Abstract Background: Aminoglycoside antibiotics are commonly prescribed for the treatment of Gram-negative infections. Appropriate dosing and therapeutic monitoring of aminoglycosides are important because these agents have a narrow therapeutic index. Aim: To audit gentamicin use at our hospital, focusing on selection of the initial dose and therapeutic monitoring practices, and to compare the results against recommendations in the existing hospital aminoglycoside guidelines, which had recently been promoted to doctors. Methods: This audit included all inpatients receiving gentamicin at The Royal Melbourne Hospital from 1 February to 12 March 2004. The principal researcher checked the drug charts of all inpatients to identify those receiving gentamicin and collected data from the medical records and the pathology database. Doses were considered ‘concordant’ if the dose given was within the recommended dosing range ±20 mg. Results: A total of 132 courses of gentamicin was included in the study. Gentamicin was prescribed for prophylaxis in 31.1% of courses. Thirty-six per cent of patients prescribed gentamicin were more than 65 years of age. Eighty-two per cent of the gentamicin used therapeutically was given as a single daily dose. Sixty-six per cent of gentamicin initial dosing was not in accordance with existing hospital guidelines. Seventy-seven per cent of gentamicin courses requiring therapeutic drug monitoring received such monitoring; however, in only 8.8% of these was the monitoring conducted according to guidelines. Conclusion: Aminoglycoside prescribing practices at our hospital are suboptimal, despite ready access to prescribing guidelines. Provision of a guideline and education sessions with doctors do not necessarily lead to widespread adoption of recommended practices. We suggest that changes to hospital systems related to prescribing and monitoring of aminoglycosides are required.