Are we really including all relevant evidence? Considerations for network meta-analysis

Abstract: Meta-analyses and network meta-analyses typically aim to include all relevant evidence. The purpose of this seminar is to discuss how the design of a (network) meta-analysis can influence the extent to which this is achieved.

The scope of the evidence space and the data admitted for analysis can both be varied to include more information. In particular, I will explore how statistical modelling approaches can be used to increase the amount of data informing treatment effect estimates. Shared parameter models, models for repeated measures and models for multiple endpoints will be examined. Example BUGS code will be demonstrated.