Assumptions for creating matrices of evidence to estimate overlap of primary studies in overviews of reviews: Protocol for a meta-research study