Population drinking and alcohol harm: what these Canadian analyses tell us

That history can continue to teach is once more demonstrated in ‘Alcohol consumption and fatal accidents in Canada 1950–98’ (Skog 2003). Time-series analyses over almost 50 years of all Canadian provinces show a strong temporal relationship between per capita consumption and alcohol trauma including fatal accidents, traffic accidents and fatal falling accidents. Skog concludes that the increases in Canadian alcohol consumption have had substantial effects on most main types of fatal accidents in Canada in the second half of the 20th century and that the strength of this association is comparable with Europe over the same time period. These results are both familiar when compared to studies published at least 20 years ago but also boldly new and confirming. This study reflects back to the early work of the French epidemiologist Ledermann (1956, 1964), who found that overall population drinking could be described via a log-normal distribution with its long right tail reflecting the very heavy drinking of a small percentage of total drinkers. Before this point, popular belief held that there were two sorts of drinkers: those who drink safely and acceptably and those who drink heavily and dependently, i.e. alcoholics. Within this statistical distribution, the arithmetic mean was considered a good indicator of the level of overall alcohol consumption in the drinking population. Accordingly, epidemiological research during the 1970s and 1980s produced a number of studies which found strong empirical relationships between average consumption of alcohol, i.e. usually per capita absolute alcohol, to health problems associated with drinking’, e.g. liver cirrhosis. While arguing against Ledermann's particular statistical assumptions, Skog (1980, 1985) himself asserted that changes in consumption in the population tend to be reflected at all drinking levels and that they can be tracked by the mean, or per-capita consumption. The demonstrated association between average per capita consumption and alcohol problems at the population level supported attention to alcohol problem prevention as a part of public health. Those who argued against a population-based alcohol policy objected to the ‘single distribution theory of prevention’, a reference to the log-normal distribution advanced by Ledermann. This counter-argument advanced an assumption that heavy drinkers accounted for all alcohol problems because most of the population drink moderately and rarely experienced any alcohol problems individually. At the time these arguments obscured a deeper policy issue: that is, whether countries were responsible for limiting alcohol consumption using price and physical availability restrictions as a means to prevent alcohol harm. If, as policy opponents argued, it was alcoholics who caused most of the harm, then identification and treatment was needed, not restrictive alcohol policy. Over the past 30 years this debate has largely and fortunately subsided. The evidence is now clear that countries with higher per capita alcohol consumption have higher rates of alcohol harm and countries with lower per capita alcohol consumption have lower rates of harm. The scientifically demonstrated effects of higher retail alcohol prices, limits on hours and days of alcohol sales and limits on alcohol purchasing ages have confirmed both the importance and relevance of a public health approach to alcohol policy and the importance of the early research on per capita consumption. In many respects, opponents of ‘a single distribution theory of prevention’ have been refuted slowly and quietly by accumulated evidence of alcohol policy effectiveness. Discussion about the actual mathematical shape of the distribution of consumption are now left to the rarified air of scientific discourse, not public health policy. The Skog analyses of Canadian data is refreshing in that we can appreciate the public health significance of higher average drinking across Canada over 50 years which has yielded greater numbers of fatal accidents. These findings come at an important time as there is increasing evidence that European countries, e.g. Sweden and Finland, which historically have had lower per capita consumption and associated lower problems have increased their consumption recently as a result of lower prices and less alcohol restrictions associated with membership in the European Union. Alcohol policy must be based at the population level and this paper demonstrates again the importance of such a focus. We can appreciate the importance of these findings to Canadian alcohol policy as well as to the rest of the world. We do not have to debate again what is the actual mathematical distribution of consumption, but rather how we can sensibly pursue reasonable and effective national alcohol policies to reduce harm.

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.