Introducing RAPTOR: RevMan Parsing Tool for Reviewers

Systematic Reviews

  ^{1    1    2    2} 

[1] 0000 0001 0601 6589, grid.21051.37, Fakultät Gesundheit, Sicherheit, Gesellschaft, Hochschule Furtwangen University, Robert-Gerwig-Platz 1, 78120, Furtwangen, Germany;0000 0004 1936 8868, grid.4563.4, University of Nottingham, Nottingham, UK;0000 0004 1936 8868, grid.4563.4, Institute of Mental Health, University of Nottingham Innovation Park, Jubilee Campus, Triumph Road, NG7 2TU, Nottingham, UK;

关键词: RevMan;    Automation;    Systematic reviews;    Data extraction;    Document classification;    Automatic document classification;    Review Manager;    Natural language processing;    NLP;    XML;   

DOI  :  10.1186/s13643-019-1070-0

来源: publisher

【 摘 要 】

BackgroundMuch effort is made to ensure Cochrane reviews are based on reliably extracted data. There is a commitment to wide access to these data—for novel processing and/or reuse—but delivering this access is problematic.AimTo describe a proof-of-concept programme to extract, curate and structure data from Cochrane reviews.MethodsOne student of Applied Sciences (16 weeks full time), access to pre-publication review files and use of ‘Eclipse’ to create an open-access tool (RAPTOR) using the programming language Java.ResultsThe final software batch processes hundreds of reviews in seconds, extracting all study data and automatically tidying and unifying presentation of data for return into the source review, reuse, or export for novel analyses.ConclusionsThis software, despite being limited, illustrates how the efforts of reviewers meticulously extracting study data can be improved, disseminated and reused with little additional effort.

【 授权许可】

CC BY