Allocation concealment

Once an allocation sequence has been obtained, great care must be taken in implementing that sequence. This is especially so if units have to be assessed for eligibility for the trial prior to allocation, as is the case for a clinical trial. If the person doing the assessment is aware of which treatment level is next in the allocation sequence, it may bias the decision on whether or not that patient enters the trial.

The best way to achieve concealment is to have the allocation sequence generated by someone not directly involved in the trial, either using a central telephone randomization system or by the pharmacy. In the latter case the treatment allocations are usually provided in serially numbered bottles of similar appearance and weight. If allocation cannot be done independently, then at least all allocations should put in serially numbered opaque envelopes. Patients then have to be accepted for the trial and write their name on the envelope before the allocation is revealed.

The purpose of allocation concealment is to prevent selection bias in terms of who gets which treatment. Bias later in the process both in how units are handled and how outcome is assessed is minimized by blinding.