Visual abstracts (VAs) are graphical representations of the key findings in manuscripts and have been adopted by many journals to improve content dissemination via social media. We sought to assess whether VAs, compared to key figures (KFs), increased reader engagement via social media using articles published in European Urology. We prospectively randomized 200 consecutive new publications to representation on Twitter and Instagram using either a VA (n = 99) or a KF (n = 101). Randomization was stratified by prostate cancer content. The primary outcome was Twitter impressions. Secondary outcomes included Twitter total engagements, link clicks, likes, and retweets, as well as Instagram likes. Analysis of covariance was conducted using the stratification variable as a covariate. We found that Twitter impressions were greater for tweets containing VAs compared to KFs (8385 vs 6882; adjusted difference 1480, 95% confidence interval [CI] 434–2526; p = 0.006). VA use was also associated with more retweets and likes (p < 0.002), but fewer full-article link clicks than KFs (60 vs 105, adjusted difference 45, 95% CI 21–70; p = 0.0004). The choice between VA and KF should depend on the relative value given to impressions versus full-article link clicks.
