Test sales do not have impact on prevalence of smoking by children

Editor—The office of trading standards in Liverpool suggests that it is possible to reduce cigarette sales to minors from 100% to 3% on the basis of test sales in which a child is recruited to attempt to purchase cigarettes while being observed covertly by a trading standards officer (R Croft, fourth WHO seminar for a tobacco-free Europe, Skovde, Sweden, June 1994). If the child is successful, the shop is prosecuted. The drop in observed sales has been interpreted as a genuine reduction in the availability of cigarettes to children. Legislation has now made test sales standard practice. But does this approach work successfully in other districts, and can it be shown to affect the availability of cigarettes to children and thus affect the prevalence of smoking?

A cross sectional survey of two schools in Gateshead was carried out in May 1995 and May 1996. The trading standards office carried out test sales in shops around one school, school A, between these dates. The aim of the survey was to assess whether the test sales had an impact on smoking behaviour. All year 10 students (ages 14-15 years) were surveyed in the two schools (224 in 1995 and 163 in 1996). Availability, source and ease of purchase of cigarettes, and prevalence of smoking were ascertained using questions derived from a questionnaire of the Office of Population Censuses and Surveys to enable comparisons to be made with national data.1 Focus groups were also carried out to explore the availability of cigarettes in more detail.

In 1995 in school A 23/59 (39%) girls and 15/58 (26%) boys smoked. This prevalence was much higher than that in school B, where 12/51 (24%) girls and 8/56 (14%) boys smoked. These prevalences did not fall greatly by 1996. Altogether 55/58 (95%) children who were regular smokers bought cigarettes from shops at least weekly. Only 3/121 (2.5%) in 1995 and 5/85 (6%) in 1996 reported ever having had someone refuse to sell them cigarettes. This is a similar figure to the one from the national survey by the Office of Population Censuses and Surveys.1 This is not surprising since the test sales resulted in no purchases and therefore no prosecutions, despite the fact that the students were still able to buy cigarettes with ease from nearby shops.

Thus this study suggests that test sales may not be a useful measure of the availability of cigarettes to children. A problem with this type of enforcement work is that a proxy outcome is being measured. Our work suggests that the number of successful test sales does not have an impact on access to cigarettes by young people or on the prevalence of smoking. Until such evidence is available it is hard to justify continuing with test sales.