Clinical practice guidelines for psychiatry are now being developed, but important deficits in the evidence base are apparent. For many of the new treatments, clinical decisions can be idiosyncratic or based on limited knowledge. There is a need not only to perform properly constructed trials, but also to make immediate use of less rigorous forms of evidence, such as clinical practice surveys. An example is a recent survey of psychiatrists' use of antidepressant drugs. Such surveys are now part of a wider movement towards a more coordinated system of practice-based outcome assessment.