Poor reporting of search strategy and conflict of interest in over 250 narrative and systematic reviews of two biologic agents in arthritis: a systematic review.

TitlePoor reporting of search strategy and conflict of interest in over 250 narrative and systematic reviews of two biologic agents in arthritis: a systematic review.
Publication TypeJournal Article
Year of Publication2009
AuthorsRoundtree AK, Kallen MA, Lopez-Olivo MA, Kimmel B, Skidmore B, Ortiz Z, Cox V, Suarez-Almazor ME
JournalJournal of clinical epidemiology
Volume62
Issue2
Pagination128-37
Date Published2009 Feb
ISSN1878-5921
KeywordsAntibodies, Monoclonal; Antirheumatic Agents; Arthritis, Rheumatoid; Conflict of Interest; Etanercept; Humans; Immunoglobulin G; Infliximab; Journal Impact Factor; Meta-Analysis as Topic; Peer Review; Publishing; Receptors, Tumor Necrosis Factor; Review Literature as Topic
AbstractOBJECTIVE: To evaluate the quality of reviews about etanercept (ETN) and infliximab (IFX), two biologic treatments for rheumatoid arthritis (RA). STUDY DESIGN: A comprehensive, systematic review, including searches of MEDLINE, EMBASE, and other electronic databases and hand-searches for published and unpublished literature. Two raters independently examined each article and identified systematic reviews as those including either a description of: (1) sources for identification and data retrieval; or (2) search strategy. They applied the quality of reporting of meta-analyses (QUOROM) instrument to systematic reviews. RESULTS: Of 3,620 total citations, 281 were identified as reviews. Of these, 26 (9%) qualified as systematic rather than narrative. Overall, few reviews described selection of sources, critical appraisal, or quantitative summary or synthesis. Systematic reviews most often failed to explain validity assessment. Several articles did not disclose authors' participation in industry-funded clinical trials. Most reviews published in high impact factor and rheumatology journals did not meet many quality standards. Significant associations existed between review type (narrative vs. systematic) and reported funding (P=0.05), conflicts of interest (P=0.005), and country of publication (P<0.0001). CONCLUSION: More than 90% of the published reviews were narrative and did not report methods and conflicts of interest in sufficient detail, raising concerns about selection and reporting bias.
DOI10.1016/j.jclinepi.2008.08.003
Alternate JournalJ Clin Epidemiol