Drug and Alcohol Dependence 109 (2010) 4–5
Contents lists available at ScienceDirect
Drug and Alcohol Dependence journal homepage: www.elsevier.com/locate/drugalcdep
Complex alcohol pharmacokinetics: A response to Moss et al. Erling Moxnes ∗ , Lene Jensen System Dynamics Group, University of Bergen, Fosswinckelsgt.6, N-5007 Bergen, Norway
a r t i c l e
i n f o
Article history: Received 8 December 2009 Accepted 11 January 2010 Available online 9 February 2010 Keywords: Alcohol Binge drinking Overshoot Misconception of dynamics Laboratory experiment Information
a b s t r a c t The commentary by Moss et al. (2010) disagrees with policy recommendations we did not give. Our main conclusions were that “Our study warrants further studies. . . [and]. . . implies a modiﬁcation of the ‘folk wisdom’ of not drinking on an empty stomach.” Our conclusions were based on a laboratory experiment where we found that juvenile subjects behaved according to a simple feedback rule when making drinking decisions. This rule led them to overshoot intended levels of drunkenness. For further research we suggested a diagnostic information treatment to test effects on real drinking behavior. Moss et al. misinterpret our discussion of the ‘folk wisdom’ and their comments strengthen our claim that this is a complex issue that requires further studies. © 2010 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Our study was inspired by studies showing that people tend to misperceive dynamic systems and choose disadvantageous policies. The Michaelis–Menten kinetic model belongs to this category of systems. We were also aware that most alcohol information campaigns have had little effect. A natural hypothesis followed: juveniles do not understand why they get ‘drunker than intended’. Our laboratory experiment supported the hypothesis about misconception, and for this purpose it was not necessary to administer alcohol. From the experiment and discussions that followed an even more intriguing insight followed. Very few juveniles (and adults) know that alcohol uptake is delayed by the stomach and upper intestines, and even when they learn about the delay (our information treatment) they do not see the implications for drinking behavior. Without this knowledge and appreciation of it, juveniles are not able to fully analyze and learn from repeated experiences of overshoots. They have to resort to external explanations such as type of alcohol, good or bad mood, food consumed, price of alcohol, etc., all of which may be of some importance for drunkenness. However, knowledge of these causes is not particularly useful if drinking behavior is dominated by a simple feedback rule that leads to overshoots under a wide range of conditions. In a practical learning process, external explanations must be considered and rejected one by one. This takes time and may explain why our respondents reported slow learning regarding their own drinking behavior.
∗ Corresponding author. Tel.: +47 55584119; fax: +47 55583099. E-mail address: [email protected]
ﬁ.uib.no (E. Moxnes). 0376-8716/$ – see front matter © 2010 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.drugalcdep.2010.01.003
The diagnostic information treatment we would like to test is one that uses a funnel analogy to explain the dynamic system, a simulator to get hands-on-experience with the dynamics, and debrieﬁngs to interpret simulated and real experiences. We speculate that such a treatment will accelerate learning; learning that eventually seems to reduce the amount of alcohol consumed irrespective of information campaigns. Hopefully, it will also inﬂuence formation of norms and collective behavior. If such an intervention leads to greater reductions in the frequency of reported overshoots than existing interventions, our hypothesis about misconception is supported also for real drinking behavior and we have identiﬁed a functional information treatment. Moss et al. (2010) present an argument that could weaken our hypothesis in real drinking situations. They rightly point out that people do not observe BAC levels, they are more likely to observe ‘felt drunkenness’. This feeling, they point out, not only depends on the BAC, but also on the rate of change in BAC. Thus, when drinking, people get an earlier indication that an intended level has been reached than in the laboratory experiment. This should work to reduce the size of overshoots. Moss et al. (2010) also point out that “. . .as BACs increase, drinkers would ﬁnd themselves increasingly unable to make the kind of judgements posited to underlie drinking decision in Moxnes and Jensen’s (2009) computer simulation. Instead, we predict an increasing dependence upon automatic processing systems, which would be inﬂuenced by past experiences of drinking, and environmental cues and pressures.” However, our hypothesized and tested drinking rule does not involve complex judgement, it is a simple feedback rule of the same type that we use when reacting to other bodily signals. Hence, as BACs increase, there should be an increasing dependence on this simple rule. In addition, alcohol could delay
E. Moxnes, L. Jensen / Drug and Alcohol Dependence 109 (2010) 4–5
the perception of felt drunkenness exceeding intended levels. This would work to increase the size of overshoots. The main point of the commentary by Moss et al. (2010) is that we “question whether eating before drinking should be promoted to young people”. However, we never discuss what one should promote. Rather we explore possible consequences of our hypothesis about misconception. When introducing the computer simulations with and without food we write: “Then we ask, what happens if juveniles drink according to the identiﬁed feedback strategy?” In their commentary, Moss et al. seem to miss the “if”. They argue as if we base our argument on a neglect of the effect of food on BACs, and they misinterpret footnote 9 to support this assumption (the footnote is only about the alcohol that is not absorbed into the blood). The simulations in Figure 7 show that we are aware of and capture the effect of food on both delay time and metabolism (beta). See Appendix A for details. Contrary to the authors’ claim, we do not refer to these well-established effects as ‘folk wisdom’. Rather it is the “age-old rule about drinking” that we refer to as folk wisdom. The simulations in Figure 8 show the negative consequences of eating if juveniles behave according to the simple feedback strategy. Perhaps we should have said explicitly that eating reduces harm for those that do not follow our hypothesized drinking strategy, that is those for whom the “if” does not apply. As Moss et al. point out, people and circumstances vary. Thus, there are most likely situations where the folk wisdom is useful, for instance in connection with drinking contests. Hence the folk wisdom does not have to be modiﬁed for all situations and not at all if our hypothesis is found to be incorrect for real drinking situations. Only if our hypothesis applies, the folk wisdom could become a double-edged sword. When a very high percentage of our respondents found the experiment interesting, we did not claim this to be proof of utility.
However, a favorable reception among students in the target group is more encouraging than the opposite. Finally, we ﬁnd it encouraging that Moss et al. ﬁnd our methodology “potentially very useful for advancing our understanding of decision making.” The points raised in their commentary make it even more important to go on with the proposed further research. Role of funding source Self-ﬁnanced by System Dynamics Group. Contributors Erling Moxnes wrote the ﬁrst draft of the response and Lene Jensen contributed to subsequent revisions. Conﬂict of interest No conﬂict declared. Acknowledgement Thanks to Prof. I. David Wheat for commenting on the manuscript. Reference Moss, A.C., Dyer, K.R., Albery, I.P., Allsop, S., Kypri, K., Erskine, J., Mackintosh, D. Alcohol pharmacokinetics, decision making and folk wisdom: a reply to Moxnes and Jensen (2010). Drug and Alcohol Dependence, this issue.