Abstract

Complex evidence synthesis (e.g. Ades & Sutton 2006) is increasingly being employed in epidemiology to combine diverse data sources, with the aim of estimating key characteristics of infectious diseases, such as prevalence and incidence. Often this is carried out in a Bayesian framework, to ease formulation of complex models and incorporate prior knowledge. Model criticism in complex evidence synthesis is a challenging and expanding area of research. I will give a broad introductory (but by no means exhaustive) overview of the literature in this area, discussing inconsistency and conflict in evidence, modelling bias, identifiability, model comparison and sensitivity analysis. I will illustrate some of the ideas with examples from my own work on HIV and pandemic influenza.

A.E. Ades, A.J. Sutton Multiparameter evidence synthesis in epidemiology and medical decision-making:current approaches. JRSS (A), 169, 5—35, 2006.