Avorn 1983.

Methods	Study design: RCT  Unit of allocation: physicians Stratification by:  geographic location Type of comparison: PEM only vs. nothing: group A: no information group B: mailed print material group C: face‐to‐face group Groups considered in review: A and B
Participants	Physicians Clinical speciality: not clear Level of training: fully trained Setting/country: not clear/US
Interventions	All members of the "print‐only" group received a letter announcing a pilot drug‐information programme.  Half of this group then received a mailed copy of a PEM patterned after the Federal Drug Administration Drug Bulletin ("bulletin" ‐ 3 issues, mailed twice each) describing alternatives to targeted drugs. The other half of the "print‐only" group received these bulletins as well as 6 PEM ("unadvertisement") printed in colour, with illustrations and references
Outcomes	1 process outcome: mean number of units prescribed / physician (all 3 drugs)
Notes	‐
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote, pg. 1460: "control and experimental interventions (described above) were then allocated randomly within each block" COMMENT: method of randomisation is not specified
Allocation concealment (selection bias)	Low risk	Quote, pg. 1460: "control and experimental interventions (described above) were then allocated randomly within each block"
Baseline characteristics similar (selection bias)	Unclear risk	Quote, pg. 1460: "the physicians in each of the study groups were comparable before the intervention in terms of the amount of the target drugs they prescribed through Medicaid, their type of specialty and their board certification" COMMENT: there are no data tables provided, neither is raw data provided in the text
Baseline outcome measurements similar (selection bias)	Low risk	Quote, pg. 1460: "the model thus controlled for differences in preintervention prescribing levels among individual physicians as well as for prescribing trends within the control group"
Incomplete outcome data (attrition bias) All outcomes	Low risk	While the authors do not give specific group‐by‐group drop‐out information, quote pg. 1460: "the dropout rates for each cause were found to be approximately equally divided among the three groups", total drop‐out was 5% overall (see drop‐out rates: pg. 1460, right column, first paragraph)
Blinding of outcome assessment (detection bias) All outcomes	Low risk	The outcome was objective
Contamination protection (contamination bias)	Low risk	Quote, pg. 1460: "If a small town contained more than one physician from our sample, all physicians in that town were randomized as a cluster to prevent cross‐contamination of information"
Selective reporting (reporting bias)	Low risk	All relevant outcomes in the methods section were reported in the results section
Other bias	Low risk	There was no evidence of other risks of bias