Answer

c—Intention to treat analysis compares all patients on the basis of the treatment group to which they were originally randomly allocated, rather than according to the treatment eventually administered. This principle is applied regardless of whether patients actually started the treatment allocated, subsequently withdrew from treatment, deviated from treatment protocol, or received a different treatment. Intention to treat analysis, therefore, maintains the comparability in potential confounding factors achieved by random allocation. Confounding factors are those that might influence outcome, such as demographics, prognostic factors, and characteristics that may influence whether someone participates in or withdraws from a trial. Providing the sample is large enough, any differences between treatment groups at the end of the study will be owing to differences in treatment and not be the result of potential confounding factors.

With the introduction of the Consolidated Standards of Reporting Trials (CONSORT) guidelines, intention to treat analysis has become standard practice when analysing results from clinical trials. The CONSORT guidelines encompass various initiatives and were developed in order to alleviate problems arising from inadequate reporting of randomised controlled trials.

Nonetheless, the intention to treat principle is not without its critics. The reasoning behind intention to treat analysis is primarily pragmatic, in that this approach probably reflects what happens in clinical practice—patients may not start, complete, or continue with their prescribed treatment. Intention to treat analysis may, therefore, provide an assessment of the practical impact of being given a treatment.

Answer a is false because patients were assigned treatment using random allocation. This approach removes allocation bias, namely systematic differences in characteristics of those assigned to treatment groups when the researcher(s) or clinician(s) chose a patient’s treatment.

Answers b and d are false by definition. Intention to treat analyses include all patients in the treatment groups to which they were originally randomly allocated, regardless of which treatment they received or whether they completed the treatment regimen.

Cite this as: BMJ 2009;339:b3603