Search | Search by Center | Search by Source | Keywords in Title
Rasmussen KG, Knapp RG, Biggs MM, Smith GE, Rummans TA, Petrides G, Husain MM, O'Connor MK, Fink M, Kellner CH. Data management and design issues in an unmasked randomized trial of electroconvulsive therapy for relapse prevention of severe depression: the consortium for research in electroconvulsive therapy trial. The Journal of Ect. 2007 Dec 1; 23(4):244-50.
BACKGROUND: The use of double-blind designs, normally the criterion standard of clinical trials, is impossible when comparing medication therapy to procedural therapies for depression. In the Consortium for Research in Electroconvulsive Therapy (CORE) trial, depressed patients recently remitted with electroconvulsive therapy (ECT) were randomly assigned to receive continuation therapy with either ECT or medications. The purpose of this article is to describe the design characteristics and challenges of the trial and of our method of dealing with the lack of double-blind outcome assessment. METHODS: The primary outcome measure was time to relapse of depression in the continuation phase. We developed a method to achieve partial blinding of depressive severity assessment. This consisted of videotaping the structured interviews, having the video tapes co-rated by personnel not involved in the patient''s care, and a videotape-tracking maneuver so that the assessor of the videotapes could be blinded to phase and type of treatment. RESULTS: We enrolled 624 patients into the initial treatment phase of the trial. Of these, 201 met criteria for randomization into the second, continuation phase. Our videotape-tracking maneuver to reduce bias in outcome assessment worked well during the trial. CONCLUSIONS: The CORE study is the first multicenter, randomized controlled trial of continuation ECT in the relapse prevention of major depressive episodes. We successfully recruited a large number of severely depressed patients into a 6 month trial and used a method of reducing bias that might result from lack of blinding.