Review of methodologies used to estimate alcohol health harms, analysis of low alcohol doses on health protection and comparison of drinking and alcohol-caused harms in Australia and Canada

December 2019

Prof Tanya Chikritzhs

Other investigators: 

Canadian Institute for Substance Use Research (CISUR), University of Victoria, Canada
Prof Tim Stockwell
Dr Adam Sherk

Project description: 

The existence of potential protective effects on human health from alcohol use at low levels is contested in the alcohol research literature. The conditions of particular interest in relation to potential protective effects include cardiovascular conditions and type 2 diabetes. Estimates of relative risk for these conditions can vary considerably among reviews, and burden of disease estimates may be substantially affected depending on the analyst’s choice of estimate. The NDRI Alcohol Team and the Canadian Institute for Substance Use Research (CISUR) undertook a collaborative study to estimate the contribution that three prevalent health conditions, ischaemic heart disease, ischaemic stroke and type 2 diabetes, make to Australian and Canadian national burden of disease estimates and implications for national drinking guidelines, when different assumptions (scenarios) regarding relative risk of death at low-dose levels are applied.

Alcohol exposure and mortality data were acquired from administrative sources and alcohol attributable fractions were calculated using the International Model of Alcohol Harms and Policies. We then customized a recent Global Burden of Disease (GBD) analysis to inform drinking guidelines internationally. Australians are more likely to self-report as abstainers but drink slightly more than Canadians, per person. Using GBD risk functions for all other alcohol-related conditions, alternative protective effect scenarios resulted in substantial differences in estimates of net alcohol attributable deaths in Australia ranging from between 2,933 and 4,570. For Canada, estimates of net alcohol attributable deaths ranged from between 5,179 and 8,024. Country-specific weighted RR functions were analysed to inform drinking guidelines: minimum risk was achieved at or below alcohol intake of 10 g/day, depending on scenario. Consumption levels resulting in ‘no added’ risk from drinking were found to be between 10 and 15 g/day, by country, gender and scenario. These recommendations are lower than current guidelines in Australia, Canada, and some other high income countries suggesting that downward revision maybe warranted.