Thread by @kidney_boy

Let's talk about the evidence behind the use of visual abstracts to communicate SCIENCE!!

This field started with @AndrewMIbrahim, the whizz kid from @UMich. He was a surgical resident exploring ways to communicate scientific manuscripts and came up with the visual abstract

opmed.doximity.com/articles/visua…

H/T @TDAungst

He realized that these would be a great way to communicate information on social media and started creating visual abstracts as part of his gig as creative director at Annals of Surgery (what do you mean your journal doesn't have a creative director?) @Atul_Gawande agreed

Ibrahim then put his theory to the test with a prospective trial where articles were tweeted with either the citation and link or citation link and visual abstract. Then after a washout period the groups crossed over.

The results were astounding pubmed.ncbi.nlm.nih.gov/28448382/

But there was an obvious pitfall to the study. Any image increases the impact of a tweet. Was the effect that @AndrewMIbrahim was seeing from the visual abstract or could any picture generate that kind of response?

So @edgarvlermamd, Sandra Oska, and I recreated Ibrahim's study but we used a triple cross-over design in order to separate out the effect of a key figure versus a visual abstract. jmir.org/2020/12/e22327

Sandra, who began the project as a second year medical student at @OUWB and is now applying for internal medicine as an M4, did the heavy lifting, picking the key figures and doing all the tweeting. We are super grateful to @AmJNephrol for cooperating with this study.

Our results were similar to @AndrewMIbrahim in direction, if not in magnitude.

In @AndrewMIbrahim's study, he found a significant increase in people clicking through to the article, we did not find that. Additionally he found 7x effect on impressions and retweets. Our effect size was smaller 2x on views and 5x on impressions.

Some of this may have to do with the fact that visual abstracts are not the shiny new thing anymore and some of the effect may have attenuated as people get used to them. Also @AmJNephrol (H-index 80) is not the Annals of Surgery (H-index 298) scimagojr.com/journalrank.php

If you want to know more about visual abstracts, I recommend the excellent review by @KidneyBea_n and @everlymd sciencedirect.com/science/articl…

Or take a look at @AndrewMIbrahim primer on visual abstracts static1.squarespace.com/static/5854aaa…