summary
Randomised clinical trials (rcts) generate knowledge that is useful in some situations but is of limited value when it comes to dealing with complex problems in clinical practice. By means of arguments drawn from acute psychiatry it is argued in this paper that a) the preconditions under which an intervention seems to work in a trial are all too often not met in the clinic; b) the problems that evidence-based interventions seek to solve tend to be unlike the problems that clinicians encounter in a day-to-day basis; and c) it is often impossible to ascertain the parameters against which the efficacy of intervention can be measured. Therefore, acute psychiatry, like all the other branches of healthcare, urgently needs not only information about 'proven efficacy' via rcts but also other kinds of information derived from other methods of research.