Using GRADE to integrate randomised and non-randomised studies in systematic reviews


Workshop session 9: Saturday, 14:00-15:30

Workshop category: 

  • Methods for conducting syntheses (including different evidence, searching and information retrieval, statistics, assessing methodological quality)


Date and Location


Saturday 16 September 2017 - 14:00 to 15:30


Contact persons and facilitators

Contact person:


Holger Schunemann
Rebecca Morgan


Cuello C1, Santesso N1, Guyatt G1, Verbeek J2
1 McMaster University, Canada
2 , Finland
Target audience

Target audience: 

Review authors, and researchers with particular interest in systematic reviews and guideline development methods

Level of difficulty: 

Type of workshop

Type of workshop : 



Objectives: To present methods for the integration of randomised (RS) and non-randomised studies (NRS) in health syntheses by using the Grading of Recommendations, Assessment, Development and Evaluation (GRADE) approach.

Description: RS are considered the best source of evidence for health syntheses and clinical practice guidelines. NRS of interventions evaluating benefits and harms are critical to many areas of evaluation, yet they are commonly disregarded or separated from RS due to confounding and bias. With better methods and tools for the conduction and assessment of risk of bias in NRS (i.e. Risk Of Bias In Non-randomised Studies of interventions [ROBINS-I]) and the increasing use of the GRADE approach to assess the overall certainty in the estimates, more opportunities to integrate NRS with RS are feasible and desirable.
The purpose of this workshop is to present and discuss different strategies on how NRS can be used as replacement, sequential, or complementary evidence for using with a body of evidence of RS in systematic reviews with real simplified examples (obtained from case studies) to practice and discuss the implications of using GRADE criteria, using Evidence Profiles, ’Summary of findings' tables, and the ROBINS-I tool.
We will also discuss different methods (e.g. sensitivity analyses) to test the robustness and implications for future research. This is a hands-on workshop with great opportunities for interaction and learning.