HomeResourcesSeeing the forest for the trees with Cochrane South Africa: Forest plots...

Seeing the forest for the trees with Cochrane South Africa: Forest plots and systematic reviews

Here’s a crash course for understanding meta-analyses and systematic reviews from Cochrane South Africa and the South African Medical Research Council.  


Resource details:

Publication title: Evidence-informed health media reporting series: how to navigate a systematic review
Author(s): Cochrane South Africa and the South African Medical Research Council
Publication date: 11 September 2020

What the webinar is about:

Cochrane South Africa’s evidence-informed health media reporting webinar series continues with this third installment on systematic reviews and forest plots. The webinar series is presented in partnership with the South African Medical Research Council, and Stellenbosch University’s Centre for Evidence-based Health Care. The webinar explained why systematic reviews are useful and how they’re compiled. The session also outlined the difference between narrative and systematic reviews and provided a brief introduction to forest plots. This third webinar was presented to the Bhekisisa team by Jimmy Volmink, the Dean of the medicine and health sciences faculty at Stellenbosch University, and Solange Durão, a senior scientist at Cochrane South Africa. 

Key take-aways from the webinar:

Systematic reviews

  • Systematic reviews help researchers assess the data and results of studies that answer the same question in a structured way.
  • How are systematic reviews set up? First, researchers methodically gather data from research that answers the question they’re investigating. The results from the research are then evaluated using a set of predetermined criteria.
  • These reviews can either be qualitative — such as literature or narrative reviews — or quantitative, where statistical data and methods are used.
  • At the heart of systematic reviews is the idea that science is cumulative. Systematic reviews allow us to evaluate the results of a given study in the context of research that has been done before.
  • It can be harmful when researchers fail to recognise the body of knowledge that a study forms part of. For example, doctors might keep administering medicine that is already proven to have serious side-effects. Or, without the proper knowledge, doctors might unnecessarily withhold medication that could help patients. 
  • Drawing similar research together using systematic reviews can help cope with information overload. 
  • Systematic reviews also help identify the gaps in existing knowledge. 
  • These reviews help researchers avoid publication bias — where studies with negative or disappointing results aren’t published. They can reduce database bias — where research isn’t comprehensive and only available or convenient databases are used, and personal bias — where the researcher’s own views or perspectives influence the selected studies. 

Forest plots

  • Forest plots are a way of visualising the results from meta-analyses.
  • These reviews systematically select studies which answer a specific question and whose results can be statistically combined. Forest plots provide a summary effect of all the included studies. 
  • Studies included in the meta analysis must have similar populations, interventions, comparisons and outcomes.
  • The effect of an individual study is graphically summarised by the small square with a horizontal line — or whiskers. The shorter the whiskers the more confident researchers are about what the study’s data is telling us.
  • The diamond at the bottom of the plot represents the combined results of the data. This captures all of the included studies. Because it captures more than one dataset, it is considered more reliable. The more data researchers have the more confident they can be. 

Additional resources:

The Cochrane team shared the following resources with us. You might find them useful too.
Key concepts for assessing claims about treatment effects
A beginner’s guide to interpreting odds ratios, confidence intervals and p-values
How to read a forest plot

Watch a recording of the webinar here.
Download Jimmy Volmink’s presentation here.
Download Solange Durão’s presentation here. 

[Please note: Information on the new coronavirus is rapidly changing. Please refer to Cochrane South Africa and the South African Medical Research Council’s websites for the latest information. Visit www.sacoronavirus.co.za for updates on South Africa’s coronavirus response.]

RELATED ARTICLES
MORE FROM AUTHOR