Blinding (Masking)
Keeping participants, clinicians, outcome assessors and/or analysts unaware of allocation, to stop knowledge of the arm distorting care or assessment.
Concealing the assigned arm after randomisation from one or more groups: participants, treating clinicians, outcome assessors and analysts. It prevents performance bias (differential care or co-interventions) and detection bias (biased outcome assessment). Labelled single/double/triple-blind — but those terms are defined inconsistently, so name who was masked.
Masking after allocation protects the care given and the outcome judged — most when outcomes are subjective.
After randomisation, a masking band across participants, clinicians, assessors and analysts, feeding into outcomes — with subjective outcomes flagged as far more vulnerable to unblinding than objective ones.
Ask who was blinded, not just whether the label says “double-blind”. Patients/clinicians unaware → care and co-interventions stay balanced (no performance bias). Assessors unaware → outcomes are scored even-handedly (no detection bias). Then ask how blindable the outcome is: pain and disability are easily swayed; all-cause mortality barely at all.
Unblinded trials with subjective outcomes tend to overstate benefit. But blinding is not always feasible (e.g. surgery) and matters less for hard objective endpoints — so a lack of blinding is judged against the outcome, not condemned automatically. A placebo arm is the usual tool that makes participant/clinician blinding possible.
Performance bias= patients/clinicians know → differential careDetection bias= assessors know → biased outcome scoring- State who was blinded; subjective outcomes need it most
PARAMEDIC-2 (NEJM 2018) — 8016 out-of-hospital cardiac arrests, adrenaline vs placebo, double-blind, placebo-controlled: treatment packs were identical apart from number and participants and study staff were masked to allocation. The primary outcome (30-day survival) is objective, so robust to unblinding — but blinding still mattered for resuscitation decisions and the more subjective neurological outcome.
- Mixing up performance bias (care delivered) with detection bias (outcome assessed).
- Forgetting unblinding via recognisable side-effects or a drug’s obvious physiological signs.
- Demanding blinding for hard objective endpoints — subjective outcomes are where it matters most.
Quick check
Which outcomes most need blinded assessment?
Answer: Subjective ones (pain, disability, symptom scores) — they are easily swayed by knowing the allocation. Objective outcomes like all-cause mortality are largely robust to unblinding.
Ready to build your plan? EMF Premium gives you all 40,000+ questions, 20 mocks and 1,215 OSCE stations from £29/month — or a one-off 3- or 6-month pass.