BACKGROUND Several studies of papers published in non-psychiatric medical journals that report on randomised controlled trials (RCTs) indicate that there is inadequate reporting of the process by which randomisation is carried out.
AIMS To examine the adequacy of the reporting of the procedure of randomisation in clinical trials of parallel design published in the British Journal of Psychiatry (BJP) and the American Journal of Psychiatry (AJP).
METHODS All issues of the BJP and the AJP published between January 1990 and December 1998 were surveyed, and papers that reported on RCTs were examined to judge the adequacy of the reporting of the process of randomisation.
RESULTS We found 183 papers which claimed to report on RCTs (73 in the BJP and 110 in the AJP). Nine (8.2%) of those in the AJP and six (8.2%) in the BJP described the technique of creating the randomisation sequence. Two (1.8%) of those in the AJP and 11 (15.1%) of those in the BJP described the mechanism of allocating treatment. One paper in the AJP and five papers in the BJP described both the generation of random numbers and allocation.
CONCLUSIONS Adequate reporting of the method of randomisation was uncommon. The RCT status of some of the papers must therefore be in doubt.