PURPOSE To investigate whether and how a multi-dimensional intervention including clinical guidelines on the choice of medical treatment in the primary and the secondary health care sector, and individual feedback to general practices about their own and other practices' prescription patterns in five Anatomical Therapeutic Chemical classification system (ATC)-groups was followed by changes in the practices' prescription pattern. METHODS Prospective historical registry study and a questionnaire study of GPs' self-reported use of guidelines and feedback. RESULTS In every ATC-group the number of prescribed defined daily doses (DDDs) kept growing after the intervention, while potential savings by DDD decreased. Individual practices' changes in the prescription pattern differed by ATC-group and practices with high potential savings/DDD before the intervention showed the greatest relative reduction in potential savings/DDD. The county's average cost/DDD for the five ATC-groups declined from above the Danish average before the intervention to a level below the average cost/DDD after the intervention. In the questionnaire study (response rate: 79%), 69% of respondents had read the guidelines and 78% reported that the feedback influenced their prescription of drugs. CONCLUSIONS The observed changes in drug costs and potential savings were not due to volume effects but a combination of price effects, including generic substitution and choice of less expensive analogues, demonstrating that it is possible to change GPs' prescription patterns without interfering with patients' access to treatment or with GPs' clinical freedom.'