The impact of low back pain systematic reviews and clinical practice guidelines measured by the Altmetric score: Cross-Sectional study

Highlights

• The way to facilitate access to scientific articles is share the results of their studies over the internet.

• In order to measure the attention attracted online, a new score named Altmetric was created.

• There are no studies that have analysed associated factors of Altmetric score for systematic reviews and clinical practice guidelines.

Abstract

Background

Although Altmetric has been widely used by researchers to monitor the audience of their articles, there are no studies that have analysed factors associated with Altmetric score for systematic reviews and clinical practice guidelines.

Objectives

1) To analyse factors that could be associated with Altmetric scores for low back pain systematic reviews and clinical practice guidelines. 2) To describe the characteristics of these articles and their Altmetric scores.

Methods

We searched for all low back pain systematic reviews and guidelines indexed on the Physiotherapy Evidence Database published between 2015 and 2017. We extracted data related to the published paper, the publishing journal, and Altmetric scores.

Results

A total of 66 systematic reviews and 5 guidelines were included. The variable impact factor (independent variable) was associated with Altmetric mentioned score (dependent variable) with a β coefficient of 15.4 (95% CI: 0.97, 29.7) ajusted to all remaining variables. The variable number of citations normalized by year of publication (independent variable) was associated with Altmetric reader score (dependent variable) with a β coefficient of 6.4 (95% CI: 4.03, 8.72) ajusted to all remaining variables. We also found that the majority of the systematic reviews and guidelines were published in English, had a descriptive title, were published as open access, included multicenter studies, and had media release generated by the publishing journal.

Conclusion

Metrics related to the number of citations, such as the impact factor are associated with Altmetric scores.

Introduction

Evidence-based practice is widely known as a decision-making process that involves not only the best scientific evidence but also clinical expertise and patient preferences.1, 2 The use of the best evidence induces the utilization of effective treatments.1, 2 Therefore, it is important that key messages from high quality research be widely delivered to clinicians, patients, and policy makers. Unfortunately, this does not always occur¹ as many clinicians still find difficulties in locating the best available evidence.3, 4 To overcome this issue, many authors and publishers are now sharing the results of their studies on the internet to reach and increase their target readership.5, 6 Currently, the use of social media is probably the most effective strategy to increase visibility of scientific articles. Accordingly, a new score named Altmetric was created to measure the attention publications attract online.⁷

The Altmetric score is composed of two independent scoring systems, the Altmetric mentioned and the Altmetric reader.⁸ The Altmetric mentioned score is calculated by the number of mentions on social media (e.g. Facebook, Twitter); mainstream media coverage; encyclopaedias (e.g. Wikipedia); online platforms (e.g. Faculty1000 and Publication Peer-Reviews); videos (e.g. YouTube); sites on questions and answers (e.g. Q&A stack overflow); and documents (e.g. policy documents). Each of these mentions receive different weights to calculate a total score (e.g. each mention on Facebook counts 0.25 points while a mention on Twitter counts 1.0 point).⁸ The Altmetric reader score is calculated by number of mentions by reference managers such as Mendeley, Connotea and CiteULike. Each of these mentions receives identical weights for all reference managers (i.e. 1.0 point for each mention).

To date, there is one systematic review⁹ and a few original research articles10, 11, 12, 13 demonstrating that number of citations and Altmetric scores are positively correlated (with correlation coefficients ranging between r = 0.30 and 0.61). In addition, Araujo et al.¹⁴ found that, in low back pain clinical trials, the number of citations and the journal’s impact factor were positively associated with Altmetric (with β coefficients of 5.2 and 3.4 points, respectively).¹⁴ These results mean that for every citation received, the Altmetric score is likely to be 5.2 points higher. Similarly, for every point of journal’s impact factor the Altmetric score is likely to be 3.4 points higher.¹⁴ However, there are no studies that have identified the Altmetric scores or factors associated with the Altmetric score for systematic reviews and clinical practice guidelines. The investigation on factors associated with Altmetric scores could bring new insights for authors and journal editors on how to better disseminate research findings to clinicians using the internet.¹

The primary objective of this study was to explore potential factors associated with the publishing journal and the published articles that could be associated with Altmetric scores for low back pain systematic reviews and clinical practice guidelines. The secondary objective of this study was to identify the main characteristics of these articles (impact factor, number of citations normalized by year of publication, language, type of title, type of access, number of centers contributing to the manuscript, and media release generated by the journal). In this study, systematic reviews and clinical practice guidelines for physical therapy interventions for low back pain were chosen by the authors because low back pain has the largest amount of evidence in the field of musculoskeletal health.¹⁵ Additionally, low back pain is extremely prevalent16, 17, 18, 19 and involves high costs.19, 20, 21, 22