The Trent Regional Health Authority funded a study in 1995 to train nurses in an accident and emergency (A&E) department to screen all adult attendees for alcohol problems with a view to identifying a sample of problem drinkers to participate in a randomised controlled trial (RCT). In the RCT identified drinkers were to be assigned either to health education plus brief counselling intervention or, as controls, to health education alone. Despite 16654 attendance's at A&E during the recruitment phase of the study only 20% of attendees were screened of whom a further 19% were identified as problem drinkers by the CAGE screening questionnaire. Less than half of the problem drinkers were, however, provided with feedback by the nurses, leaving a small group of 264 eligible for entry to the RCT. The great majority of this subgroup refused an initial appointment at the specialist clinic and so the trial was abandoned. A number of in-depth interviews were undertaken with the nurses in an attempt to understand ways in which the overall conduct of the study might have been improved. This paper outlines in some detail some of the reasons for the lack of success with the study which include; general environmental factors that undoubtedly led to stress and poor morale amongst the nursing team, the differences in perception between managers and clinical nurses concerning the value of research and the inadequacy of the initial training programme. The paper concludes that there are problems in the NHS which do not provide a helpful backcloth to the successful conduct of health services research.