Avoiding misleading reporting on systematic reviews & meta-analyses

Slides from a short presentation at the World Conference of Science Journalists in 2019, with links to backgrounder posts.

Transcript

1. Avoiding Misleading Reporting on Systematic Reviews & Meta-analyses World Conference of Science Journalists Lausanne 2 July 2019 Hilda Bastian @hildabast hildabastian.net
2. Studies are a collection of parts of uneven quality & quantity [+ spin]
3. Pitfall 1: Exaggerating the quantity of evidence https://well.blogs.nytimes.com/2016/06/27/putting-breast-cancer-on-a-diet/ 1
4. •  82 studies, published 31 meta-analyses •  84% of those analyses combined 20 or fewer studies – 58% had 10 or fewer •  The most studies in any one meta-analysis was 24 https://academic.oup.com/annonc/article/25/10/1901/2801213
5. Common error: exaggerating the amount of evidence https://well.blogs.nytimes.com/2016/06/27/putting-breast-cancer-on-a-diet/
6. 21 studies with approx 42,000 women
7. 2 . . .
8. Sherrington et al (2019). Exercise for preventing falls in older people living in the community. https://www.cochranelibrary.com/cdsr/doi/10.1002/14651858.CD012424.pub2/full 8 studies with fewer than 2,700 people
9. From the Summary of Findings table:
10. Pitfall 2: Overestimating the strength of evidence
13. From 2008 Cochrane review: https://doi.org/10.1002/14651858.CD001321.pub4
14. From 2008: Update in 2012:
15. 5 tips for understanding data in meta-analysis: bit.ly/5TipsData 5 key things to know about meta-analysis: bit.ly/MetaBasics Another 5 things to know about meta-analysis: bit.ly/MetaTraps Systematic reviews & meta-analyses 5 step checkup: bit.ly/MetaCheckup