The relationship between citations, downloads and alternative metrics in rheumatology publications: a bibliometric study

Abstract

Objective Scientific journals and authors are frequently judged on ‘impact’. Commonly used traditional metrics are the Impact Factor and H-index. However, both take several years to formulate and have many limitations. Recently, Altmetric—a metric that measures impact in a non-traditional way—has gained popularity. This project aims to describe the relationships between subject matter, citations, downloads and Altmetric within rheumatology. Methods Data from publications in Rheumatology were used. Articles published from 2010 to 2015 were reviewed. Data were analysed using Stata 14.2 (StataCorp, College Station, TX, USA). Correlation between citations, downloads and Altmetric were quantified using linear regression, comparing across disease topics. Relationship between downloads and months since publications were described using negative binomial regression, clustering on individual articles. Results A total of 1460 Basic Science and Clinical Science articles were identified, with the number of citations, downloads and Altmetric scores. There were no correlations between disease topic and downloads (R2 = 0.016, P = 0.03), citations (R2 = 0.011, P = 0.29) or Altmetric (R2 = 0.025, P = 0.02). A statistically significant positive association was seen between the number of citations and downloads (R2 = 0.29, P < 0.001). No correlations were seen between Altmetric and downloads (R2 = 0.028, P < 0.001) or citations (R2 = 0.004, P = 0.445). Conclusion Disease area did not correlate with any of the metrics compared. Correlations were apparent with clear links between downloads and citations. Altmetric identified different articles as high impact compared with citation or download metrics. In conclusion: tweeting about your research does not appear to influence citations.

What are the variables associated with Altmetric scores?

“…The study designs of the included articles were mixed research designs [ 15 – 28 , 30 – 32 ] and randomized controlled trials [ 11 , 29 ]. The research fields of these articles included biomedicine [ 20 ], burn care [ 31 ], ecology and conservation [ 30 ], emergency medicine [ 19 ], engineering and technology, gastroenterology and hepatology [ 26 ], general medicine [ 18 ], joint arthroplasty [ 29 ], medical education [ 15 ], medical and natural sciences [ 25 ], multidisciplinary [ 22 ], oncology [ 24 ], physiotherapy [ 11 , 16 ], plastic surgery [ 17 ], psychiatry [ 23 ], radiology [ 32 ], rheumatology [ 21 ], social sciences and humanities, solid organ transplantation [ 27 ], and spine [ 28 ]. The main objective of the included studies was to assess the association between Altmetric scores and variables such as citation counts (i.e., number of citations), journal impact factor, access counts (considered the sum of HTML views and PDF downloads), papers published as open access, time since publication, and/or press releases generated by the publishing journal.…”

“…Different types of analyses were conducted in the included studies: correlation analysis [ 15 , 17 – 19 , 22 , 23 , 26 – 29 , 31 , 32 ], regression analysis [ 11 , 16 , 21 ], boosted regression trees analysis [ 30 ], principal component analysis, and factor analysis [ 20 , 25 ]. The main results of the included studies demonstrated that the variables citation counts (i.e., number of citations), journal impact factor, access counts (considered the sum of HTML views and PDF downloads), papers published as open access, time since publication, and press releases generated by the publishing journal were associated with Altmetric scores.…”

“… Altmetric reader score are associated to citation counts. 9 Chen et al, 2019 [ 21 ] (Rheumatology) To analyze the relationship between citation counts and Altmetric score. Year of search strategy: between 2010 and 2015.…”

Do you tweet?: trailing the connection between Altmetric and research impact!

“…Some studies found associations between social media activity and citation levels, 4 7 8 while others could not confirm this. 9 10…”

“…An analysis using data from manuscripts published from 2010 to 2015 in Rheumatology did not find a connection between the Altmetric score and citation rate. 9 This discrepancy could be due to the analysis of different journals, more probable however, due to the increasing usage of social media by Rheumatologists in recent years. A survey among young Rheumatologists in 2016/2017 showed that more than 80% were active social media users 13 and during the EULAR meeting 2018 #EULAR2018 was used by over 10 000 tweets.…”