgical intervention using an intention come treat analysis, surgery appears to have actually a falsely low mortality (table 1). Secondly, intention to treat analysis allows for non-compliance and deviations from policy by clinicians. There are, the course, exceptions. Some varieties of deviations from randomised allocation may happen only in ~ the trial setting and would certainly not be expected in regimen practice. Because that example, in a psychological comparing active and placebo vaccination there is the potential because that placebo vaccine to be mistakenly administered in location of active, however this might not occur external the trial and so require not it is in accounted for in estimates of potential efficacy. However, most varieties of deviations from protocol would proceed to occur in routine practice and also so have to be consisted of in the estimated advantage of a change in therapy policy. Intention to treat analysis is thus most perfect for pragmatic trials4 of performance rather than for explanatory investigations of efficacy.

Deviations from randomised allocation often result in absent outcome data. A full application that the intention come treat technique is possible only when complete outcome data are accessible for every randomised subjects treatment must constantly be required to minimise absent responses and to follow up those that withdraw from treatment, however this is particularly important because that the implementation of one intention to treat analysis.5 No consensus exists about how missing responses need to be handled in intention come treat analyses, and different approaches might be proper in various situations. Practice additionally varies over dealing with of false inclusions (subjects found after randomisation not to accomplish the entry criteria). Thus, over there is no single meaning of one intention come treat analysis, and also the expression seems to have different definitions for various authors.6 We brought out a survey of newly published reports to examine current application that intention come treat analysis.


We identified all reports the randomised regulated trials published in 1997 in four significant medical journals:, Lancet, JAMA, and also New England newspaper of Medicine. All except the New England newspaper of Medicine have adopted the CONSORT statement,7 which calls for that authors indicate whether analyses were performed on an intention come treat basis. The total variety of randomised managed trials was acquired by Medline searches because that publication form “randomized regulated trials” within each journal and cross checked versus the Cochrane managed trials register.8 The journals were then hand searched to identify trials which report an intention come treat analysis. For articles in the and also Lancet, we likewise carried out a complete text find for “intention to treat” or “intent to treat” ~ above the net (,

All trials the reported an intention come treat analysis were then separately assessed by both authors. We thought about deviations from arbitrarily allocation, false inclusions, and absent response. For each trial we taped whether each of this occurred, and if so, the method of analysis and whether this an approach was explicitly stated. The assessment of absent response was minimal to the primary outcomes if any type of were specified. Any type of uncertainties or disagreements between the 2 assessments were solved by consensus.

See more: What Does The Ghost Of Christmas Past Represent, Ghost Of Christmas Past


About half of all the randomised managed trials reported an evaluation explicitly described as intention to treat (table 2), with comparable proportions in each journal. A full of 119 reports the randomised managed trials consisting of an intention to treat analysis were assessed. Table 3 summarises your characteristics.