Monday, December 16, 2013

If You Build It, They Will Drink

If there were lingering doubts about the effect of alcohol availability on alcohol consumption, a host of new studies seem to lead the reader to the same conclusion: that increases in availability are correlated to increases in consumption. In other words: if you build it (bar, liquor store, etc), we will drink. What's more, in many cases, it's not just drinking that will happen. So-called alcohol outlet density has been linked to interpersonal and intimate partner violence, adolescent consumption and beliefs about alcohol, and even alcohol-attributable deaths. On the other hand, raising the minimum prices or implementing taxes on alcohol sales seems to go a long way in reducing these potential harms.

The journal, Addiction, has published several of these studies online in the past few weeks. Gruenewald and colleagues analyzed survey data from 50 California cities with populations between 50,000 - 500,000. They found "greater on-premise outlet densities were related to greater drinking frequencies and volumes, and use of on-premise drinking places" (like bars and restaurants).  The researchers concluded that, in addition to characteristics of the individual drinkers (e.g. "impulsivity, risky driving), alcohol availability is correlated with consumption and related problems.

Also in-press at AddictionPaschall and colleagues analyzed the same sample, but instead focused on adolescent drinking. Some 1478 California youths, aged 13-17, responded to survey questions about past-year alcohol consumption, perceived availability, and questions related to underage enforcement and parental views toward drinking. The answers to these questions were then compared against alcohol outlet (bar) density, public policy, law enforcement activity and city demographics. The authors found that adolescent behaviors and attitudes were significantly affected by their environments. For example, past-year alcohol use was positively correlated to bar density and inversely correlated to "the comprehensiveness and stringency of local alcohol policies". In addition, higher rates of adult drinking were associated with greater increases of past-year adolescent drinking over the three-year study period.

Over in Alcohol and Alcoholism, Grubesic and colleagues studied the association between outlet density in Philadelphia and violent crime. Once again, the researchers found consistent association between the two. Here is a pair of maps, the first showing assault density, the second showing outlet density:

Contrary to the popular belief, no association was found between assault density and "transportation nodes and risky retailers". However, alcohol expenditures and general commercial activity were "positively and significantly" associated with assault density. 

The connection between intimate partner violence (IPV) and alcohol outlet density seems to be well established. In 2012, Conradi and colleagues reported that the density of bars in California was positively associated with IPV-related emergency department visits between 2005-2008. Then, earlier this year, Waller and colleagues found alcohol outlet density to be positively correlated to male-to-female physical - but not sexual - IPV among a national sample. Finally, in March, Zhao and colleagues showed that alcohol outlet density was associated with an increase in alcohol-attributable deaths in British Columbia between 2002-2009. In fact, they calculated that a 10% increase in private liquor stores was associated with a 2.45%, 2.36% and 1.99% increase in acute, chronic and total alcohol-associated (AA) mortality rates.

Interestingly, the single policy that seemed to have the biggest impact in turning these numbers around: raising the minimum price for alcohol. A 10% increase in the minimum price was associated with a 31.72% reduction in "wholly AA deaths". Pretty big numbers. As Dr W observed recently, "raising taxes on alcohol would do more for public health than all the treatment in the world.

What do you think?

1 comment:

  1. Correlation does not equal causation. Interesting results none the less, but I would need to see the study on how the new liquor store lead to new problems, or how closing the liquor store lead to lower problems. Also, is the cost data based on raising the cost, or on simply noticing that fewer problems occur where alcohol costs more? They are not the same thing. Retailers often charge more when people can afford to spend more. I would be for higher taxes on alcohol if that were shown to be effective (problems shown to go down after tax raised, same neighborhood. NOT problems shown to be less in neighborhood with higher taxes - again, not the same data set at all) especially if the revenue were used to fund treatment instead of flowing to the general fund.
    M. Whiting, MD