COMMENTS ON SKOG (2006): THE PREVENTION PARADOX ALWAYS LOOKED TO ME LIKE A SLEIGHT OF HAND

Anyone who has worked to raise awareness of alcohol-related problems and/or to promote effective prevention policies will be familiar with the need to broaden perspectives beyond a justifiable concern for people with severe dependence or ‘addiction’. In the last 30 years a series of government and expert reports in different countries have highlighted the pervasive nature of alcohol problems in society, their diversity and the need for broad-based responses, whether in terms of treatment services (e.g. the UK Department of Health & Social Services 1978; Institute of Medicine 1990) or public health policies that regulate price and physical availability (Bruun et al. 1975; Edwards et al. 1994; Babor et al. 2003). The prevention paradox arguments summarized with such brilliant clarity by Skog (2005) speak directly to the central question as to how widely distributed are the harms from alcohol in the general population: the more widely distributed then the stronger the case for broad-based responses. There is no doubt that strong arguments and convincing evidence are needed to persuade nervous policy makers to impose restrictions on so popular a product as alcohol (Stockwell & Crosbie 2002). The prevention paradox idea—that the many low-risk drinkers are somehow responsible for most of the alcohol-related harm—was proposed by Kreitman (1986) as an argument for policies intended to reduce everyone's alcohol consumption a little. Edwards et al. (1994) also incorporated this argument as part of their scientific case for population-wide alcohol controls. Skog summarizes well and fairly the outcomes of a series of critical analyses of the prevention paradox utilizing survey data from Europe (Skog 1999; Gmel et al. 2001) and Australia (Stockwell et al. 1996) which show that when the measures of harm are of acute problems associated with intoxication, then (a) these are best predicted by measures of drinking to intoxication (‘binge’ drinking) rather than overall volume of drinking; and (b) most of the episodes of harm are experienced by the large number of people whose overall volume of drinking is low but whose drinking pattern is ‘spikey’, i.e. they are occasional ‘binge drinkers’ (although the precise proportions will vary with methods and populations). Does this mean more or less support for population-wide or ‘universal’ alcohol policies than with Kreitman's (1986) or Edwards et al.'s (1994) interpretation of the prevention paradox? My colleagues and I suggested that the correct interpretation of the data provided a stronger case for effective regulation (both targeted and universal) because it was straightforward, easy to explain and did not have recourse to a mysterious scientific idea that defied common sense (Stockwell et al. 1996). To defend population-level controls with evidence that ‘light’ drinkers contribute most harm, without also explaining that actually it is only those among them (and there are many) who occasionally drink to excess, still seems to me like a sleight of hand. In the following years our group developed a range of national- and state-based measures of serious alcohol-related harm (deaths, hospital episodes, road crashes, violence) and ‘risky’ drinking patterns (Chikritzhs et al. 2003). These have been used to raise awareness of alcohol-related harms in Australia through wide dissemination of the National Alcohol Indicator reports. One indicator is the percentage of all alcohol consumption that puts the individual drinker's health and safety at risk: we estimated that at least 61% of all consumption reported in a 2001 Australian survey was consumed on ‘risky’ drinking days when more than 60 g was consumed by men and 40 g by women (Stockwell et al. 2004). These kinds of data provide a stronger case for universal regulatory strategies than the abstract and implausible idea that ‘light’ drinkers are the main problem. I agree completely with Skog's assertion that a ‘non-conservative’ interpretation of the paradox data is not warranted, i.e. that they support universal instead of targeted prevention strategies. The Perth group's recent Prevention Monograph also recommends a ‘balanced’ approach between universal and targeted prevention strategies in the prevention of harms from substance use (Loxley et al. 2004). However, being able to show that many young adults drink regularly in excess of the above risk levels (Chikritzhs et al. 2003), that more do so when drinking on licensed premises (Donnelly & Briscoe 2003) and that most alcohol is consumed in a hazardous fashion (Stockwell et al. 2004) strengthens the case for those ‘universal’ strategies for which there is strong empirical evidence (Babor et al. 2003; Loxley et al. 2004). I also question Skog's scepticism about the prospects for strategies that target high-risk drinkers. Even ‘rationing’ is alive and well in some Australian communities where liquor licensing restrictions have been introduced in response to extreme levels of alcohol-related problems (Gray & Saggers 2005). I also disagree that effective policies must always be unpopular (Babor et al. 2003). Having accurate data on local patterns of drinking and serious related harms can help to shape universal and targeted strategies so they are understood as being fair, effective and well-justified (Loxley et al. 2004), e.g. increased taxes on alcohol in order to fund treatment and prevention programmes (Chikritzhs et al. in press), tax advantages to lower-strength beers (Gruenewald et al. 1999; Stockwell & Crosbie 2002), restricted trading hours of premises with a record of violence (Donnelly & Briscoe 2005), stricter enforcement of underage drinking laws (Grube 1997) and random breath testing (McKnight & Voas 2004), to mention a few.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.