Work out blinding compliance rate instantly with clear inputs, formula shown and shareable results.
Blinding is tested by asking participants to guess their allocation. In a two-arm trial chance is 50%, so 58 correct guesses out of 100 gives a blinding index of 0.16 - some unblinding, worth flagging. An index near zero means guesses are no better than chance, which is what successful blinding looks like.
Blinding index
index = (correct guess rate - chance rate) / (1 - chance rate)
Chance rate
chance = 1 / number of arms
Side effects are the main clue, along with symptom improvement. Active drugs with distinctive effects are notoriously hard to blind.
CONSORT encourages reporting it. A broken blind does not invalidate a trial but must be considered when interpreting subjective outcomes.