This letter is a reply to Barry A., Dennis M. Methodological and practical limitations to visually determining intoxication. Addiction 2014; 109: 851–853

We are grateful for the letter from Barry & Dennis [1] because it helps to clarify the kind of measurement validation needed to support enforcement of laws prohibiting service to intoxicated bar patrons. In fact, although in our paper [2] we discuss the development and validation of a measure of intoxication based on observable signs, addressing drunkenness of bar patrons does not require a measure that is valid across the spectrum of levels of consumption—the only validation issue for the enforcement approach we are proposing is that if someone is identified as being very drunk, they actually are very drunk.

In particular, when enforcing laws against overserving, enforcement agents are often put into the position of having to prove, after the fact, that the patron who was served was, in fact, intoxicated. Without this proof, the charge against the licensee is often dropped, leading to frustration on the part of enforcement agents and ultimately a lack of enforcement. One way around this, as we suggest in the paper, is to set the threshold quite high using gross signs of intoxication (e.g. staggering, slurring words, losing balance) to increase specificity (i.e. so that no non-intoxicated people are identified erroneously as intoxicated). This will, of course, result in reduced sensitivity (i.e. more false negatives); however, that some people who have high blood alcohol concentrations (BACs) may not be identified as visibly intoxicated is irrelevant—to make the charge stick, it is only necessary to show that the person identified as intoxicated was, in fact, intoxicated. With current technology, such observable criteria could easily be backed by video recording of the person's behaviour.

The arguments by Barry & Dennis regarding why intoxication cannot be measured reliably based on observable signs relate almost entirely to the fact that people who have high BACs may not appear to be intoxicated, thereby resulting in false negatives. Brick & Erickson [3] were also concerned with this problem—that is, misinterpreting people as being less intoxicated than they actually are—and the inherent risks this poses for safe driving. This is a very important concern for preventing drinking and driving, because such individuals are likely to have significant impairment when they drive despite the lack of overt signs of intoxication. Thus, we agree entirely with Barry & Dennis that some people who have consumed large amounts of alcohol may indeed appear quite sober because of individual differences and the tolerance they have developed. However, the main concern in terms of intoxication of bar patrons is identifying patrons who are obviously intoxicated. Failing to identify people as being intoxicated because they are well behaved and not showing any visible signs of intoxication (despite high BACs) is of much less concern.

Existing laws regarding serving intoxicated patrons are rarely, if ever, enforced and it is unlikely that random breath checks will ever be implemented to address simple intoxication. Therefore, we disagree with Barry & Dennis that setting a threshold for intoxication based on observable criteria could ‘serve as a hindrance’ to building future initiatives to prevent intoxication. If this approach can result in systematic and successful enforcement of serving laws, at least for the most obviously drunk, this would be a substantial step forward from the current situation.

Declaration of interests

In the past 5 years, the Centre for Public Health (M.A.B., K.H.) has received a grant from Drinkaware to undertake an independent study of drinking behaviours among students and M.A.B. has provided them with independent medical advice. Drinkaware is an independent UK-wide charity supported by voluntary contributions from the alcohol and supermarket industries and governed through a memorandum of understanding between the Department of Health, Home Office, Scottish Executive, Welsh Assembly Government, Northern Ireland Office and Portman Group. There are no other conflicts of interest.


  1. Top of page
  2. References