Intended for healthcare professionals


Failure of an intervention to stop teenagers smoking

BMJ 1999; 319 doi: (Published 09 October 1999) Cite this as: BMJ 1999;319:934

Not such a disappointment as it appears

  1. Donald Reid, chief executive
  1. UK Public Health Association, London SW1P 2HW

    Papers p 948

    Given the recent upturn in teenage smoking,1 would the innovative West Midlands prevention programme, described in this week's issue (p 948)2 be the magic bullet so many have been waiting for? Alas, as the authors have convincingly shown, it turned out to be a blank. This is not surprising, since the methods used did not appear to correspond with the findings from decades of research into “effective” antismoking programmes for schools.

    Successful programmes have usually been based on the social influences theory, which involves persuading teenagers to develop the skills and commitment to resist cigarettes.3 Since success depends on working with socially interactive groups, the individualised computer component of the West Midlands programme would have had little to contribute.

    The programme's class lessons component focused on the “stages of change” model of behaviour change, which was developed from studies of adults who stopped smoking. It is difficult to understand the application of this to the different process of preventing teenagers from starting to smoke. It is even more difficult to believe that it would be preferable to tried and tested approaches based on the social influences theory.

    Nevertheless, the programme might have been expected to have had an effect on existing teenage smokers. Its failure even in this group reinforces evidence that the acquisition and shedding of a smoking habit in the teenage years is essentially chaotic. Unlike adult quitting, it does not follow any readily definable stages.4 The stages of change model is therefore unlikely to be relevant.

    But, paradoxically, a positive result from either part of the trial might have led to a greater disaster. It has proved relatively easy to obtain favourable results from school antismoking interventions under research conditions—with their budgets for training and the prestige conferred on schools by participating.5 But, as follow up studies in Minnesota and Britain have shown, 6 7 the favourable effects from the original trials disappear in later years. Teachers soon start to take short cuts with the protocols, while the pressure for examination success causes schools to reduce the time available for the programmes.5

    The intervention reported this week by Aveyard et al was doomed from the start by the requirement that participating teachers should undertake a two day course beforehand. Schools across Britain are unlikely to release one or more teachers for two days' training each year on a topic which, to them, ranks below alcohol, drugs, and sex education in priority. And, short of a ruthlessly enforced decree from the government, few schools will allocate six lessons to smoking in a year (as required here) except as part of a trial.

    So it is no surprise that, despite massive efforts since the 1980s to disseminate “effective” programmes requiring training and additional classroom time in the UK and the US, there has been little change in teenage smoking on either side of the Atlantic.5 Schools simply cannot sustain complex programmes of this kind in the face of competing pressures.

    But if the results of this trial had been positive the temptation to launch a massive dissemination programme would probably have proved irresistible. Once the initial enthusiasm had worn off, any early effects would have dissipated just as they did with earlier programmes.5 And any NHS funding for the programme would have been at the expense of more effective interventions for adults, such as publicity and face to face advice from health professionals. 8 9

    There are no magic bullets to be found in school antismoking programmes: the methods that worked in the early trials had a delaying effect only,5 and none have been capable of dissemination on a large scale. Is it too much to hope that this experiment marks the end of attempts to find a quick fix, school based solution to the problem of teenage smoking? If it is, these disappointing findings will be of greater benefit to public health than they appear.


    1. 1.
    2. 2.
    3. 3.
    4. 4.
    5. 5.
    6. 6.
    7. 7.
    8. 8.
    9. 9.