Critical Thinking in Critical Care Medicine
Blinding
When, who and what?
Blinding is the intervention that prevents people from knowing which group each patient has been assigned to. It maintains the prognostic balance of groups in randomized trial and it allows for unbiased assessments of prognosis or diagnostic tests in cohorts. Your new test results or your prognosis will not be biased by knowing a what the reference standard showed and vice-versa.
Blinding is a key part of any randomized control trial. It minimizes bias from placebo effect or nocivo effect - as a patient that knowledge alone might change how you feel about the treatment - you might feel better with the intervention, and you might feel worse if you got the control.
As a clinician, ignoring what treatment the patient is getting avoids voluntary or involuntary co-interventions, that might bias results one way or another. Even paying more attention to one group vs the other might change prognosis.
But blinding doesn't stop there, you can blind people that collect data, people that determine if an outcome is present or not, statisticians and even the writers of the journal articles.
Even objective outcomes can be affected! Small things such as how many times you call a patient to ask how they are doing or how hard you look for death certificates might affect your loss to follow up or how outcomes are determined.
Beware of terminology such as single or double blinded... they only ensure the reader is the one blinded to what happened in the study! We should ask for journals to specify exactly who they blinded and how.
Blinding of the patient and clinician usually happens by making the placebo and the intervention similar in characteristics. Of course there are cases in which blinding is hard, such as the presence of an ECMO machine or a dialysis machine. Additionally, some drugs cause side effects that might alert an astute clinician as to which group is getting what.
Blinding is usually confused with allocation concealment by novice learners, a good tool to differentiate them is the who, when and what questions:
Allocation Concealment | Blinding | |
Who is kept in the dark? | The enroller | Patient, physician, statistician, data collectors, adjudicators of outcomes |
When? | During randomization, before and during enrolment. | After randomization and when the patient is already assigned to a group. |
What do we keep unknown? | The list used for randomization | The group assignments |