Unhealthful News 211 – Study offers insight into how addiction makes it easier to quit smoking

I just realized that the state university system here (Penn State) might well be one of the best contributors to useful research about tobacco after the usual suspects (Rodu, Lund, Snowdon, BAT, etc.). A huge majority of research published on smoking and other tobacco use is either completely useless (telling us what we already know, or being so convoluted that it is impossible to translate into useful information) or is just pure advocacy disguised as research. So a few informative studies is enough to put a person or institution into the top ten. 

I have to say I did cringe when I saw this headline in the Penn State house-organ magazine/feed:

Mobile technology helps explore nicotine addiction

I was anticipating something that told us nothing about addiction, and was really just a glorified quit-line with someone gathering some useless noisy data and calling it research. Indeed, the reporter included some of the usual pablum in the article, but the research itself was interesting. The mobile phones were used to get instant feedback from smokers who were trying to quit about their emotions, desire to smoke, and smoking status, which likely give much better quality data than a long survey a week in. The reporting of the research included this gem:

“One thing that really stood out among the relapsers is how their urge to smoke just never dropped, in contrast to those who were successful in quitting for a month — their urge dropped quickly and systematically — almost immediately upon quitting,” said Stephanie Lanza, scientific director of The Methodology Center at Penn State. “That was surprising to see.”

That actually tells us something potentially useful about addiction, as promised by the title. However, it is necessary to think a bit about what addiction really means to see what it is saying.

I usually argue that the word “addiction” has no place in scientific discussions, at least not if it is not defined. As typically used, it is the worst kind of terminology: It seems like it means something significant, but if you start drilling down on it, it seems to not mean anything special after all. It is purely a “I know it when I see it” phenomenon in its common usage, and different “I”s definitely see it differently. This makes for terrible science, and use of the undefined term as if it is meaningful always strikes me as a convenient easy clue that an author really does not know how to think scientifically. 

Sometimes “addiction” is used as a synonym for “use”, or for use with an element of condemnation (as in “use of something that I [the author] disapprove of”). Sometimes it invokes a theme of users not wanting to quit, though no distinction is offered that defines “addictive” as meaning anything other than highly desirable. Perhaps the most common use of the term is to refer to some overwhelming urge, but that is clearly not what it means (or else breathing, thirst, and wanting to see your child are addictions). A small minority of those who use the term do manage to get closer to some meaningful concept, that it has something to do with a non-typical pattern of preferences and behaviors, though they seldom nail it down.

The one group of researchers that I am aware of that have declared a meaningful formal definition are the Chicago School economists.  This should not be surprising, since the phenomenon clearly has to do with individual preferences and consumer behaviors — the stuff of microeconomics, not of the biological or epidemiological side of the health sciences. (Note: Yes, in saying this I am declaring that the DSM-type wanderings do not constitute a definition of “addiction” (or “dependence” or the like). Take a look at them. They are just lists of vague considerations.  It is basically a technical-seeming way of saying, “you will know it when you see it; look for characteristics like….” There is no there there. It is much closer to something like “what makes a novel good” than a scientific definition.)

To simplify down to its essence, the Chicago definition of addiction focuses on how past consumption increases the marginal value (i.e., how much you want it) of later consumption. That is, most goods have diminishing marginal returns: if you have consumed a lot lately, you are not so interested in more right now (this is the standard mathematical assumption in economics, and is generally true so long as you choose “a lot” and “lately” right). So if you have just eaten, your desire to eat more is lower than it was before you ate. If you spent yesterday hiking, you are over it for a while and can settle in to work. But for some things, the desire is actually increased by recent consumption, contrary to the standard Econ 101 assumption. So if you did not smoke today, yesterday, or any time in the last decade (or ever), your desire to smoke is pretty low; but if you smoked a pack yesterday, your desire to smoke today is going to be pretty high.

Some readers are probably already picking holes in this construction. If you smoked five minutes ago, your desire to smoke right now is less than it was when you were lighting up that last one, not more. The eating example is mostly right, but if you just ate one bite of baklava (or whatever your favorite dessert is), your urge to eat more is probably increased. The “increasing marginal returns” definition has some complications and exactly when it starts and stops have to be defined ad hoc. Lots of things — probably most any desirable consumption good — are addictive on a short time scale after a small amount of consumption (“I just ate one bite of this and so want a second”) or a long term scale after frequent habituation (“I had never wanted to try that restaurant, but I started eating there every Thursday and I don’t plan to stop”), and nothing is addictive at every time and quantity (“since I did a couple of lines of coke five minutes ago, I don’t want any more right now”). But for some values of time and quantity marginal returns are increased, and when that occurs more than is typical, it is interesting.

This makes it a potentially useful conceptualization and basic structure of a definition, but not a complete definition or anything you can push too hard on. (The attempts by Chicago economists and others to do empirical work based on the definition are, thus, pretty much a joke — and I say this as someone who has done such work.)

Still, if you start with the concept, but do not take the details too literally, you can do some interesting science. Start by realizing that the nature of addiction, so defined, depends on the exact pattern of time and quantity that creates increasing returns, and probably the intensity of that increase. That makes it meaningful and worthy of empiricism. Then look back to the quote from the article. 

From that quote, I think we can conclude that those who have the easiest time quitting are the ones who were mainly motivated by addiction, and get an estimate of the parameters of that addiction. That is, the reason they wanted to smoke today as much as they do is because they smoked yesterday. Once that is gone — once they have not smoked for just a few days — the desire to smoke, which was caused by the increasing marginal returns effect, drops off, and quitting is easy.

Who does not quit? Those whose benefits from smoking are not mere addiction. For them, being abstinent for a few days or even a few weeks does not change the sign of their cost-benefit calculation, which says “I am better off smoking than not”. It is, of course, possible that their marginal value did drop after a few days of abstinence, but not so much that they decided that quitting was better.

The basic point here is not news to those of us (an extremely small minority among those who study tobacco and such) who understand that smoking and other tobacco use is, first and foremost, a free consumer choice, and the departures from that are quite interesting but fundamentally secondary. But the value of this result is probably lost on the ignorant masses who claim that there is something fundamentally different about smoking compared to other consumption choices.  They are like the astronomers who assumed Earth was a fixed center or chemists before the discovery of isotopes, who generated enormous quantities of completely pointless study results.

Of course, it is possible the the Penn State researchers might have said something dumb, like declaring the non-quitters to the ones they found to be more addicted. But perhaps not. Perhaps they did the cleanest possible science, not even using the word addiction, and simply reported their results. Maybe they even got it right. I have to say that I have chosen to not go look at the actual research so that I do not have to burst this little bubble of believing that someone did some useful novel research about tobacco/nicotine use, and drew accurate conclusions from it.

There are reasons for optimism. The key researchers were methodologists, not public health people. This provided observations like:

“To me, the biggest innovation here is looking at how something like baseline dependence is predictive of that behavior over time or (specifically) the urge to smoke over time,” said Lanza. “It’s now expressed as a function of time. Instead of saying ‘if you’re higher on dependence you’re going to have a higher urge to smoke over time,’ you can now depict how that association between baseline dependence and urge to smoke varies with time in a very fluid and naturalistic way.”

In other words, it is an attempt to measure addiction based on the one almost-concrete definition of it in the scientific literature, something that has to do with preferences over time as consumption changes. This is genuinely useful stuff, helping us understand people’s motivations and thus (*gasp*) helping them make rational choices rather than treating them like agricultural research crops, an object to be manipulated. It would be over-interpreting this one bit of research to out-and-out declare “this confirms the conventional wisdom that the addictive component of smoking wears off in a day or two, and so those who are motivated to smoke mostly by addiction and who want to quit just need to get over that hump; those with other motives are an entirely different story, and their choice to smoke cannot be explained by addiction, at least not on this time/quantity scale”. But it definitely points in that direction, and promises some interesting work to come.

“Our goal is to work hand-in-hand with tobacco (and other) researchers, to help them understand these really intricate processes that are happening,” said Lanza. “We want to really understand addiction and how to break addiction, so that interventions can be targeted and adaptable.”


Well, maybe they will get another study or two out before they abandon science.


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s