The previous two chapters showed that pooling a meta-analysis of separate studies is pretty meaningless, if the results are significantly heterogeneous across studies. Instead, a careful investigation of the potential causes of heterogeneity has to be accomplished. Differences in age groups, co-morbidities, co-medications, gender differences and other characteristics are common causes. Subgroup comparisons are commonly applied for the purpose. However, in the past few years multiple regression analysis has been increasingly used as an alternative approach. The main study outcome is the dependent variable and the potential causes of heterogeneity are the independent variables. Advantages of this method include that the effects of multiple factors can be studied simultaneously, and that confounders and interacting-factors can be adjusted (see also the Chaps. 28 and 30 ). In the present chapter an example was used of a recently published meta-analysis from our group (Atiqi et al. 2009).
CITATION STYLE
Cleophas, T. J., & Zwinderman, A. H. (2012). Meta-regression. In Statistics Applied to Clinical Studies (pp. 391–396). Springer Netherlands. https://doi.org/10.1007/978-94-007-2863-9_34
Mendeley helps you to discover research relevant for your work.