QUALITY, LANGUAGE, SUBDISCIPLINE AND PROMOTION WERE ASSOCIATED WITH ARTICLE ACCESSES ON PHYSIOTHERAPY EVIDENCE DATABASE (PEDRO)

Parma Yamato T.¹, Arora M.^2,3, Stevens M.¹, Elkins M.^4,5, Moseley A.¹

¹The George Institute for Global Health, The University of Sydney, Musculoskeletal Division, Sydney, Australia, ²John Walsh Centre for Rehabilitation Research, Kolling Institute of Medical Research, Northern Sydney Local Health District, Australia, ³Sydney Medical School Northern, The University of Sydney, Sydney, Australia, ⁴Centre for Education & Workforce Development, Sydney Local Health District, Australia, ⁵Sydney Medical School Central, The University of Sydney, Sydney, Australia

Background: Several factors can influence the selection of research articles for answering clinical questions. FUTON bias (Full Text On Net), NAA bias (No Abstract Available), and free access to the full article or abstract have each been identified as possible influences on the selection of research articles when searching for evidence. Factors that may relate to the number of accesses of articles include the research design, year of publication, trial quality, language of publication, area of physiotherapy, and being promoted to users.

Purpose: To quantify the relationship between the number of times articles are accessed on the Physiotherapy Evidence Database (PEDro; www.pedro.org.au) and the article characteristics. A secondary aim was to examine the relationship between accesses and the number of citations of articles.

Methods: All articles available on PEDro from August 2014 to January 2015 were included. We extracted the research design, year of publication, PEDro score, source of systematic review (Cochrane or non-Cochrane), language, subdiscipline of physiotherapy, whether articles were promoted to PEDro users, citation, and journal impact factor. The number of times each article was downloaded or accessed was downloaded from log files and Google Analytics. Three predictive models were examined using multiple regression analysis (all articles, systematic reviews only and clinical trials only). The number of citations was downloaded from Google Scholar for a random sample of 1000 articles.

Results: There were 29,313 articles indexed on PEDro between August 2014 and January 2015. Seven factors predicted more accesses: synthesis research (i.e., guidelines and reviews), recent articles, English language, promotion to PEDro users, being a Cochrane review, and (for trials) having a higher PEDro score. The musculoskeletal, neurology, orthopaedics, sports, and paediatrics subdisciplines were associated with more accesses, while the ergonomics and occupational health subdiscipline was associated with fewer accesses. We also found that there was no association between number of accesses and citations.

Conclusion(s): The number of times an article is accessed on PEDro is partly predicted by the degree of synthesis and quality.

Implications: The PEDro resource can be used to quickly guide users to high-quality clinical research about the effects of physiotherapy interventions. Ranking of search results by the degree of synthesis (guidelines, then reviews, then trials) and quality (Cochrane before non-Cochrane for reviews, descending total PEDro score for trials) may facilitate access to the best and most sought-after evidence.

Funding acknowledgements: This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Topic: Professional issues

Ethics approval: Ethics approval was not required in this study.

All authors, affiliations and abstracts have been published as submitted.