Salameh

Appraisal of: Salameh JP, McInnes MDF, Moher D, Thombs BD, McGrath TA, Frank R, Dehmoobad Sharifabadi A, Kraaijpoel N, Levis B, Bossuyt PM. Completeness of Reporting of Systematic Reviews of Diagnostic Test Accuracy Based on the PRISMA-DTA Reporting Guideline. Clin Chem. 2019 Feb;65(2):291-301. DOI: 10.1373/clinchem.2018.292987

Reviewers:

Caroline Higgins

Full reference:

Short description:

This study aims to evaluate the quality or “completeness” of reporting in published systematic reviews of diagnostic test accuracy (DTA) using the recently released Preferred Reporting Items for Systematic Reviews and Meta-Analyses diagnostic test accuracy extensions: PRISMA-DTA and PRISMA-DTA for Abstracts. Authors searched MEDLINE to identify DTA systematic reviews published between October 2017 to January 2018. A set of 100 reviews were included in this study and assessed using 26 item PRISMA-DTA checklist (the “full checklist of 27 items minus the item referring to PRISMA-DTA for abstracts”) and the 11 items on PRISMA-DTA for Abstracts checklist. Overall, this study found that recent DTA systematic reviews lacked information readers require to fully understand the validity and generalizability of their results: mean number of items reported was 18.6 of 26 items (71%; SD = 1.9) with a range from 12.0 to 23.0.

Of interest to information specialists, this study found 87 of the 100 reviews reported information sources searched, however suboptimal reporting of date last searched (33/100) and full search strategy (42/100). Regarding PRISMA-DTA for Abstracts, findings indicate a moderate reporting of databases searched (63/100) and infrequent reporting of last date searched (42/100).

Limitations stated by the study author(s):

Authors identified the potential for extractor subjectivity during data extraction (i.e., potential for extractors to employ different thresholds to what could be considered satisfactory reported or not). This acknowledgement points to this study not undertaking an inter-rater reliability analyse to understand the degree of consensus between the six extractors.

Limitations stated by the reviewers:

The search strategy used to identify potential studies for inclusion does not appear to be sensitive or comprehensive: (1) MEDLINE was only information source searched; (2) strategy used was not from a publication reviewed by the ISSG Search Filter Resource.

Study type:

Single study