A Crossover Randomized Trial of Visual Abstracts versus Plain-text Tweets for Disseminating Orthopaedics Research

Abstract Background Scientists, surgeons and trainees are increasingly taking an active role on Twitter to find, disseminate and exchange knowledge. The purpose of this study was to determine if peer-reviewed journal articles shared on Twitter using visual abstracts (VAs) improve user engagement compared to plain-text tweets. Methods A two-arm randomized controlled trial with crossover was performed. Manuscripts from Journal of Arthroplasty were allocated to one of two arms and disseminated via the journal Twitter account (@JArthroplasty) as either a text-based tweet or a VA. The primary outcome was online engagement (a composite of retweets, replies and likes) at 7 and 30 days after posting. Univariate analysis for non-parametric and parametric data was performed using Mann-Whitney test or Student t-tests, respectively, alpha was set at 0.05. Results 20 in-press manuscripts were randomized to standard tweets (10), or VAs (10) the same day of online publication. The mean number of engagements was higher in the VA group at seven (412 ± 216 vs 195 ± 133; p=0.016) and 30 days (495 ± 204 vs 244 ±162; p=0.007). After the crossover, similar results were reported. Overall, VAs attracted a significantly greater number of engagements than standard tweets. The majority of engagement for both plain text tweets and VAs occurred shortly after the tweet is posted. Conclusion Online, public engagement with orthopaedic research is generally low. However, when VAs are used to communicate research through social media outlets such as Twitter, the overall research engagement significantly increases compared to plain text tweets.