The rise of evidence-based practice has highlighted the importance of effective recruitment to randomised controlled trials if studies are to be adequately powered and valid. However, there are also increasing concerns about patient preferences and choice within trials. The current authors are involved in a trial of a programme to provide self-care skills training for people with long-term health conditions (the 'Expert Patients Programme'), and during the design stage there were significant concerns about the impact of patient preferences on the feasibility and validity of the study, because recruitment required that patients risk randomisation to a waiting list control group. This paper first details the issues raised in this trial, and the methods used to attempt to overcome them. Data on recruitment rates, loss to follow up and patient decision making about participation are presented, which suggested that initial assumptions about the impact of patient preferences were incorrect. Key lessons concerning preferences in this patient population and related issues in trial administration are then discussed.