Looking for Correlation, Where None Exists (again)

By Paul

You will of course remember a ridiculous “study” from last year (covered neatly by Clive Bates here) that spread far and wide about alcohol being related to e-cig use. That one of course came from up north – Liverpool to be exact, and who do we know in Liverpool that doesn’t like e-cigarettes?

Well it would seem some researchers over in Texas don’t like e-cigarettes much either (to be fair, not many researchers in the US like them, but that’s by the by). Strangely enough, this particular study made it to the journal around the same time as the one from Liverpool. Coincidence? I think not. What is strange is why it is only now getting press-time. By press time I mean an article in the Daily Fail (where else?).

Boldly titled: “Electronic Cigarette Use Among College Students: Links to Gender, Race/Ethnicity, Smoking, and Heavy Drinking” – of course, I could quite easily end the post with this one simple phrase – kids that are drawn to certain “risky” behaviors will be drawn to other risky behaviors, there’s nothing noteworthy about that – but where’s the fun in that?

Overview

Objective

I am immediately asking myself – why? – I mean, what exactly is the point of researching something like e-cigarettes alongside drinking? As I mentioned above – those that are likely to try either smoking, drinking or vaping is likely to try any or all of the others – which can be extended to illicit drugs, but that’s not the point of this study.

Methodology

Oh. It’s one of those studies.

Conclusions

Oh joy, another one of those conclusions where they think that A must equal B, which in turn must lead to C. Are they right? Well, let’s have a look.

Early on the authors set the tone for the study:

Erm, since when is “heavy drinking” defined as 4 or more (for females) and 5 or more for the males? That doesn’t make particularly good sense to me. Are they five strong ales? Five strong lagers? Five pints of Guinness? I’m not even going to touch on the “selective” citations in that paragraph.

As with any US study on e-cigarettes, the one question they ask regarding “use” is inherently wrong – “How many times in the past 30 days have you used an e-cigarette?” – as has been discussed heavily at GFN, the right questions need to be asked to gather valid, useable information. What kind of information can you glean from answers ranging from – “I have never used an e-cigarette in my life, not even once” to “I have used an e-cigarette every day or almost every day during the past 30 days” – especially if the “tried it once or twice” style answers lead the researchers to classify that individual as a “current user” instead of someone who has experimented with the product. Do these researchers forget what it is like to be young?

Add in a simple yes/no for the “during the past 30 days have you smoked all or part of a cigarette” and you’ve got yourself a recipe for disaster. For instance from those two variables, the researchers created four “nicotine use groups” :

  • non-current users (not using either e-cigarettes or cigarettes) (n=456)
  • smokers (n=59)
  • vapers (n=38)
  • dual users (n=46)

It is useful to categorise in this way as long as the baseline information is of any actual use, instead the researchers used the US default – past 30 day – and of course self-reporting answers, the latter of which is one flaw in the approach.

Interestingly, the authors used the “e-cigarette use” question to calculate the rate of “ever-use” which would of course include experimentation. In this case, they “discovered” 175 had “ever” used an e-cigarette. In an effort to understand the cigarette/e-cigarette use more clearly, the researchers then had the participants answer three yes/no questions:

  1. Has there ever been a period in your life when you smoked cigarettes every day for at least 30 days?
    1. which would of course be classified as “smoker” rather than former smoker
  2. Have you smoked at least 100 cigarettes in your entire life?
  3. Have you used an e-cigarette on at least 100 occasions in your entire life?

For those that answered “yes” to questions one or two, they were then asked for more detail on their smoking – one wonders why they didn’t do that initially.

Notably in this table, the majority of the 599 participants were female (65%) which would tend to suggest that (for this particular campus) the males were less inclined to take the survey, although 211 did. However, as noted in the study text, neither sex or ethnic origins play a significant role in behavioral risk.

Interestingly, and contrary to the headlines generated by this study (E-cigarettes ‘ARE a gateway to smoking for young people – AND their use is linked to a greater risk of alcohol abuse’), there were more “heavy drinkers” among the non-nicotine users group than the other groups combined.

The following comment:

Can readily, and handily be debunked just by looking at Table 1. Out of 248 self-confessed “heavy drinkers” a stonking 59.7% of them were **non-nicotine users **and only 40.3% were either smokers, vapers or dual-users. Hardly a case of correlation there.

Of course, there are several limitations to this study (aren’t there always), first and foremost it’s a highly localised sample so cannot in any way be extrapolated to the wider US or global populations – despite the headlines claiming otherwise. The study focused on cigarettes & e-cigarettes – the authors “recommend” that future studies (sigh) examine a broader range of “tobacco use” – because of course, the US thinks that e-cigarettes are a tobacco product (rolleyes).

The authors conclude:

Unless I missed a section, dual-use came out as 46 out of 599 – hardly 55%! Also, the conclusion of “e-cigarette use appears to be linked to heavy drinking” isn’t substantiated by the data.

Plain and simply, this study is nothing more that a click-bait generator, useful for only one thing – spreading more confusion.

But then, it’s never been about health has it?