'Risk of bias' assessment of Joyner 2020
Domain	Assessed outcomes	Authors' judgement	Support for judgement
Representative study group (selection bias)	Not available	Low	Large population size, prospective study, interim analysis of first 5000 patients
Outcome detectors blinded to intervention (detection bias)	Adverse events	High	Not blinded, awareness of intervention can bias assessment of subjective outcomes
Complete outcome assessment/follow‐up (attrition bias)	Adverse events	Unclear	Preliminary results; only serious adverse events assessed, 4‐h follow‐up
Well‐defined study group (reporting bias)	Adverse events	Unclear	Study population well described, intervention scarcely described
Well‐defined outcome (reporting bias)	Adverse events	Unclear	Preliminary results; serious adverse events only
Well‐defined risk estimates (analyses)	Adverse events	Not available	No analyses performed
Important prognostic factors or follow‐up taken adequately into account (confounding)	Not available	High	Not adjusted for confounding factors