Generalized anxiety patients were randomly allocated to Cognitive-Behaviour Therapy, Diazepam or Placebo and managed in a primary care setting. Treatments were balanced for degree of psychologist/patient contact. A range of outcome measures, including patient self report, psychologist assessor and general practitioner ratings were used. Large variations within group response to treatment emerged. At the end of active treatment the superiority of Cognitive-Behaviour Therapy was suggested. Post-study psychotropic prescription and psychological treatment was assessed at a 12-month follow-up. The Cognitive-Behaviour group revealed the lowest incidence of subsequent treatment interventions.