Allocation Concealment

We know Randomization is important. But Allocation Concealment is the lock that protects it.

The Cheating Doctor

Imagine a trial for a new surgery. The doctor has a list of random assignments:

1. Surgery
2. Medicine
3. Surgery

The next patient is a young, healthy man. The doctor thinks, "He would do great with surgery." He looks at the list. The next slot is "Surgery." Perfect. He enrolls him.

The next patient is old and frail. The doctor thinks, "Surgery might kill him." He looks at the list. The next slot is "Medicine." Perfect. He enrolls him.

This destroys the randomization. The surgery group gets all the healthy people. The medicine group gets all the sick people.

The Solution

The assignment must be hidden until after the patient is enrolled.

• Opaque Envelopes: The doctor opens a sealed envelope only after the patient signs the consent form.
• Central Web System: The doctor enters the patient details into a computer, and the computer spits out the assignment.

The Diagnosis

Allocation Concealment is distinct from Blinding.

• Allocation Concealment: Protects the assignment process (before the trial starts).
• Blinding: Protects the assessment process (after the trial starts).

You need both.