Drawing conclusions from network meta-analysis




Poster session 2 Thursday: Evidence synthesis - methods / improving conduct and reporting


Thursday 14 September 2017 - 12:30 to 14:00


All authors in correct order:

Yepes-Nuñez J1, Brignardello-Petersen R1, Bonner A1, Schunemann H1, Guyatt G1
1 McMaster University, Canada
Presenting author and contact person

Presenting author:

Juan Yepes-Nuñez

Contact person:

Abstract text
Background: Systematic reviewers having conducted a network meta-analysis (NMA) face a daunting task of succinctly and judiciously drawing conclusions from evidence that includes estimates of the relative effectiveness of all possible pairwise comparisons (with point estimates and confidence or credible intervals), an assessment of the certainty of each of these estimates (also known as quality of the evidence), and rankings. To date, guidance on how to combine all this information to draw appropriate conclusions about the relative merit of the management options under consideration remains limited.

Objectives: To present an initial approach to draw conclusions from NMA, combining the estimates of effect, rankings, and certainty in the evidence.

Methods: The GRADE project group on NMA is developing an approach to integrate all the information to draw conclusions from NMA. Brainstorming sessions will be followed by the presentation of the approach to other experts during research meetings. The feedback from these experts will be incorporated and the approach will modify and tested in other examples. We will repeat these steps as many times as necessary.

Results: Conclusions thus far include the need to assess rankings in the context of certainty of the evidence. In particular, high ranking of a particular treatment may be misleading if the contributing evidence is of low or very low certainty. We will present further insights regarding how to draw conclusions from NMA, and will discuss the strengths and limitations of our approach. We will provide guidance on how to draw conclusions from NMA, and to get feedback on our work.

Conclusions: Making appropriate conclusions from NMA requires considering all the pieces of information. Guidance on how to combine the effects estimates, ratings of certainty of the evidence, and rankings, is crucial.