Meyer 2016.

Methods	Setting: Germany, population‐based Recruitment: nationwide random sample of general population using a random digit‐dialling procedure
Participants	1462 daily smokers with no intention to quit in the next 6 months; 48.5% female, average age 39.4, average cpd 19.8
Interventions	∙ Abstinence Intervention ‐ 3 computer‐tailored counselling letters and self‐help manuals that targeted smoking abstinence, sent just after baseline and at 3 and 6 months. Letters were tailored according to the principles of the transtheoretical model (TTM) of behaviour change and were generated by a fully automated computer expert system ∙ Reduction Intervention ‐ 3 counselling letters and self‐help manuals that targeted reduced smoking, sent at baseline and at 3 and 6 months. Letters were tailored in the same way as in the abstinence intervention, but specifically for reduction of smoking as the target behaviour ∙ Minimal assessment ‐ assessment only
Outcomes	Strictest: 6 months' continuous abstinence at 24 months Other: 12 months Validation: none
Notes	Funding: "Funding was gained from the German Federal Ministry of Education and Research (grant no. 01EB0120, 01EB0420, 01EE1406F) and the Social Ministry of the State of Mecklenburg‐Vorpommern (grant no. IX311a 406.68.43.05)" Declaration of interest: "The Project is part of the German research network EARLINT (EARLy substance use INTervention) and was supported by the research consortium on addiction, AERIAL. Funding was gained from the German Federal Ministry of Education and Research (grant no. 01EB0120, 01EB0420, 01EE1406F) and the Social Ministry of the State of Mecklenburg‐Vorpommern (grant no. IX311a 406.68.43.05). None of the authors have other relevant financial disclosures"
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "The participants were randomized to the three study conditions via a computer‐based procedure. To increase the power of the comparisons between both intervention groups, we used a disproportional randomization algorithm (Dumville et al., 2006), setting the allocation probability to 36.8% for each intervention group and 26.4% for the assessment‐only control group"
Allocation concealment (selection bias)	Low risk	Not specified, but no face‐to‐face contact with researchers
Blinding (performance bias and detection bias) All outcomes	Unclear risk	Not specified, but no objective measure of results
Incomplete outcome data (attrition bias) All outcomes	High risk	Quote: "The number of participants lost to follow‐up rate was significantly higher (at month 12: chi2‐test, df = 2, p < 0.001; at month 24: chi2‐ test, df = 2, p < 0.001) in the intervention groups (at month 12: 23% reduction group, […] at month 24: […] 27% abstinence group) compared with the assessment‐only control group (at month 12: 9%; at month 24: 16%)"