Reliability and applicability of the revised Cochrane risk-of-bias tool for randomised trials (RoB 2): low inter-rater reliability and challenges in application