Blinding strategies in the conduct and reporting of a randomized placebo-controlled device trial

Blinding is a fundamental design strengthening feature in randomized clinical trials. Blinding is particularly important in trials whose outcome measures include subjective assessment of signs and symptoms, such as trials investigating the post-thrombotic syndrome, a frequent chronic complication of deep vein thrombosis.

To determine whether strategies used to blind site investigators, research coordinators and patients were successful in a specific device trial, a multicenter trial of active versus placebo elastic compression stockings worn for 2 years to prevent post-thrombotic syndrome in patients with a first deep vein thrombosis (the SOX Trial).

Patients were randomized to the active or placebo stocking intervention, which were indistinguishable in appearance at baseline. Replacement stockings were shipped directly to patients' homes and were not worn to any of the study visits during the study. Guesstimates of treatment group assignment were completed by site investigators, research coordinators and patients at the end of study follow-up. Statistical assessments of blinding were performed using the James and Bang blinding indices.

Overall rates of correct responses were 10.4% for site investigators, 17.8% for research coordinators and 29.4% for patients. James blinding index values suggest that blinding was achieved for site investigators, research coordinators and patients. The treatment specific Bang blinding index values suggest that blinding was achieved for site investigators and research coordinators, but detected possible unblinding and opposite guess for patients in the active and placebo elastic compression stocking groups, respectively.

Post-study assessment of blinding, as was done for the SOX Trial, cannot distinguish between degree of blinding and hunches about an intervention's efficacy. However, as rates of post-thrombotic syndrome along with adverse events in the SOX Trial were similar between treatment groups, it is unlikely that hunches would have interfered with our end-of-trial assessment for blinding.

Blinding strategies used in our trial were successful overall and appeared to be most effective for site investigators and research coordinators. For patients, there may have been some degree of unblinding in the active stocking group. However, as the trial results were negative with active elastic compression stocking showing no benefit over placebo elastic compression stocking, this potential unblinding has minimal impact on the overall conclusions of the trial.