Background: Febrile illnesses are common among children in Ghana and are often diagnosed as malaria, thus overlooking urinary tract infection (UTI) as a possible cause of fever. Aims: To determine the prevalence of UTI among febrile children,5 years and to estimate the sensitivity, specificity and positive and negative predictive values of urine dipstick as a screening test. Methods: From March to July 2004, children aged 3-60 months attending an outpatient clinic at Komfo Anokye Teaching Hospital, Kumasi were systematically screened for UTI using Combi 10 dipstick (CyBow TM). All dipstick-positive and a sample of dipstick-negative urines underwent microscopy and culture (i.e. gold standard) from clean-catch or catheterised urine. Results: Of 1393 children (median age 20 months), 112 (8%) had a positive dipstick and 29 of these (25.9%) had UTIs; 118/ 1278 (9.2%) children with a negative dipstick had urine cultured, one of whom (0.8%) had a UTI. The prevalence of UTIs was 2.1% (30/1393) and was higher among females (RR 3.99, 95% CI 1.76-9.04). 70% of UTIs were in children,2 years of age (p=50.08). The sensitivity, specificity and positive and negative predictive values of dipstick were 96.7%, 58.8%, 26.1% and 99.2%, respectively. Use of dipstick as a screening test for UTI was comparable to microscopic analysis for pyuria. 90% of all UTIs were clinically misdiagnosed (70% as malaria). Escherichia coli was the predominant isolate (60%). Co-trimoxazole and ampicillin, commonly used to treat uncomplicated UTIs at first level in Ghana, showed 0% and 8.3% in-vitro sensitivities, respectively. Ciprofloxacin and cefuroxime (widely used at regional/ tertiary level) showed good sensitivities, 99.0% and 86.2%, respectively. Conclusions: Urine dipstick should be promoted as a screening test for UTI. First-line use of cotrimoxazole and ampicillin for UTI should be reviewed.