Meta-Analysis – Definition, Purpose And How To Conduct It

The number of studies being published in biomedical and clinical literature is increasing day by day. The massive abundance of research studies in these areas makes it rather difficult to synthesise all data and accumulate knowledge from various studies. All of this delays certain clinical decisions and conclusions to be made.

To determine the validity of a hypothesis, it is necessary to look into multiple studies rather than just one. For this purpose, systematic reviews or narrative reviews have been used to synthesise data from multiple studies, which often leads to an objective approach as different people can have different opinions. Meta-analysis, on the other hand, provides an objective and quantitative approach to combining evidence from various studies.

What Is Meta-Analysis?

The meaning of meta-analysis is a statistical method of combining results from numerous studies on a certain research question. The term was first used in 1976 and can be used to determine if the effect reported in the literature is real or not.

To conduct a quality meta-analysis, you need to identify an area in which the effect of treatment is uncertain. It is also recommended that you collect as many studies similar to the effect as possible so that you can compare them and get a better picture of it. This assists the researcher in understanding how big or small the effect is, and how different the results are from other studies.

Purpose Of Meta-Analysis In Research

The purpose of meta-analysis is more than just combining results from studies to give a statistical assessment. It also helps to point out:

• Any potential reasons for variations and differences in results, also known as heterogeneity in meta-analysis. Some popular reasons for this might be differences in sample size, or differences in analysis methods in research.
• The real estimate of the effect size that is reported in literature than any individual study. Combining multiple studies reduces research bias and the chances of random errors.

Meta-Analysis In Applied And Basic Research

Basic research involves seeking knowledge and gathering data on any subject, whereas applied research is more experimental and uses methods to solve real-life problems. Meta-analyses are used in both types of research that is applied and basic research.

• Pharmaceutical companies use meta-analyses to gain approval for new drugs, such as antibiotics for bacterial infections. Even regulatory authorities use this research method to gain approval for different processes. Hence, meta-analysis is used in medicine, crime, education and psychology for applied research.
• In terms of basic research, it is used in various fields such as sociology, finance, economics, marketing and social psychology. An example of meta-analysis in basic research is studying the effect of caffeine on cognitive performance.

Strengths and Limitations Of Meta-Analysis

There are many key benefits of meta-analysis in research studies, as it is a powerful tool. Here are some strengths of it:

• A meta-analysis takes place after a systematic review, which means the end product will be reliable and accurate.
• It has great statistical power, as it combines multiple studies rather than one individual study, which might suffer from a lack of sufficient data.
• This analysis can confirm existing research or refute it. Either way, it gives a confirmatory data analysis.
• It is regarded highly in the scientific community, as it provides an objective and solid analysis of evidence.

Challenges Associated With Meta-Analysis

Meta-analysis also has certain challenges that result in limitations while carrying out this statistical quantitative approach. Some challenges faced by meta-analyses are:

• It is not always possible to predict the outcome of a large-scale study. This is because meta-analysis mostly rely on small-scale studies which do not represent the broad population.
• A decent meta-analysis can not make up for flawed or bad research designs. Thus, it can not control the potential for bias to arise in studies. Therefore, it is urged to only include research with sound methodologies known as “best evidence synthesis”.

How To Conduct A Meta-Analysis

Before conducting a meta-analysis and defining the research scope, it is necessary to evaluate the number of publications that have grown over the years. It can be quite hard to scan and skim through a large number of studies and literature reviews, which is why it is necessary to define the research question with care, including only relevant aspects. Here are steps on how to perform a meta-analysis:

• Formulate a research question that showcases the effects or interventions to be studied. This is mostly a binary question, such as “Does drug X improve outcome Y” in clinical studies.
• Conduct a systematic review that analyses and synthesises all data related to the one research question.
• Gather all data such as sample sizes and research methods used to indicate data variability. All decide which dependent variables are allowed.
• The selection of criteria is also a crucial step as it is necessary to understand whether published or unpublished studies are to be included or not. Based on the research question, it is important to choose studies that are quality-based and relevant.
• Choosing the right meta-analytic methods and meta-analysis software to be used in meta-analysis is another significant step. Some methods used are traditional univariate meta-analysis, meta-regression and meta-analytic structural equation modelling methods.
• While evaluating the data, it is necessary to use a meta-analysis forest plot, which is the graphical representation of the results of a meta-analysis studies. Its visual representation helps understand the heterogeneity among studies and helps compare the overall effect sizes of an intervention.
• The final step of literature meta-analysis is to report the results. They should be comprehensive and precise for the reader’s understanding.

Meta-Analysis Vs Systematic Review

A systematic review is a comprehensive analysis of existing research, whereas a meta-analysis is a statistical analysis or combination of results from two separate studies. Here’s how the two differ from each other:

Frequently Asked Questions