OBJECTIVE To determine the number of laboratory confirmed urinary tract infections in children and to ascertain general practitioners' practices and attitudes towards their investigation and management. METHODS Prospective one year survey of urine specimens submitted for bacteriological investigation; review of radiology department records; questionnaire survey of general practitioners. METHODS Gloucester health district. METHODS 57,432 children aged < 15 and 7143 children aged < 2 registered with a general practice in Gloucester health district and their 195 general practitioners. RESULTS 4317 urine specimens were submitted from children aged < 15, of which 563 from 442 children were culture positive. The rate (number/100 children/practice) of culture positive urine specimens in these children varied more than 10-fold between general practices, and this correlated closely with the rate of referral of urine specimens for investigation. A follow up specimen to check for clearance of infection was taken in 22% (125/563) of infections. Of the 821 specimens submitted from children aged < 2, 103 from 89 children were positive. Of these children, 28 underwent radiological imaging. Most general practitioners would aim to obtain bacteriological confirmation of urinary tract infection on weekdays but only a minority said they would do so at weekends. They were apparently more likely to refer boys and younger children for renal tract imaging after a first urinary tract infection. CONCLUSIONS Urinary tract infection in children was underdiagnosed, and after a confirmed infection only a minority of patients received renal tract imaging or microbiological follow up. Greater awareness of the importance of investigation and management of urinary tract infection in children is needed, and the practical difficulties faced by general practitioners must be resolved.