INTRODUCTION: The Cardiff Model for Violence Prevention is a data sharing approach aiming to reduce violence-related presentations to emergency departments (EDs). This model has not previously been tested with patients attending EDs in a major Australian urban setting, nor has the use of media reporting of results and letters to venues without a local violence prevention board been tested. OBJECTIVE: To determine this approach's effect on alcohol-associated presentation rates during high-alcohol hours (HAHs) in a major ED servicing the nightlife district in central Melbourne. DESIGN: Between July 2017 and June 2020, this prospective study asked people aged≥18 years attending ED about their alcohol consumption and last drink location. From December 2018, quarterly letters were sent to the top five venues identified as the last drink location and press releases were sent to all media outlets. FINDINGS: A significant gradual decrease in monthly injury attendance rates during HAH was observed (coefficient = -0.001, p= 0.017), however, no step change was found. Venue-level data highlighted a short-term drop in attributed events. DISCUSSION: Sharing Last Drinks ED data via stand-alone media reporting and letters to venues was associated with a small but significant reduction in alcohol-related injury presentation rates. However, the impact waned as media stopped reporting findings. This study highlights the importance of including a violence prevention board. CONCLUSION: The model continues to have promise for reducing alcohol-related harm, however, engagement from government agencies free from alcohol industry funding is important for long-term success.