Bearcroft 1994.

Methods	Study design: C‐RCT Unit of allocation: GP practices Type of comparison: PEM only vs. nothing group A: no mailing group B: mailing of guidelines + background information
Participants	Physicians Clinical speciality: general practice/family medicine Level of training: fully trained Setting/country: general practice/UK
Interventions	The PEM consisted of a mailed package including: guidelines for referrals for chest radiography that were advisory only and relevant background information.  Guidelines for referrals for chest radiography were developed after a previous study involving the prospective analysis of 2017 consecutive chest radiograph referrals. The presenting indications were compared with the subsequent radiological findings and those indications with a particularly low yield were identified. These guidelines, therefore, were specifically relevant to local practice and they highlighted those groups of patients in whom, based on the previous study, significant abnormalities were uncommon. They were advisory only and included a general reminder that a good clinical history, together with a presumptive diagnosis, would allow a more helpful, accurate and patient‐specific report
Outcomes	4 process outcomes: x‐ray requests not meeting guideline requirements x‐ray requests with inadequate patient history recorded clinical diagnosis reported smoking history
Notes	‐
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote, pg. 56: "GP practices were allocated using a random number table into either the study or control group"
Allocation concealment (selection bias)	Low risk	COMMENT: the unit of allocation is by GP practice and allocation is performed on all units at the start of the study
Baseline characteristics similar (selection bias)	High risk	No baseline characteristics were reported
Baseline outcome measurements similar (selection bias)	Unclear risk	No information is provided
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	This is not specified: while it is implied by it being a prospective analysis of all GP requests for chest radiography, it is not specified whether any of the records were missing after baseline
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	While an attempt was made to blind the outcome assessors, quote, pg. 56: "the reporter was unaware from which group of GPs the request originated", this was not complete, quote, pg. 56: "the majority of the examinations performed were then reported by one of two radiologists (PWPB and JS)", and no quantification of this "majority" was provided
Contamination protection (contamination bias)	High risk	Quote, pg. 58: "in addition, there may have been crossfertilization between study and control groups as GPs meet professionally and socially. Such an effect would be conservative, leading to a reduction in the overall difference"
Selective reporting (reporting bias)	Low risk	All relevant outcomes in the methods section were reported in the results section
Other bias	Low risk	There is evidence of potential unit of analysis error