Objectives: Urinary tract infections (UTIs) are one of the leading reasons for antibiotic prescriptions in emergency departments (EDs), with half of these antibiotics being inappropriately prescribed. Our objective was to assess the impact of a computerized decision support system (CDSS) on compliance with guidelines on empirical antibiotic prescriptions (antibiotic and duration) for UTIs in EDs. Methods: A multicentre prospective before-and-after controlled interventional study was conducted from 19 March to 28 October 2012. All adults diagnosed with community-acquired UTIs (cystitis, pyelonephritis or prostatitis) at three French EDs were included. The antibiotic therapy was considered compliant with guidelines if the antibiotic and the duration prescribed were in accordance with the national guidelines. Data were collected using electronic medical records. Paired tests were used when comparing periods within each ED and global analyses used multivariate logistic mixed models. Results: Nine hundred and twelve patients were included during the 30 week study period. The CDSS was used in 59% of cases (182/307). The CDSS intervention improved the compliance of antibiotic prescriptions in only one ED in a bivariate analysis (absolute increase +20%, P ¼ 0.007). The choice of the antibiotic was improved in multivariate analyses but only when the CDSS was used [OR¼ 1.94 (95% CI 1.13 –3.32)]. The CDSS also changed the initial diagnosis in 23% of cases, in all three EDs. Conclusions: The CDSS only partially improved compliance with guidelines on antibiotic prescriptions in UTIs.