Assessing heterogeneity in meta‐analyses of Helicobacter pylori infection‐related clinical studies: a critical appraisal

Abstract

OBJECTIVE: To critically assess the meta-analyses of Helicobacter pylori infection-related clinical studies, particularly the handling of between-study heterogeneity. METHODS: A qualitative, all-language, systematic literature search was performed in Medline, PubMed, BioMed Central and Embase up to February 2003, supplemented by a manual search of major relevant journals. Assessment was according to modified criteria for literature searching, eligibility criteria, validity assessment, data extraction and presentation. Five parameters were used to assess the quality of the meta-analyses in handling between-study heterogeneity. RESULTS: Of 84 potentially relevant citations, 47 were systematic reviews and of them 38 were meta-analyses. Of these 38 studies, 15 (39.5%) had conducted a literature search of multiple databases and 34 (89.5%) had conducted a supplementary manual search. The eligibility criteria were clearly presented in 81.6% of studies, but the quality of the primary studies was assessed in only 26.3%. The process and strategy for data extraction was reported in 57.9% of all studies; 19 (50%) studies planned statistical tests of between-study homogeneity and the results were reported in 18, but the level of statistical significance was reported in only 11 (57.9%). The selection of and justification for a statistical model was presented in 39.5% and 26.3% of studies, respectively. Among the 11 meta-analyses in which statistical between-study heterogeneity was reported, 54.5% ignored the statistical findings and proceeded to pool the study results. The implications of between-study heterogeneity were discussed in only 8 studies. CONCLUSIONS: Many methodological flaws were identified in the meta-analyses of H. pylori-related clinical studies, particularly for assessing, reporting and interpreting between-study heterogeneity. This warrants consistent and urgent adherence by reviewers and journal editors to the methodological guidelines for meta-analyses.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 6 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.