Format

Send to

Choose Destination
J Clin Epidemiol. 2019 Jan;105:60-67. doi: 10.1016/j.jclinepi.2018.08.022. Epub 2018 Sep 22.

GRADE approach to rate the certainty from a network meta-analysis: avoiding spurious judgments of imprecision in sparse networks.

Author information

1
Department of Health Research Methods, Evidence and Impact, McMaster University, 1280 Main St W, Hamilton, ON L8S 48L, Canada.
2
Evidence-Based Practice Center, Mayo Clinic, 200 1st Street SW, Rochester, MN 55905, USA. Electronic address: Murad.mohammad@mayo.edu.
3
Department of Health Research Methods, Evidence and Impact, McMaster University, 1280 Main St W, Hamilton, ON L8S 48L, Canada; Department of Family and Community Medicine, Schwartz/Reisman Emergency Medicine Institute, University of Toronto, 200 Elizabeth Street, Toronto, ON M5G 2C4, Canada.
4
Department of Health Research Methods, Evidence and Impact, McMaster University, 1280 Main St W, Hamilton, ON L8S 48L, Canada; Evidence-Based Dentistry Unit, Faculty of Dentistry, Universidad de Chile, 200 1st Street SW, Rochester, MN 55905, USA.
5
Department of Health Research Methods, Evidence and Impact, McMaster University, 1280 Main St W, Hamilton, ON L8S 48L, Canada; Department of Medicine, McMaster University, 1280 Main St W, Hamilton, ON L8S 48L, Canada.
6
Department of Medicine, UHN and Mt Sinai Hospital, 200 Elizabeth Street, Toronto, ON M5G 2C4, Canada; Institute of Health Policy, Management and Evaluation, University of Toronto, 4th Floor, 155 College St, Toronto, ON M5T 3M6, Canada.

Abstract

When direct and indirect estimates of treatment effects are coherent, network meta-analysis (NMA) estimates should have increased precision (narrower confidence or credible intervals compared with relying on direct estimates alone), a benefit of NMA. We have, however, observed cases of sparse networks in which combining direct and indirect estimates results in marked widening of the confidence intervals. In many cases, the assumption of common between-study heterogeneity across the network seems to be responsible for this counterintuitive result. Although the assumption of common between-study heterogeneity across paired comparisons may, in many cases, not be appropriate, it is required to ensure the feasibility of estimating NMA treatment effects. This is especially the case in sparse networks, in which data are insufficient to reliably estimate different variances across the network. The result, however, may be spuriously wide confidence intervals for some of the comparisons in the network (and, in the Grading of Recommendations Assessment, Development, and Evaluation approach, inappropriately low ratings of the certainty of the evidence through rating down for serious imprecision). Systematic reviewers should be aware of the problem and plan sensitivity analyses that produce intuitively sensible confidence intervals. These sensitivity analyses may include using informative priors for the between-study heterogeneity parameter in the Bayesian framework and the use of fixed effects models.

KEYWORDS:

GRADE; Network meta-analysis; certainty; clinical practice guidelines; evidence-based medicine; imprecision; meta-analysis; quality of evidence

PMID:
30253217
DOI:
10.1016/j.jclinepi.2018.08.022
[Indexed for MEDLINE]

Supplemental Content

Full text links

Icon for Elsevier Science
Loading ...
Support Center