Abstract
In evidence synthesis using meta-analysis or network meta-analysis, researchers commonly encounter between-study heterogeneity, which may arise from the differences in the conduct of the study. A random effects model is often considered appropriate for allowing for this variability. However, in the case where there are only limited number of studies available for pooling the results, a fixed effect model assuming this variability does not exist is often used because there are too few studies to estimate the between-study heterogeneity.
NICE new methods and processes manual published in January 2022 states that in the case with few included studies in evidence synthesis it may be preferable to use external information in the form of informative prior distributions to help estimating the between-study heterogeneity.
This talk will introduce the 3-stage elicitation approach allowing for uncertainty to be represented by an informative prior distribution to avoid making misleading inferences in meta-analysis or network meta-analysis. The method is flexible to what judgments an expert can provide and is applicable to all types of outcome measures.
Bio
Kate Ren joined HEDS in 2011 after completing a PhD in Probability and Statistics specialising in Bayesian methods in clinical trial design at the University Of Sheffield. Kate specialises in the development and application of Bayesian method in health economics, and the elicitation of experts' beliefs.
https://www.sheffield.ac.uk/scharr/people/staff/kate-ren