Guide to Conducting Meta-Analysis

Before knowing the systematic steps to meta-analysis, let us first know what it means by Meta-analysis.

What is a meta-analysis?

In simpler terms, meta-analysis is a quantitative study comparing the results of two or more different primary studies with conflicting results using a statistical procedure. It is used to establish a statistical significance.

Steps to meta-analysis:

There are four steps to conducting a meta-analysis.

  1. Framing a research question

The first step is to frame a clear and interesting question. As it is a quantitative study, the PICO framework is being used to formulate the question (where PICO stands for Population, Intervention, Control, and Outcomes).

  1. Searching the literature

After framing the question, the next step is to search all the databases to find a sample that is similar to our study. The most relevant way is to use a keyword search as it yields almost accurate results. Those results will help us to understand our topic much better.

Inclusion and Exclusion criteria

After finding out a list of studies, the researcher now decides which studies will be included and excluded from the process of meta-analysis.

  1. Choosing a meta-analytical method

There are four meta-analytical methods, three fixed-effect methods, and one random-effects method. Choosing the correct meta-analytical method should depend on the framed research question.

The three fixed-effect methods are

  • Mantel-Haenszel method – This test is also known as the Cochran–Mantel–Haenszel test (CMH) and is used in the analysis of stratified or matched categorical data.
  • Peto Method – This method can only be used to pool odd ratios. It is a sum of ‘O-E’ statistics where ‘O’ stands for the observed number of events and ‘E’ for the expected number of events.
  • Inverse Variance Method – In the inverse variance method, the weight given to each study is the inverse of the variance of the effect estimate.

The random-effects method is an assumption that the observed data can vary across studies.

  1. Finding out the result

The last step is to find out the results using one of the meta-analytical methods and report them.

While reporting the results, the researcher uses tables and figures that include all the information showing effect sizes, number of observations, errors, etc. Along with that, whatever data is present in the tables and figures should also be explained in the results or discussion section. The researcher should always mention the heterogeneity measures so that the readers don’t find it false.

Conclusion

This editorial makes you understand what meta-analysis is, along with the basic step-by-step process to conduct a meta-analysis. The methods here are a general structure that includes framing a research question, searching the literature, choosing the right meta-analytical method, and at last, finding out the result and reporting it.