Over the limit? Testing non-linear associations between alcohol outlets and young adults' alcohol consumption

Introduction and Aims: By specifying a threshold at which the number of liquor licences has the most impact on local populations, authorities can work to restrict licence approvals and help prevent alcohol-related harm. Design and Methods: Raine Study Generation 2 participants reported their alcohol intake at 22 years (n = 843) and liquor licences within 1600 m of participants' homes were mapped. Analyses examined associations between licences (all licences, on-premise licences, liquor stores) and alcohol intake (g ethanol per day). Two models were fitted: (i) forced a straight-line relationship; and (ii) allowed a curved relationship via restricted cubic splines. Results: The straight-line and curved models showed significant relationships with all licences (P = 0.002 and P = 0.002 respectively) and on-premise licences (P = 0.006 and P = 0.01 respectively), but not liquor stores (P = 0.065 and P = 0.13 respectively). The straight-line model indicated that alcohol consumption increased, on average, by 0.15 g per day for each additional licence and 0.17 g per day for each additional on-premise licence. The curved model indicated that consumption increased by around 0.4 g per day for each additional licence from 0 to 10, but increases were negligible for additional licences beyond 10. The curved model provided a better overall fit to the data than the straight-line model (R² 9.52% vs. 9.18%), but the improvement in fit did not quite reach statistical significance (P = 0.08). The curvature was similar, but less pronounced for on-premise licences (R² 9.11% vs. 8.95%; P = 0.23). Discussion and Conclusions: Results suggest a possible saturation point at which additional licences have a smaller effect on the alcohol intake of 22-year-olds living in metropolitan Perth.