: Amid concerns about increasing alcohol-related violence in licensed premises, Queensland introduced a system of risk-based licensing (RBL) in 2009, the first of five Australian jurisdictions to do so. Under RBL, annual license fees are supposed to reflect the risk of harm associated with the outlet's trading hours and record of compliance with liquor laws. The objective is to improve service and management practices thereby reducing patron intoxication and related problems. Using police data, we defined cases as assaults that occurred during so-called 'high-alcohol hours', and compared a pre-intervention period of 2004-2008 with the post-intervention period 2009-2014. We employed segmented linear regression, adjusting for year and time of assault (high vs. low alcohol hours), to model the incidence of (1) all assaults and (2) a subset that police indicated were related to drinking in licensed premises. We found a small decrease in all assaults (β = -5 per 100,000 persons/year; 95% CI: 2, 9) but no significant change in the incidence of assault attributed to drinking in licensed premises (β = -8; 95% CI: -18, 2). Accordingly, we concluded that the results do not support a hypothesis that RBL is effective in the prevention of harm from licensed premises. There may be value in trialing regulatory schemes with meaningful contingencies for non-compliance, and, in the meantime, implementing demonstrably effective strategies, such as trading hour restrictions, if the aim is to reduce alcohol-related violence.
Keywords: alcohol policy; alcohol-related harm; assaults; evaluation; liquor licensing; police data.