We have just been alerted to the trial by Hollis et al(1) and felt
that other conclusions could be drawn if the totality of the evidence on
the effectiveness of the TTM-based interactive computer program was
considered. We are the authors of the European study described in the
discussion that found no effect of a similar intervention when tested in
schools(2;3).
The authors criticise our trial for imbalance of study arms, but this
is just wrong. The most important predictor of smoking at follow up was
smoking at baseline. In the intervention group, 13.3% were regular (1
cigarette per week minimum) smokers and in the control group it was 12.8%,
while the corresponding figures for never smokers were 51.8% and 54.8%.
Furthermore, we adjusted for age, sex, ethnicity, mother, father, sibling,
best friends smoking habits, baseline smoking habit, baseline stage, and
socio-economic deprivation. The real test of imbalance is if the
unadjusted and adjusted odds ratios differed by much. For example, in the
main analysis, the unadjusted odds ratio for regular smoking at follow up
for TTM versus control was 1.08 for the effectiveness of the intervention,
which changed to 1.14 on full adjustment. The authors say that follow up
was short, but we followed participants one and two years from baseline,
which was four and 16 months from the end of the intervention(2;3). It is
most unlikely that intervention effects would occur more than 16 months
from the end of the intervention if they had not been apparent earlier.
The authors also imply that engagement with the intervention might
also explain the difference in the apparent effectiveness of the
intervention in the two trials. This also is an unlikely explanation. In
our report cited by Hollis et al, we show that 70.2% of baseline non-
smokers and 55.6% of baseline smokers thought the session was both
interesting and useful, declining to 59.6% and 44.9% on second use(4).
Hollis et al recorded no comparable data, but these high levels of
engagement in our trial suggest lack of efficacy was not due to this
cause. We argued for reasons described in the paper cited by Hollis et al
that non-engagement with the intervention was a risk factor because it
reflected non-engagement with schooling generally(4).
At the time of our original trial report, Prochaska put the lack of
efficacy down to adolescents having too few sessions on the interactive
computer program, arguing that adolescents are more resistant to
intervention than are adults(5). The results of Hollis et al suggest
that, as with adults(6), one or two sessions with the expert system is
sufficient for an effect, so this too does not explain the lack of
efficacy we observed.
We suspect that the difference between the trial results occurs
because of two factors. The first is the context. Visiting the doctor is
momentous for an adolescent and discussing smoking is difficult, whereas
responding to a computer in a class with your friends is much less
daunting. A previous trial showed that brief advice from a health
professional about smoking given to adolescents has an effect similar to
that observed in adults(7). The second is human interaction. In our
trial, adolescent s did not discuss their own smoking behaviour. In
Hollis et al, nearly everyone had an additional brief discussion with a
counsellor in addition to the computerised intervention and the results of
the computer sessions were also discussed. Taken together, these trials
therefore point towards the issue of context and human interaction as key
contributors to the efficacy of this intervention package. However,
interpreting these results would be easier if more data were available and
it is. Redding and Velicer were involved in a clinical trial of the
expert system in school pupils, comparable with ours. The trial was
completed in 2000(8), but the effects of the intervention have not yet
been reported.
Reference List
(1) Hollis JF, Polen MR, Whitlock EP, Lichtenstein E, Mullooly JP,
Velicer WF et al. Teen reach: outcomes from a randomized, controlled trial
of a tobacco reduction program for teens seen in primary medical care.[see
comment]. Pediatrics 2005; 115(4):981-989.
(2) Aveyard P, Cheng KK, Almond J, Sherratt E, Lancashire R,
Lawrence T et al. A cluster-randomised controlled trial of an expert
system based on the transtheoretical ("stages of change") model for
smoking prevention and cessation in schools . BMJ 1999; 319:948-953.
(3) Aveyard P, Sherratt E, Almond J, Lawrence T, Lancashire R,
Griffin C et al. The change-in-stage and updated smoking status results
from a cluster-randomized trial of smoking prevention and cessation using
the transtheoretical model among British adolescents. Prev Med 2001;
33:313-324.
(4) Aveyard P, Markham WA, Almond J, Lancashire E, Cheng KK. The
risk of smoking in relation to engagement with a school-based smoking
intervention. Social Science & Medicine 2003; 56:869-882.
(5) Prochaska JO. Stages of change model for smoking prevention and
cessation in schools. BMJ 2000; 320:447.
(6) Velicer WF, Prochaska JO, Fava JL, Laforge RG, Rossi JS.
Interactive versus noninteractive interventions and dose-response
relationships for stage-matched smoking cessation programs in a managed
care setting. Health Psychol 1999; 18(1):21-28.
(7) Walker Z, Townsend J, Oakley L, Donovan C, Smith H, Hurst Z et
al. Health promotion for adolescents in primary care: randomised
controlled trial. BMJ 2002; 325(7363):524.
(8) Plummer BA, Velicer WF, Redding CA, Prochaska JO, Rossi JS,
Pallonen UE et al. Stage of change, decisional balance, and temptations
for smoking. Measurement and validation in a large, school-based
population of adolescents. Addict Behav 2001; 26:551-571.
Conflict of Interest:
None declared
eLetters: ![[Read eLetters]](/icons/misc/sectionDOWN.gif){kind=icon}
