Marteau 2012.

Methods	Design: RCT Country: UK Recruitment methods: participants were recruited through NHS primary care practices. Smokers were identified through practice registers and sent a letter offering assistance to quit and an invitation to participate in the trial. Setting: smoking cessation clinics in primary care
Participants	Inclusion criteria: smoking ≥ 10 cigarettes per day; wanting to quit smoking; aged ≥ 18 years Exclusion criteria: none stated Participants randomised: 633 participants (mean age 47.3 years (SD 13.3); 54.3% women; 90.2% white
Interventions	Aim of intervention: to increase adherence to NRT by informing participants that their oral dose is tailored based on an analysis of their genotype, rather than their phenotype (FTND score). Intervention: communicating different means of tailoring prescribed medication, delivered by trained research nurses. Behavioural support (based on withdrawal‐orientated therapy) and nicotine patches were provided (with the patch dose tailored in relation to cigarettes per day) to all participants. Participants were also prescribed an oral NRT product of their choice. The dose of oral NRT in the intervention arm was tailored based on gene variant. Participants were given both forms of NRT 1‐day prequit and told the basis for their dosage. They were also provided with a personalised booklet and an appointment card documenting the dose of NRT to use daily and giving the reason for the dose. The rationale for the dose was reiterated at each subsequent clinic. Behavioural support was offered twice prior to quit day, weekly afterwards for 4 weeks and then at 8 weeks. Quit day was set 2 weeks and 1 day after baseline. Support sessions lasted 10–30 minutes, depending on progress and stage of quit attempt. Control: same content apart from the dose of oral NRT and the corresponding communication of the rationale was tailored based on FTND score.
Outcomes	Primary adherence outcome (continuous data): proportion of all prescribed NRT taken over 28 days, assessed at 28 days of treatment period (ITT data). Checked by tablet counts of medication used. Other adherence outcomes: proportion of all prescribed NRT taken over 7 days; proportion of participants showing no use of NRT; proportion of participants showing use of NRT beyond 28 days. Secondary outcomes: biochemically validated prolonged abstinence at 28 days; biochemically validated prolonged abstinence at 6 months; anxiety assessed using the short‐form Spielberger State‐Trait Anxiety Inventory‐6.
Notes	Phenotype arm was regarded as control arm as it is more similar to standard care. Funded by Medical Research Council, UK. 1 author reported having completed consultancy and research on smoking cessation for pharmaceutical companies. The remaining authors declared no conflicts of interest.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Random sequence was computer generated (p. 4, paragraph 2). No evidence of systematic differences between arms, with no reported differences in baseline demographic and smoking characteristics (p. 6, Table 1).
Allocation concealment (selection bias)	Low risk	Allocation was conducted from a central isolated location, separate from trial co‐ordination and participant recruitment (p. 4, paragraph 2). The randomisation sequence was revealed sequentially and concealed from the trial team, nurses and participants (p. 4, paragraph 3).
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Outcome assessors for primary outcome were not blinded, but because tablet counts were used it was unlikely that this constituted a clear risk of bias. Outcome assessors for longer‐term follow‐up were blinded to allocation (p. 4, paragraph 3). Primary adherence outcome was checked by tablet counts of medication used (p. 3, paragraph 5). Abstinence outcomes were biochemically validated (p. 3, paragraph 13).
Incomplete outcome data (attrition bias) All outcomes	Low risk	ITT analysis was used and reported with non‐respondents at follow‐up treated conservatively as non‐adherent and continuing smokers (p. 4, paragraph 11). There were no differences in attrition between arms (p. 7, paragraph 3) with numbers being reported, the overall number of participants lost being < 50% and the difference in percentage followed up between groups being < 20%.
Selective reporting (reporting bias)	Low risk	Study was preregistered including specified outcomes and these were unchanged in study report (ISRCTN14352545). This is also clear in a published protocol.
Consistency in intervention delivery	Low risk	Standardised script used, detailed in the published protocol (p. 3, paragraph 3).
Summary risk of bias	Low risk	Summary risk of bias assessed as low.