​

Prof Richard Riley gave a talk on using IPD meta-analysis to estimate predictors of treatment effect (treatment-covariate interactions), including demonstration of non-linear relationships. ​The part on IPD meta-analysis starts at 42 mins 42 seconds, and is intended for a broad audience. The corresponding paper is now published [PDF]:

​

​

​

​

​

​

​

​

​

​

​

​

​

Prof Richard Riley  also gave the following more statistical talk about IPD meta-analysis methods, including more technical details about modelling treatment-covariate interactions, was given to the Japanese region of the International Biometrics Society. The corresponding paper is now published [PDF]: