summary Evidence-based medicine (ebm) is an algorithm that integrates information technology and the results of outcome research into clinical work. ebm, if well understood, is a valuable and critical bottom-up method that promotes a dialectical process by which the practitioner is able to arrive at an informed treatment choice; the method starts from the individual patient and involves discussion with the patient and critical reading of the relevant literature. However, there is a risk that ebm will become an ideology unless we are aware of its assumptions and epistemology. It is not valuefree, but is driven by an implicit objectivist view of man. ebm attaches greater value to biomedical sciences that to the humanities, to empiricism than to hermeneutics and to quantitative research than to qualitative research. ebm focuses on internal validity, but is in danger of neglecting factors such as qualitative changes that are difficult to measure, ecological validity and effectiveness and may hinder the generalisation of research results to clinical practice. The limitations of ebm are even more poignant in psychiatry, because psychiatry is highly sensitive to context and values. ebm is a blessing if we use it critically and remain aware of its limitations and underlying philosophy and if we supplement it with strategies such as value-based medicine (vbm). ebm, however, does not teach us what is useful or valuable, whereas vbm helps us to explain and weigh up the values that are at stake.