posted on 2015-10-09, 10:59authored bySze Huey Tan, Nicola J. Cooper, Sylwia Bujkiewicz, N. J. Welton, D. M. Caldwell, Alexander J. Sutton
OBJECTIVES: To present graphical tools for reporting network meta-analysis (NMA) results aiming to increase the accessibility, transparency, interpretability, and acceptability of NMA analyses. STUDY DESIGN AND SETTINGS: The key components of NMA results were identified based on recommendations by agencies such as the National Institute for Health and Care Excellence (United Kingdom). Three novel graphs were designed to amalgamate the identified components using familiar graphical tools such as the bar, line, or pie charts and adhering to good graphical design principles. RESULTS: Three key components for presentation of NMA results were identified, namely relative effects and their uncertainty, probability of an intervention being best, and between-study heterogeneity. Two of the three graphs developed present results (for each pairwise comparison of interventions in the network) obtained from both NMA and standard pairwise meta-analysis for easy comparison. They also include options to display the probability best, ranking statistics, heterogeneity, and prediction intervals. The third graph presents rankings of interventions in terms of their effectiveness to enable clinicians to easily identify "top-ranking" interventions. CONCLUSIONS: The graphical tools presented can display results tailored to the research question of interest, and targeted at a whole spectrum of users from the technical analyst to the nontechnical clinician.
Journal of Clinical Epidemiology, 2014, 67 (6), pp. 672-680
/Organisation/COLLEGE OF MEDICINE, BIOLOGICAL SCIENCES AND PSYCHOLOGY/School of Medicine/Department of Health Sciences
The file associated with this record is under a 12-month embargo from publication in accordance with the publisher's self-archiving policy, available at http://www.elsevier.com/about/open-science/open-access. The full text may be available in the publisher links provided above.