Background: An advance directive is a statement of a person's preferences for treatment, should he or she lose capacity to make treatment decisions in the future. Aims: To evaluate whether use of advance directives by patients with mental illness leads to lower rates of compulsory readmission to hospital. Method: In a randomised controlled trial in two psychiatric services in inner London, 156 in-patients about to be discharged from compulsory treatment under the Mental Health Act were recruited. The trial compared usual psychiatric care with usual care plus the completion of an advance directive. The primary outcome was the rate of compulsory readmission. Results: Fifteen patients (19%) in the intervention group and 16 (21%) in the control group were readmitted compulsorily within 1 year of discharge. There was no difference in the numbers of compulsory readmissions, numbers of patients readmitted voluntarily, days spent in hospital or satisfaction with psychiatric services. Conclusions: Users' advance instruction directives had little observable impact on the outcome of care at 12 months.