Background. There are no published criteria for improving the quality of effectiveness of randomized controlled trials (RCTs) in psychiatry.

Method. The authors review and systematize the relevant literature on effectiveness trials, with particular reference to psychiatry.

Results. In planning effectiveness RCTs in psychiatry, seven sets of issues need to be carefully considered: (i) study question (i.e. is the study question expressed in an answerable way?); (ii) reference population (i.e. what is the reference group or subgroup to which the trial results should be generalized?); (iii) patient sample (i.e. how far does the sample reflect the target population?); (iv) study settings (i.e. how representative are the study settings of routine clinical sites?); (v) study interventions (i.e. is the study intervention manualized, acceptable to patients and suitable for widespread use?); (vi) control condition (i.e. are the key characteristics of the control condition well described, and do they vary within and between sites?); and (vii) bias (e.g. attrition, blinding, concealment, consent and contamination).

Conclusions. More effectiveness trials are needed which have sufficient statistical power to provide precise answers to assist clinicians in making treatment decisions. The development of effectiveness trials in psychiatry, both for studies of individual treatments and for service evaluations, may be enhanced by carefully considering and justifying decisions in relation to each of the seven key headings proposed here.