Post-randomisation exclusions: the intention to treat principle and excluding patients from analysis

摘要

Trial investigators can exclude patients' data from analysis, without risking bias, when ineligible patients are mistakenly randomised into a trial Similarly, data on patients who were prematurely randomised and so did not receive an intervention can be excluded, as long as allocation to treatment arm cannot influence the likelihood that patients receive the intervention Data should be included in the analysis when patients are randomised before information is available to confirm their eligibility and when the eligibility criteria are too broad and some patients don't have the condition of interest. But investigators can do a secondary analysis that excludes such patients Although excluding patients from analysis in certain circumstances does not bias the results, investigators should adhere to the highest standards of methodological design and trial execution to minimise post-randomisation exclusions