A household survey of 1160 Western Australian adults was used as a basis for exploring drinkers' reports about the settings in which they drank alcohol and their experiences of alcohol related harm. Of the 873 drinkers identified, 7.9% had experienced some form of acute alcohol related harm over the previous 3 months. Violent incidents were the most common of these and drink-driving offences the least. Such harm was significantly more likely among drinkers who variously drank 'heavily', were male, single, under 25 years of age and/or who drank on licensed premises. Regression analyses revealed that even when demographic characteristics of the drinkers were controlled for licensed premises were significantly more likely to be the settings used prior to harm occurring. Bar staff continuing to serve 'obviously intoxicated' customers was the most powerful predictor of harm. Premises which offered discounted drinks or permitted crowding also tended to be those where intoxication was permitted but these variables were not directly associated with an increased risk of harm. These findings lend further weight to the view that prevention efforts should focus on licensed drinking environments and, in particular, the practice of continuing to serve obviously intoxicated customers.

The concept of the Alcohol Dependence Syndrome has been influential in the field of alcohol studies in the 1980s. The Severity of Alcohol Dependence Questionnaire (SADQ) is one of a generation of alcohol problem scales developed to measure degree of dependence rather than presence or absence of 'alcoholism'. This paper describes the development of a form of the SADQ for community samples of drinkers (SADQ-C) and its relationship to a brief scale designed to measure impaired control over drinking. In a sample of 52 problem drinkers, SADQ and SADQ-C correlated almost perfectly (r = 0.98). In a larger sample of 197 attenders at a controlled drinking clinic, Principal Components Analysis revealed one major factor accounting for 71.7% of the total variance. High internal reliability was indicated with a Cronbach's Alpha of 0.98. Application of this instrument in a random survey of Western Australian households is then described. It was necessary to remove items relating to 'reinstatement of dependence' for this sample. A single major factor was identified by principal components analysis, accounting for 69.1% of the total variance. In both the clinic and the community samples SADQ-C scores correlated highly with Impairment of Control scores. The findings are interpreted as supporting the view that there is a single dimension of alcohol dependence upon which all persons who drink alcohol with any regularity may be located.

Kreitman's discussion of the preventive paradox in relation to the prevention of alcohol problems has had profound implications for alcohol policy and has generated considerable controversy [1]. It is argued here that although Kreitman should be credited with the important observation that alcohol-related harm is not confined to a few dependent drinkers, none the less an apparent paradox is not an ideal platform from which to recommend policy. Furthermore, Kreitman's own data and data from an Australian survey of drinking are used to demonstrate that a commonplace truth underlies his apparently paradoxical findings. It is shown that the preventive paradox disappears when consideration is given to the amount of alcohol consumed on either (i) the day of highest alcohol intake out of the last four, or (ii) the day on which acute alcohol-related harm occurred. Episodic heavy consumption by people whose average alcohol intake can be classified as 'low' or 'medium' risk contributes to the bulk of such experiences of harm. It is suggested that the importance of intoxication as a public health and safety issue has been neglected. This neglect is compounded when public education campaigns and prevention policy are only based on average rates of alcohol consumption. Advice regarding the low risk levels of consumption for different types of harm should form one component of a comprehensive harm reduction policy. Other elements of such a policy should include a variety of other measures of proven effectiveness in relation to reducing levels of intoxication and related problems.