Papageorgiou et al (Reference Papageorgiou, King and Janmohamed2002) are quite right to point out that advance directives (or advance statements) have potentially beneficial effects on the processes of care, but the best way of implementing and evaluating them is far from clear. They chose a randomised controlled trial (RCT) to evaluate the effectiveness of advance statements, and used the number of compulsory admissions a year later as the main outcome measure. They found that advance statements had little impact on the outcome of care.

Different research methodologies exist to answer different types of research question, and while RCTs may be appropriate for establishing the effectiveness of an intervention, they provide little information as to the best way of implementing and delivering an intervention, especially complex interventions such as advance statements, which serve the ethical purpose of trying to preserve individual autonomy. In view of this we have to consider the power relationships between service users and mental health services. The authors appear to be aware of these relationships. The booklets in which their patients wrote their directives clearly stated that patients' wishes could be overridden by compulsion. This raises many questions.

Who ‘recruited’ patients into the study? How did recruitment take place? What steps were taken to inform service users about the pros and cons of advance statements? How were service users, professionals and structures of care such as the Care Programme Approach process prepared for advance statements? These questions concern power, values and interest. Do professionals really consider advance statements to be helpful, and take their implementation seriously? If not, how might this affect the way patients respond when asked whether they want to write an advance statement? The discussion in the paper indicates that staff may have had a ‘lack of sustained awareness’ of advance statements over the follow-up period. Our experience in Bradford indicates that a considerable amount of developmental work with mental health professionals and service users is necessary if advance statements are to be implemented.

The Medical Research Council (2000) has prepared a framework for use of RCTs for complex interventions, which sets out four stages of development. It starts with pre-clinical justification for the intervention, followed by modelling (defining the intervention and understanding the relationships between the component parts), and concluding with long-term implementation of the intervention. A definitive RCT should take place only after the first two stages have been completed. Our experience in Bradford is that advance statements are complex interventions that require lengthy developmental work if they are to stand a chance of success. Papageorgiou et al make no reference to what, if any, developmental work took place before the introduction and evaluation of advance statements, making it difficult to draw conclusions about their effectiveness or otherwise.