Crone 2003.
| Methods | Randomized controlled cluster trial | |
| Participants | Country: Netherlands. 26 schools providing lower secondary education. Schools recruited through community health services. 14 of 54 health services provided names of schools. 26 schools were recruited but unclear about number of schools approached 'First grade students' from 154 classes participated (mean age 13 years). 1444 in intervention, 1118 in control at baseline. Minority of the sample were of non‐Dutch ethnicity |
|
| Interventions | As well as usual drug prevention/education programme, the intervention classes received 3 lessons on knowledge, attitude and social influence, followed by class agreement not to start smoking or stop smoking for 5 months. Admission to competition to win a prize dependent on classes completing registration, having < 10% smokers after 5 months and producing a photo expressing the idea of a non‐smoking class. Competition prizes (monetary prizes EUR 220 – EUR 450) available to 6 classes with < 10% smoking and ‘a photo best expressing a non‐smoking class’. Control classes received the usual drug prevention/education programme; in 7 schools this was the national drug education programme |
|
| Outcomes | Baseline (October 1998) 8 months follow‐up (June 1999) 20 months follow‐up (June 2000) Self‐reported smoking behaviour, intentions and attitudes measured at each time point. Smoking defined as including ‘experimenting’, weekly and daily smoking. No biochemical verification of smoking |
|
| Notes | Theoretical basis: Intervention based on ‘social influence model’, which was not described in detail | |
| Risk of bias | ||
| Bias | Authors' judgement | Support for judgement |
| Random sequence generation (selection bias) | Low risk | Schools recruited, then randomized. Coin tossing by independent person. Intervention group had a significantly lower proportion of boys, older participants, and non‐Dutch participants at baseline. These were adjusted for in the reported analyses. No significant difference in smoking prevalence between groups at baseline |
| Allocation concealment (selection bias) | Low risk | Allocation by independent person |
| Blinding of participants and personnel (performance bias) All outcomes | Low risk | Not blinded but unlikely to affect outcome |
| Blinding of outcome assessment (detection bias) All outcomes | Low risk | Students told that the jury of the competition were not informed of the results of the study and that registration for the competition was conducted independently of the study evaluation |
| Incomplete outcome data (attrition bias) All outcomes | High risk | High number lost to follow‐up: in intervention group, 907 participants were lost to follow‐up (1444 at baseline to 537 at 20 months). In control group, 714 participants were lost to follow‐up (1118 at baseline to 404 at 20 months). 1 school dropped out in intervention; 2 schools dropped out in control Statistically significantly different distribution of baseline measures of SES, ethnicity, religion, age and smoking among those who were followed up at 8 months and were lost to follow‐up. Non‐response was higher among smokers, especially in the control group. Similar comparisons were not made at 20 months |
| Selective reporting (reporting bias) | Unclear risk | Insufficient reporting to make a judgement on prespecified outcomes |