ABSTRACT
In meta-analysis, independent studies of the same problem will differ in many ways, leading to statistical variation in results across studies. It is often of interest to investigate whether specific characteristics of the different studies explain such variation. This is often undertaken using the techniques of meta-regression. This chapter introduces meta-regression, describing standard meta-regression models with a single covariate and multiple covariates, and outlining the methods used to estimate the parameters of these models. The methods are illustrated using examples. The relationship between meta-regression and subgroup analyses is discussed, as are practical issues such as selecting appropriate covariates, recognizing the possibility of aggregation bias, and controlling spurious findings due to chance.
