Harmonization of explanations for common judgments about the quality of evidence in summary of findings tables

Article type
Authors
Langendam M1, Mustafa RA2, Santesso N2, Carrasco A2, Moustgaard R3, Ventresca M2, Heus P4, Brozek JL2, Lasserson T5, Schünemann HJ2
1The Dutch Cochrane Centre, The Netherlands
2McMaster University, Canada
3Nordic Cochrane Centre, Denmark
4Dutch Cochrane Centre, The Netherlands
5Cochrane Editorial Unit, UK
Abstract
Background: Assessing the quality of the body of evidence (QoE) using the GRADE criteria, justifying and documenting these assessments are desirable for new Cochrane Reviews as part of Summary of Findings (SoF) tables (MECIR standards). In an evaluation of SoF tables, we observed variability in how authors of Cochrane Reviews report their assessments in footnotes and comments. Harmonization of explanations for common judgments is one of the aims of this Methods Innovation Fund project.

Objectives: To provide templates of standardized wording of footnotes and comments in SoF tables in order to enhance their usefulness and facilitate interpretation.

Methods: Standardized footnotes and related guidance are currently being developed based on a database of footnotes and comments from SoF tables published in the Cochrane Library Issue 3, 2012 (n = 502). The footnotes and comments will be aggregated by pre-defined and newly discovered themes. Preliminary pre-defined themes are justification for judgments about each of the GRADE QoE domains (reasons for downgrading and not downgrading as well as upgrading the evidence), judgments across domains, explanations in the face of single RCTs, no meta-analysis and refraining from grading the evidence. The method for evaluating the usefulness of footnotes is in development but will include assessments by experienced GRADE users as well as by a representative sample of users of SoF tables. Selected anonymized examples will be included in training materials and GRADEpro software.

Results: Preliminary results indicate that Cochrane authors used different styles in reporting footnotes and sometimes provide reasons for their judgments that contradict current guidance (e.g. for single study evidence and upgrading of RCTs). Final results will be presented at the Colloquium.

Conclusions: Knowing how authors approach footnotes is a useful indicator of possible areas where more training or guidance is needed, e.g. in the Cochrane and GRADE Handbook and GRADEpro software.