tobacco packaging

The Lancet, Volume 384, Issue 9939, Pages 233 - 234, 19 July 2014

<Previous Article|Next Article>

doi:10.1016/S0140-6736(14)61210-1Cite or Link Using DOI

Copyright © 2014 Elsevier Ltd All rights reserved.

Standardised packaging and tobacco-industry-funded research

Ashok Kaul a, Michael Wolf b

We recently assessed the (possible) effect of plain packaging on the smoking behaviour of young Australian individuals (aged 14—17 years).1 Our conclusion was that there is no evidence that plain packaging has lowered smoking prevalence among young Australians.

Our study1 has been criticised by Anthony Laverty and colleagues (April 19, p 1384).2 They state that “in view of the short time span since the measure was introduced, the variability in the measure, and the small sample size” failing to find any evidence for a plain packaging effect “is neither an unexpected nor a meaningful conclusion”.2 On the basis of a reasoning that is not explained in sufficient detail, they further claim that a reduction of 1·25 percentage points “would be required to be statistically significant using this analysis”.2

First, any actual reduction will only turn out to be statistically significant with a certain probability, and this probability is known as the power of the test. Therefore, the authors need to attach a power (number) to the specific effect of 1·25 percentage points (unless they have a power of 1 in mind, which is unrealistic).

Second, an effect as large as 1·25 percentage points is not needed to be detected with any reasonable power. For example, power against a reduction of 0·5 percentage points is about 0·65; power against a reduction of 1·0 percentage point is about 0·80; and power against a reduction of 1·25 percentage points about 0·85.1 Power of 0·8 is a commonly accepted industry standard,3 so even the power against a reduction of only 0·5 percentage points is not unreasonably low.

Laverty and colleagues1 conclude that “the lesson from Australia is that the tobacco industry's struggle against standardised packaging will not cease and it is essential to guard against continued misrepresentation of the evidence”. The data we have worked with are publicly available, and our analyses are described in detail and can be replicated. So, which evidence is supposed to be misrepresented is unclear to us.

Full-size image (24K) Associated Press

MW and AK have received fees from Philip Morris International.

References

1 Kaul A, Wolf M. The (possible) effect of plain packaging on the smoking prevalence of minors in Australia: a trend analysis. Working Paper ECON 149, Department of Economics, University of Zurich. http://www.econ.uzh.ch/static/workingpapers.php?id=828. (accessed June 26, 2014).

2 Laverty AA, Watt HC, Arnott D, Hopkinson NS. Standardised packaging and tobacco-industry-funded research. Lancet 2014;383: 1384. Full Text | PDF(173KB) | PubMed

3 FDA. Guidance for Industry—Diabetis mellitus: Developing drugs and therapeutic biologics for treatment and prevention.http://www.fda.gov/downloads/Drugs/Guidances/ucm071624.pdf. (accessed June 26, 2014).

a Saarland University, Department of Economics, 66123 Saarbrucken, Germany

b University of Zurich, Department of Economics, 8032 Zurich, Switzerland