A group treatment program, designed for post-adolescent women with a range of eating problems, was developed, implemented, evaluated and then transported to a different treatment setting to test its effectiveness. The treatment program was aimed at overcoming some of the difficulties traditionally associated with clinical trial treatment protocols. In particular, the inclusion of sub-clinical and comorbid clients, the integration of two empirically validated theoretical approaches (cognitive-behavioural therapy and interpersonal therapy), and the implementation of the program in both a generalist outpatient clinic staffed by postgraduate counselling students and a hospital specialist clinic, enabled treatment outcome to be assessed under more naturalistic conditions. Twenty-eight women (Mean age = 33 years), met in six small groups at the University clinic, and completed the 10-week group program, which incorporated both cognitive-behavioural and interpersonal components. Outcome was determined by measures of behaviour change, changes in eating psychopathology and measures of general psychopathology. In addition, treatment generalisability was assessed by running the group program in an alternative setting, with an additional 12 clients, led by two clinically trained therapists. Group data indicated improvement on most outcome measures at both posttreatment and follow-up, for treated clients (in both settings) relative to a waiting-list control group. The variables which were most affected by treatment, however, were not necessarily those specifically targeted by the intervention. The results demonstrate the need to gather empirical data on psychotherapy as it is actually practiced, with a broad range of clients, therapists and using integrated treatment approaches.