The Association Between Altmetric Attention Scores and Public Engagement in the Medical Literature.

INTRODUCTION

With the advent of social media and the associated increase in connectivity between scientists and the lay public, the Altmetric Attention Score has been created as a way to measure these interactions between scholarly publications and media dissemination. Little is known, however, whether these types of media exchanges measured by Altmetrics may serve as a proxy for public engagement. As such, we have sought to determine whether or not an association exists between Altmetric scores and public engagement, as measured by article citation in a health policy document.

METHODS

The top 100 highest scoring articles in the medical and health sciences with respect to Altmetric Attention Scores were selected from each of 3 y (2014, 2015, and 2016). Each article was then matched to an article from the same year and journal with the highest Relative Citation Ratio (RCR) for comparison. Bivariate analysis compared article groups with respect to citation in a public policy document, open-access status, and funding status, as well as Altmetric and RCR scores. A multivariable model was then constructed to identify significant factors associated with citation in a public policy document. Finally, a contour plot was generated in order to estimate the interaction between Altmetric Scores and RCR and their comparative effects on the probability of inclusion in a health policy document.

RESULTS

Of the 600 articles included in the analysis, 286 (48%) had been cited by a public policy article. The only difference that existed between the cohorts was for funding status, with 55 articles (40%) in the RCR cohort having received funding compared to 81 (60%) in the Altmetric cohort (P = 0.011). On bivariate analysis, both Altmetric (P = 0.0018) and RCR (P < 0.0001) scores were independently predictive of policy citation. In a multivariable model, the interaction between Altmetric Scores and RCR with respect to policy inclusion was significant (OR = 1.22; 95% CI = 1.08-1.38) and a contour plot demonstrates that either high Altmetric score or RCR alone is sufficient to generate a high probability of policy inclusion.

CONCLUSIONS

Scholarly article Altmetric Scores may serve as a novel means to explore public engagement in scientific research and health policy. In addition, journals that aim to impact public policy through article dissemination may benefit from engagement in social media avenues in addition to traditional citation pathways in order to encourage broader inclusion.