Evaluating the Quality of Meta-Analyses in the Ophthalmology Literature

Abstract

Purpose:: The purpose of the study was to assess the methodological quality of meta-analyses published in five leading Ophthalmology journals.

Methods:: We searched the MEDLINE database electronically to identify meta-analyses published in five leading ophthalmology journals between 2002 and 2006. The main inclusion criterion was the reporting of pooled quantitative results of studies in ophthalmology. All meta-analyses were retrieved and independently assessed for quality by three reviewers using a 10-item scale from the Overview Quality Assessment Questionnaire.

Results:: Nineteen meta-analyses were identified to be included in this evaluation. Overall agreement between the three independent reviewers was good (67% agreement). The scientific quality of the meta-analyses, as assessed by the 10-item scale, was generally low (mean score= 4.5 ± 2.4, maximum possible score=7). Overall, more than half of the meta-analyses (53.3%) had major flaws (score ≤4), including poor descriptions of statistical analyses and the omission of information on the quality of the studies that were pooled. The agrrement between the three reviewers about the presence of major flaws in the studies (score ≤4) was 100%.

Conclusions:: This evaluation of the quality of meta-analyses published in ophthalmology journals between 2002 and 2006 suggested major methodological flaws. Meta-analyses with low quality may limit the validity of the reported results. More efforts should be made to improve the quality of meta-analyses in ophthalmology journals by following published guidelines for the conduct and reporting of meta-analyses.

Keywords: clinical (human) or epidemiologic studies: biostatistics/epidemiology methodology