O'Keefe 2022
Appraisal of: O'Keefe H, Rankin J, Wallace SA, Beyer F. Investigation of text-mining methodologies to aid the construction of search strategies in systematic reviews of diagnostic test accuracy-a case study. Res Synth Methods. 2022 Jul 15. https://doi.org/10.1002/jrsm.1593.
Reviewer(s):
Caroline Higgins
Julie Glanville
Full Reference:
O'Keefe H, Rankin J, Wallace SA, Beyer F. Investigation of text-mining methodologies to aid the construction of search strategies in systematic reviews of diagnostic test accuracy-a case study. Res Synth Methods. 2022 Jul 15. doi: 10.1002/jrsm.1593. PMID: 35841125.
Short description:
Using a DTA systematic review project as a case-study, an information specialist investigated the benefits and limitations of sixteen text-mining products. A sample of five key studies meeting the review’s inclusion criteria were used. Text-mining tools were evaluated in several domains: ease of use, potential for introducing biases, identification of useful terms, and new terms impact on search results.
Authors reported text-mining applications identified eleven relevant, previously missed publications, demonstrating the value of text-mining tools in search strategy development. Some drawbacks included lowering the original search strategy’s precision, and issues with input/output formats (e.g., some tools did not support downloading of results; input formats required were not commonly used by information specialists). Of interest to information professionals, authors assigned an overall scoring to each tool. Two applications, Text Analyzer [1] and Yale MeSH Analyzer [2], scored highest with “extremely useful” rating.
[1] The Yale MeSH Analyzer 2022. http://mesh.med.yale.edu/
[2] Text Analyzer. 2022. https://www.online-utility.org/text/analyzer.jsp
Limitations stated by the author(s):
This demonstration project used a small sample of publications to test and the review was in a niche subject area (post-mortem techniques in the perinatal population).
Limitations stated by the reviewer(s):
None.
Study Type:
Single Study
Related Chapters:
Tags:
diagnostic accuracy
text-mining tools