Double-blindness procedures, rater blindness, and ratings of outcome. Observations from a controlled trial

We determined whether blindness in a double-blind randomized controlled trial of alprazolam and exposure therapies in patients with panic disorder and agoraphobia was maintained in assessors and patients, what were the factors related to "unblinding," and whether unblinding was associated with clinical outcome.

In 129 patients with panic disorder and agoraphobia who were randomized to alprazolam-exposure, placebo-exposure, alprazolam-relaxation, or placebo-relaxation conditions, blindness was tested at the end of treatment by the independent assessors' and patients' classification of the treatment condition.

Assessors' classifications were correct in 82% of the alprazolam group and 78% of the placebo group; corresponding figures for patients' classifications were 73% and 70%, respectively. Factors associated with unblinding included drug side effects but not assessors' ratings of treatment outcome.

Judgment of the validity of the outcome of a randomized controlled trial is easier if the report notes not only the use of a double-blindness procedure but also details how blind the raters remained and how any unblinding affected their ratings of clinical outcome.