Application of the systematic review and bibliometric network analysis (SeBriNA) methodology contextualizes evidence. Part 2: rituximab for non-Hodgkin's lymphoma

Abstract

Objective: We conducted a systematic review and bibliometric network analysis (SeBriNA) of rituximab for non-Hodgkin's lymphoma.

Study design and setting: We searched three primary data sources (1997-2003) for five document types: original research, reviews, guidelines, editorials, and media reports. We conducted cumulative meta-analysis on three outcomes (mortality, tumor response, safety) and used GRADE (Grading of Recommendations, Assessment, Development, and Evaluation) to classify evidence quality. Direct citation relationships between original research documents and other documents were analyzed and visually represented.

Results: Of 6,798 documents, 757 met inclusion criteria. The 317 original research documents represented 209 study clusters and 8,483 evaluated patients. Of 209 study clusters, 2.9% were randomized controlled trials (RCTs) and reported data on outcomes of interest. The quality of evidence was moderate. We identified 1,571 direct citations to the 317 original research documents. The first RCT reporting relevant outcomes appeared in 2000, whereas the first guidelines appeared in 1999. Of 212 media reports, 92% cited no original research.

Conclusions: Of 757 rituximab documents, RCTs of comparisons and outcomes represented <3% of original research. In contrast, review articles, guidelines, editorials, and media reports each outnumbered the relevant original research. The SeBriNA review facilitated the analysis, contextualization, and interpretation of these complex relationships.