Introduction: The quality of controlled trials is often in question. The method of randomisation, the enforcement of blinding, adherence to study protocol and the adequacy and suitability of the control or comparison group, all have a potential influence the validity of the results and conclusions of controlled trials. In trials evaluating interventions for psychiatric disorders, little is known about the potential impact of these and other aspects of trial design and conduct. A reliable quality assessment instrument for use in all trials in psychiatry is therefore needed.

Objective: To develop an instrument to evaluate the quality of clinical trials in the psychiatric field. This instrument has been designed for use in systematic reviews and for guidance for referees and trialists.

Methods: The MEDLINE database (1966 -1996) was searched using the Cochrane Collaboration search strategy for trials relevant to the Depression, Anxiety and Neurosis Group and the results downloaded to the specialist register of trials. 52,539 references were located by this search. These were divided into 5 year periods, and from each period 5 trials were selected using a random numbers table. A number of methodological criteria was chosen based on previous research on trial quality. These have been applied to a pilot sample by 3 independent reviewers. Through discussion of problems and discrepancies the quality criteria were clarified and refined. The selected trials were then rated and the reliability of the crude and refined instruments were tested. The final instrument was then verified by an independent expert.

Results/Discussion:The final results and conclusions will be presented at the Cochrane Collaboration in October 1998.