Meta-analysis: common misconceptions

In this talk, I explain some common mistakes made when undertaking and interpreting meta-analyses. Firstly, I explain the assumptions behind the commonly used ‘random-effects meta-analysis’ model, and describe why it produces a summary result that may mislead clinical practice. Secondly, I consider the use of meta-regression to estimate ‘effect-modifiers’, which are factors that modify treatment effect. I explain why this can produce misleading conclusions for stratified medicine, due to ecological bias. Finally, I discuss why individual participant data (IPD) meta-analyses should not automatically be viewed as the ‘gold-standard’, because many biases can also affect their conclusions. The talk is intended for a wide audience: real examples will be used throughout to illustrate the issues, and there will be minimal use of equations.

Admission: Free and open to all with no ticket required. Entry is on a first come, first served basis. Speaker(s): Dr. Richard Riley, University of Birmingham