The primary objective of this quality improvement project was to measure and reduce the number of oxycodone immediate-release tablets dispensed to overnight stay surgical patients at discharge. The secondary objective was to reduce the proportion of inappropriate oxycodone immediate-release prescriptions at discharge. Interrupted time series analysis was performed in four surgical wards of St Vincent's Public Hospital, Sydney. The baseline period was from January 2005 to August 2013. Interventions and followup occurred until July 2017. Baseline audit of oxycodone immediate-release tablet numbers showed prescribing increased significantly with a monthly linear trend of 1.8 (95%CI = 1.4-2.3; p = 0.001) tablets/100 surgical admissions from January 2005 to August 2013. Four sequential interventions produced no significant change in the primary objective. At the end of the first month of a fifth intervention, comprising audit-feedback plus individual academic detailing, the average number of oxycodone tablets decreased by 77 (95%CI 39-115) tablets/100 surgical cases, and the postintervention linear trend was a monthly reduction of 3.2 (coefficient À3.2 (95%CI À4.5 to À1.8); p = 0.001) tablets/100 surgical admissions. Baseline audit showed 27% of oxycodone prescriptions to be inappropriate. Following our intervention, this dropped to 17% (p = 0.048), and then to 10% (p = 0.002) after 3 years.