The predictive validity of quality of evidence grades for the stability of effect estimates was low: a meta-epidemiological study

The limited predictive validity of the EPC approach to GRADE seems to reflect a mismatch between expected and observed changes in treatment effects as bodies of evidence advance from insufficient to high QOE.