Main

Since 1984, the second Wednesday of March has been known as No Smoking Day (NSD) in the United Kingdom. There is evidence3,4 that a large number of smokers (up to 2 million) make an attempt to abstain from smoking, or smoke less, on this day. We therefore decided to test the hypothesis that nicotine withdrawal causes deficits in real-world psychomotor performance by comparing a measure of performance on NSD itself with performance on Wednesdays before and after NSD.

Our measure was the number of non-fatal accidents at work reported to the Health and Safety Executive in Great Britain in industries in which the executive is the enforcing authority. Workplace accidents resulting in a major injury, or in a person having to take more than three days off work, have, by law, to be reported to this body in Great Britain. The national Health and Safety Executive statistics are based on reports under the Reporting of Injuries, Diseases and Dangerous Occurrences Regulations5.

Our assumptions were as follows: first, there will be many more people in the working population suffering from nicotine withdrawal on NSD itself than on Wednesdays before and after; and second, deterioration in psychological function as a result of nicotine withdrawal causes an increased chance of an accident at work.

We compared the average number of reported accidents occurring during NSD week with the average for the two weeks before NSD week (our ‘before’ measure) and the average for the two weeks after NSD week (our ‘after’ measure). Table 1shows the means and standard errors of these measures for the ten years (1987-96) for which daily data were available. Planned paired-sample Student's t-tests (one-tailed) indicated that there were significantly more reported accidents on NSD than on the Wednesday in the ‘before’ measure (t(9) = −2.15, P=0.03), or on the Wednesday in the ‘after’ measure (t(9) = −5.95, P < 0.001).

Table 1 Numbers of reported non-fatal accidents in 1987-96
Full size table

Visual inspection of data in Table 1shows that, in contrast with NSD, there were actually fewer reported accidents on the Monday, Tuesday, Thursday and Friday in NSD week than on those days in the ‘before’ measure. This indicates that there was nothing about the week of NSD generally that was generating an excess of reported accidents. Moreover, although there are fewer reported accidents generally in the ‘after’ measure (perhaps because of better weather or more people being on holiday as spring progresses), the difference in the number of reported accidents on NSD compared with on the Wednesday in the ‘after’ measure is larger than for all the other days. Finally, there are also large day-of-the-week effects, for which there may be many underlying causes.

Of course, there may be some other factor associated with NSD that underlies the effect seen. It is difficult, however, to imagine what it might be, especially given that the actual date of NSD can vary between 8 and 14 March across years. We stress that the NSD effect, if real, should not be construed as indicating that cessation attempts on No Smoking Day are a bad idea, although it may suggest that wider use of nicotine replacement might be beneficial. Our findings could be tested with data from other countries (for example, from the Great American Smokeout) and by examining other variables for which accurate day-by-day data are available (such as road traffic accidents).