Applied meta-analysis with R
The book is built as a thorough, step by step introduction to meta-analysis in the R environment. No prior knowledge of \(R\) is required as the authors present the \(R\) commands in parallel with the statistical notions. Through style and numerous examples it is accessible at an undergraduate and graduate level, being as well a good reference for researchers (biomedical, but not restricted to) dealing with real data. The book is structured in ten chapters and commences with an introduction to the R environment and commands. Technical details such as installing \(R\) and the necessary packages are included. The first chapter also contains introductory notions on simulations and data analysis. The second chapter reviews the research protocol for meta-analysis. Using examples from the biomedical field, the authors highlight the data abstraction and extraction of the research question steps. The third chapter presents fixed and random effects in meta-analysis, focussing on the influence of hypotheses and effect sizes and on weighting schemes (for fixed effects), and on the DerSimonian-Laird estimator of \(\tau^2\) (for random effects). The chapter concludes with a detailed example (the amlodipine trial data) including full discussion on decisions and results. The fourth and fifth chapters present the meta-analysis of binary and continuous data, respectively. The analysis of discrete data is presented using the risk-ratio, risk difference and odds ratios (Woolf's method). The Mantel-Haenszel and Peto methods are also discussed. The analysis of continuous data is presented on a step by step implementation in R (using mainly the meta library).NEWLINENEWLINEThe sixth chapter reviews methods to assess the heterogeneity of the data. First, the authors present methods to quantify the heterogeneity using the \(r^2\), the \(H\) and the \(I^2\) indexes. Next, two examples are discussed extensively: the Cochrane Collaboration Logo data and the Tubeless vs Standard PCNL data. The seventh chapter introduces meta-regression and is built on a comparative analysis (on examples) of meta-regression and weighted regression. The examples which are based on the Ischaemic Heart Disease (IHD) data set and the attention deficit/ hyperactivity disorder (ADHD) data set. The eighth chapter discusses two possible approaches for the analysis of real biomedical data: the individual patient level and the meta analysis, and presents again, through examples, the pros and cons of both. The chapter concludes with a simulation study on continuous outcomes. The nineth chapter continues the comparison by presenting data for rare events. The book concludes with the tenth chapter which overviews other R packages for meta-analysis. Packages for obtaining correlation coefficients or for multivariate meta-analysis are also presented.
- An adjusted random-effects model for binary-data meta-analysis
- Meta-analysis. Methods for health and experimental studies
- Meta-analysis of Binary Data Using Profile Likelihood
- Handbook of meta-analysis
- Working with meta-data from C/C++ code in R: the RGCCTranslationUnit package
- Meta-analysis with R
- Applied meta-analysis with R and Stata
- Relative efficiency of using summary versus individual data in random‐effects meta‐analysis
- A serious flaw in nutrition epidemiology: a meta-analysis study
This page was built for publication: Applied meta-analysis with R
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2836476)
