Kottke 1989.

Methods	Study design: RCT Unit of allocation: physicians Type of comparison: PEM only vs. nothing group A: no intervention group B: reception of educational patient material group C: workshop + patient education materials Groups considered in review: A and B
Participants	Physicians Clinical speciality: general practice/family medicine Level of training: fully trained Setting/country: general practice/US
Interventions	The PEM studied in this report was a smoking cessation manual entitled 'Quit‐and‐Win' that could be used as an instructor's manual, as a self‐help guide, or as 1 part of a comprehensive intervention. The physicians were advised to give a copy to any patient who smoked. They were told that their supply of Quit‐and‐Win booklets would be replenished when required
Outcomes	5 process outcomes: patients have been asked by physician if he/she smokes smoking patients who reported being asked by physician to quit smoking smoking patients who were asked to set a quit date smoking patients who were given a follow up appointment smoking patients who received supportive materials 5 patient outcomes: PEM only vs. % of patients who reported an attempt to quit smoking (more than 24 hours without smoking) duration of smoking cessation (in days) month of quit attempt % of patients who reported not smoking at the time of interview smoking patients who agreed to quit smoking
Notes	2 separate PEM analysis for all 10 points: PEM only vs. no intervention PEM only vs. workshop
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote from correspondence with the author: "believe that we assigned the physicians using a computer random generator"
Allocation concealment (selection bias)	Unclear risk	No information was provided
Baseline characteristics similar (selection bias)	Low risk	TABLE 1 and quote, pg. 2103: "neither the mean age of the physicians, the size of the clinics nor the patient load…differed significantly among the three groups" COMMENT: even if professionals were well balanced, patients did not have all baseline characteristics similar
Baseline outcome measurements similar (selection bias)	Unclear risk	Outcomes were not collected at baseline
Incomplete outcome data (attrition bias) All outcomes	Low risk	The proportion of patient‐smokers was similar between groups, and the percentage reached at 1 year for follow‐up was similar. Quote, pg. 2103:  "patients who either could not be contacted or refused to be interviewed were assumed to be continuing to smoke and were assumed not to have made any cessation attempts"
Blinding of outcome assessment (detection bias) All outcomes	High risk	This is a self‐report assessment by patients who were not blinded
Contamination protection (contamination bias)	Low risk	Quote, pg. 2102: "to prevent contamination from having physicians of the same practice in different trial groups, all physicians in the same practice were either moved to the most intense level of intervention to which any of them had been originally randomized or, if not yet randomized at the time this problem was discovered, added to the group to which their partner(s) had been randomized"
Selective reporting (reporting bias)	Low risk	All relevant outcomes in the methods section were reported in the results section
Other bias	High risk	COMMENT: the primary outcome measure was 102 question questionnaire for patients, making this outcome measure susceptible to LOW validity