Background and objective: We examined whether the use of visual abstracts on social media platforms can improve comprehensive social media and conventional metrics such as total altmetric attention score, full text page views and citation counts (study outcomes) through retrospective cohort study.
Methods: We included all original research articles (Total 307 articles: N = 69 in visual abstract group and N = 238 in control group) published between July 2018 and January 2019 in the JAMA, BMJ and the NEJM and used negative binomial regression to adjust for article characteristics.
Results: Adjusted analysis showed no significant differences between articles with and without visual abstracts in the altmetric attention score (p = 0.37) and in number of page views (p = 0.44). Citations in the Web of Science core collection were found to be statistically significant favouring control group (p = 0.028). We also found no significant differences in altmetric attention score and page views after stratification for article type [randomised controlled trial (RCT) vs. non-RCT]. Citations counts were found to be borderline significant for RCT (p = 0.04) and non-significant for non-RCT.
Conclusion: Visual abstracts might not be effective in disseminating scientific research. We should look at other innovative ways to improve the visibility of the research.
Keywords: abstracting and indexing; information dissemination; publishers and publishing; research impact; social media; statistics.
© 2021 Health Libraries Group.