Risk of bias table for MRPs following studies after approval by a REC.

*Easterbrook 1992 reported publication rates for completed and approved studies separately. We just refer to the completed sample in this review. NA: Not applicable.+ means low risk of bias; - means high risk of bias;? means unclear risk of bias.Follow-up time:>24 months after study completion: +. -. MRPs which judged their follow- up rather on approved than completed studies. Although these MRPs fulfilled the 24 month follow-up criteria, we judged them to have an unclear risk of bias:?.Methodology used to identify journal publication: electronic search and author contact: +. only author contact (with response rate of ≥80%): +. only author contact (with response rate -; only database search (in 1 database): -. only database search (in>1 database): +. methodology not given:?.Adjustment for confounders: if an analysis for factors associated with journal publication was carried out: +. if no analysis was carried out: -.Matching criteria: if ≥2 matching criteria given: +. if -. matching criteria not given in MRP:?. if only author contact was used to identify journal publication: NA (not applicable).Research status: completed protocols: +. approved protocols: -.Sampling method: all trials, random sample or consecutive trials: +. if only investigator responded to questionnaire for this sample: -. sampling method not given (e.g., without the word “all”):?.Risk of bias table for MRPs following studies after approval by a REC.