THE LEVEL OF COVERAGE OF THE PHYSIOTHERAPY EVIDENCE DATABASE (PEDRO) BY GENERAL BIBLIOGRAPHIC DATABASES

L. Carballo-Costa1, R. Costas2, A. Moseley3
1University of A Coruna, Department of Physiotherapy, Medicine and Biomedical Sciences, A Coruna, Spain, 2Centre for Science and Technology Studies (CWTS). Leiden University, Leiden, Netherlands, 3Institute for Musculoskeletal Health, The University of Sydney and Sydney Local Health District, Sydney, Australia

Background: The Physiotherapy Evidence Database (PEDro) is considered the most important and comprehensive bibliographic database for physiotherapists, selecting and indexing clinical practice guidelines, systematic reviews, and randomised controlled trials evaluating physiotherapy interventions. We present a large-scale comparison of PEDro with other three major bibliographic data sources; two multidisciplinary databases (Web of Science and Dimensions), and one focused on health sciences (PubMed) to study the coverage of articles indexed in PEDro by these databases.

Purpose: The aims of this study are:
  1. To characterise the coverage of articles indexed in PEDro by these bibliographic databases.
  2. To identify which topics identified in articles indexed in PEDro are covered by these other databases.

Methods: All trials, reviews and guidelines indexed in PEDro and published between 1996 and 2018 were included in the analysis (n=38,717). The Digital Object Identifier (DOI) and/or PubMed Identifier (PMID) were extracted to match articles with the Web of Science, Scopus, Dimensions and PubMed databases. PEDro was compared in a pairwise manner with each of the other data sources. Data about the title and abstracts of the articles indexed in PEDro were extracted to study the coverage of the topics. Free VOSviewer software (version 1.6.15) was used for the text mining of title and abstracts, and to create the visualisations.

Results: A total of 35,193 articles with DOI and/or PMID were matched to one or more of the databases. Dimensions indexed the highest number of PEDro articles (35,115; 90.7%), followed by Scopus (33,654; 86.9%), PubMed (33,170; 85.7%), then Web of Science (29,317; 75.7%). The coverage of topics varied between the databases, but those related to back pain, neck pain and stroke were less covered by the other databases.  3,524 (9.1%) of articles with no DOI or PMID (indexed only by PEDro) covered topics related to acupuncture, neurology and pain. The main graphs can be seen in http://doi.org/10.6084/m9.figshare.12735473.

Conclusion(s): Articles indexed in PEDro are not entirely covered by the other databases analysed. The coverage of topics varies slightly between the different databases, with PEDro indexing articles on topics not well covered in other databases.

Implications: This information is useful to clinicians and researchers as they can consider PEDro as a very good source of articles about physiotherapy interventions. However, there are other relevant sources of information that can be considered as complementary, especially for researchers interested in applying different bibliometric and qualitative methodologies. Policymakers could use these data to support decisions about financial support to maintain this fundamental resource in physiotherapy.

Funding, acknowledgements: This study has not received any financial support.

Keywords: Bibliometrics, Databases coverage, Physiotherapy Evidence Database (PEDro)

Topic: Research methodology, knowledge translation & implementation science

Did this work require ethics approval? No
Institution: University of A Coruna
Committee: University of A Coruna
Reason: This is a bibliometric study.


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

Back to the listing