Joseph 2004.

Study characteristics
Methods	Design: Cluster‐randomized controlled trial Setting: 20 Veterans Affairs Medical Centers, USA Recruitment: Calls to patients who visited primary care provider in the past 6 weeks
Participants	Pre‐intervention 4254 adults who smoked who had visited their primary care provider in the past 6 weeks. av.age 64, 95% M, av. cpd not reported Post‐intervention 1424 adults who smoked who had visited their primary care provider in the past 6 weeks. av.age 64, 97% M, av. cpd not reported 575 (280 in the intervention group and 295 in the control) participants made up the cohort of people who smoked and who were contacted both pre‐and post‐intervention
Interventions	Intervention: • Providers (the site‐based principal investigator and 1 other key advocate from each site) received a 2‐day training meeting ‐ Emphasis on options for increasing identification of people who smoked in the computerized patient record system ‐ Promotion of treatment in the primary care setting rather than use of referral‐based care ‐ Encouragement of removal of formulary restrictions to prescription of smoking‐cessation aids and provision of materials to address Pharmacy and Therapeutics Committees • Sites were visited by the interventionist for 2 or 3 days to provide academic detailing of the implementation strategies that was sensitive to local hurdles Control: no information provided on the care this group received or did not receive
Outcomes	PPA at 12m Validation: None Quit attempts Measures of provider implementation: Ask, Advise, Assist and Assist‐Meds
Funding Source	Grant from the Veterans Administration Health Services Research and Development Service: CPG 97‐039
Author's declarations of interest	Not reported.
Notes	Strategy: Provider training + EMR prompts + outreach facilitation Level: Provider + Practice Comparison type: Multicomponent vs. standard care
Risk of bias
Bias	Authors' judgement	Support for judgement
Sequence Generation	Unclear risk	No details reported.
Allocation concealment	Unclear risk	No details reported.
Blinding of outcome assessors All outcomes	Unclear risk	Smoking status was self‐reported and the number of contacts in the control group is not reported
Incomplete outcome data All outcomes	Low risk	At site level, there was no loss to follow‐up (n = 0/20) at 1 year. At patient level, it was not the intention of this study to follow up the same participants from the outset. Patients were randomly selected at baseline and at 1 year and surveyed.
Recruitment bias (cluster RCTs only)	Unclear risk	Participants were affiliated with the practice before randomization. QUOTE: "...among a sample randomly selected from patients who had seen their primary care provider within 6 weeks..."
Balanced baseline characteristics? (cluster RCTs only)	Low risk	QUOTE: "There were no significant differences between subject characteristics in the 2 treatment groups"
Adjustment for clustering in analysis? (cluster RCTs only)	High risk	No adjustment for cluster nature of data reported