• ARGH! ARGH ARGH ARGH!!!!! – Reporting science isn’t that hard.

    Friends just sent me this. “Beers Implicated in Emergency Room Visits“. That’s from the NYT:

    Nationwide, roughly a third of all visits to emergency rooms for injuries are alcohol related. Now a new study suggests that certain beverages may be more likely to be involved than others.

    The study, carried out over the course of a year at the Johns Hopkins Hospital in Baltimore, found that five beer brands were consumed most often by people who ended up in the emergency room. They were Budweiser, Steel Reserve, Colt 45, Bud Ice and Bud Light.

    Three of the brands are malt liquors, which typically contain more alcohol than regular beer. Four malt liquors accounted for nearly half of the beer consumption by emergency room patients, even though they account for less than 3 percent of beer consumption in the general population.

    The Atlantic tells us, “Study: Bud Is the King of Beers…at the Emergency Room“. TIME says, “Budweiser Drinkers Are Most Likely to End Up in Emergency Room“. CBS News reports, “Study: Many alcohol-related ER visits involve Budweiser“. ARGH!!!!!

    This is the study. “Alcohol Brand Use and Injury in the Emergency Department: A Pilot Study“. Did you see those last three words? A PILOT STUDY. Here’s the objective (emphasis mine):

    In an urban emergency department on weekend nights in 2010 and 2011, 105 interviews assessed feasibility of collecting alcohol brand consumption data from injured patients who drank within 6 h of presentation

    The point of this study was to see if it was possible to collect data from injured patients. It was NOT TO MAKE CONCLUSIONS OF WHAT THEY DRANK. From their conclusions (emphasis mine):

    Our pilot study demonstrated that collecting alcohol type and brand data in a hospital emergency department is feasible, if labor-intensive. Physicians were welcoming of and cooperative with the research team, and the hand-held netbook methodology enabled the research team to conduct the survey in a very brief and unobtrusive manner. Securing patient agreement to participate in the study improved substantially when the research team wore white lab coats.

    The limitations of the study are clear (emphasis mine):

    Limitations of our study included the small sample size, limited to patients presenting at a single urban hospital emergency department; the temporal difference between the data on beverages consumed, collected in a single neighborhood in a single city in 2010–2011, and the market share data, reported nationally for 2010; and the use of self-report data. Surveys of alcohol consumption often find that the population under-reports its alcohol consumption, with surveys capturing as little as 30%–60% of the market as reported in sales data (Duffy & Waterto, 1984). Research team members noted a tendency to report small amounts of alcohol consumed relative to the respondents’ earlier states of intoxication and injury severity. Given the vulnerable position of respondents, social desirability bias may also have played a consistent role in under-reporting the amount of alcohol consumed (Davis, Thake, & Vilhena, 2010). We did not record number of refusals in the study, but did find that refusals were infrequent after the research team donned white laboratory coats.

    While our study cannot definitively identify problematic beverage types or brands beyond the neighborhood and city in which the data were collected, it did establish that these data can be collected, and can provide insight into the alcohol consumption of an emergency department population, compared to the share of brands and types of alcohol in the broader marketplace.

    Please do not write articles on how beer compares to other alcohols in terms of ED visits based on this study. Please don’t tweet about it. And please, please don’t email me about it. I’m already screaming enough today.



    • “I’m already screaming enough today.” This line was worth the price of entry only because after reading it I was yelling at the screen. Thank you.

    • I first read the article title as “Bears Implicated in Emergency Room Visits“. I thought that it would refer to Stephen Colbert’s ThreatDown which always has bears as the Number 1 threat.

      “I believe all God’s creatures have a soul — except bears, Bears are actually Satan’s children.”

    • You have my sympathy. Sheesh.

      But I see lots of problems with the proposed study, even above the ones they mention. A grad student friend transferred from Yale to a state school in another state when said friend’s advisor got a better offer. She reported that the amount of beer consumed was roughly the same, but that at Yale, you had to drink imported and pretentious brands and at her new school, the object was to find the cheapest beer. So it’s not the brand but the culture. I’m just not sure that there’s really anything there to measure and too many variables. To say nothing of where the subjects did their drinking: home, bars, parties.

      Also, the “almost no one refused to cooperate when the researchers wore white lab coats” bit sounds to me to be crossing a line into unethical research behavior: the subjects will have recently been in a serious accident involving alcohol, and sleazing then into agreeing to participate when they are in a vulnerable/hurt state stinks.

      • The fact that the study occurred in a single hospital in a single neighborhood means that the results were almost certainly biased by subculture. If it were a college neighborhood vs. an ethnic neighborhood, they’d get different results. You’d have to repeat this in lots of places to get anything meaningful.

        Of course the authors are aware of this, but it didn’t stop the media.

    • Well the NY Post struck with this: Budweiser drinkers most likely to end up in emergency room.

      And they labeled it a story. Sorry add to the collective pain


    • Their goal is to attract readers, they did that right.