Allocation concealment

Allocation concealment means that the person recruiting and enrolling a participant cannot know or predict which group that participant will be assigned to until enrollment is complete. Without it, a recruiter who foresees the next assignment might, consciously or not, delay a sicker patient to steer them into the treatment arm, which reintroduces the selection bias that randomization exists to prevent. It is distinct from blinding: concealment protects the assignment step, while blinding hides assignment after it has happened. When reading a trial, look for methods like a central randomization service, or sequentially numbered, opaque, sealed envelopes; be wary of assignment by birth date, chart number, or alternation, which are predictable. Trials with poor concealment tend to overestimate treatment effects. Example: a phone-in randomization center reveals the assignment only after the patient is registered, so recruiters cannot game it. Ask how the sequence was concealed at the point of enrollment.