Meta-analysis

Abstract

The previous chapter has discussed how to gather and evaluate existing evidence from epidemiological studies. Here further consideration is given to summarizing the identified evidence in such a way that it can be used for decision-making, including approaches to control for chance and potential bias. Meta-analysis refers to the statistical analysis of results from individual studies for integrating the findings. There are other terms related to meta-analysis such as quantitative review, combined analysis, pooled analysis, or quantitative synthesis. Some of them use different methods, for example, meta-analysis of published data considers each study as a unit of analysis while individual patient data analysis includes the original data from each study on a participant level. This chapter describes how to numerically summarize data through performing a meta-analysis using data from a systematic review of epidemiological studies. It also considers possible bias, reporting guidelines, and statistical software available for meta-analysis.