Executive summary

Background
Alcohol consumption is strongly linked with premature mortality. The current study aimed to determine the strength of association between national per capita alcohol consumption with homicide and suicide rates in Australia across the period 1979-2009.

Methods
Data on all Australian homicide and completed suicides that occurred between 1 January 1979 and 31 December 2009 were obtained from the Australian Bureau of Statistics (ABS). Homicides were defined by codes E960-969 from ICD9 (until 1996) and codes X85-Y09 from ICD 10 (1997-2009). Completed suicides were defined by codes E950-E959 (ICD 9) and codes X60-X84 and Y870 (ICD 10). Annual population estimates, stratified by gender, were obtained from the ABS (ABS, 2008; ABS, 2010), and used as the denominator to obtain per capita rates of homicide and suicide. Annual data on apparent consumption over the study period were obtained from the ABS (ABS, 2012a). These data present per capita estimates for total alcohol consumption of pure alcohol (L), as well as estimates for per capita beer, wine and spirits consumption. Annual mean unemployment rates for the period were obtained (Australian Government Department of Education, Employment and Workplace Relations, 2012), as well as and crude divorce rates (per 103) (ABS, 2012b).

Results
Beer consumption was positively correlated with homicide rates (r=0.70), while wine (r=-0.74) and spirits (r=-0.86) consumption rates were negatively correlated. These patterns were also true of both genders, and remained significant after controlling for unemployment and divorce rates. The relationship of alcohol consumption to suicide was narrower than that observed for homicide. Beer consumption was not significantly related to suicide rates (r=-0.20), while wine (r=-0.60) and sprits (r=-0.47) were negatively correlated. The absence of a correlation between suicide rates and beer consumption was due to a low association with male suicide (r=-0.01), compared to the significant association with female rates (r=0.64).

Conclusions
There were significant and contrasting associations with beverage type and gender, although these associations appeared stronger for homicide than suicide. For any society, the circumstances in which a particular beverage is consumed, and by whom, would appear more important in relation to violence than the alcohol concentration of that beverage.