OBJECTIVE: To assess the effectiveness of feedback on the number of pathology tests ordered by general practices. DESIGN: Randomised controlled trial. SETTING: General practices in two primary care groups in Leicestershire, UK. SUBJECTS: 58 general practitioners in 17 practices received guidelines then feedback at 3-month intervals for 1 year about the numbers of thyroid function, rheumatoid factor tests and urine cultures they ordered, and 38 general practitioners in 16 practices received guidelines then feedback about lipid and plasma viscosity tests. MAIN OUTCOME MEASURES: Numbers of tests requested per thousand patients registered with each practice. RESULTS; There were no changes in the numbers of tests per thousand requested in either of the study groups for any of the tests. CONCLUSIONS: Feedback did not have an influence on test ordering by general practitioners in this study. More intensive strategies may be required to change the use of laboratory tests.