Article type
Year
Abstract
Objectives: To apply the current guidelines for conducting meta-analyses of observational trials to the Environmental and Occupational Health literature.
Methods: We identified relevant articles in MEDLINE using the keywords: meta-analysis, environmental exposure, occupational exposure, occupational disease, and textword: meta-analysis. All abstracts were reviewed and full articles were retrieved if they presented either a meta-analysis of environmental health or occupational health studies or were methodological papers that discussed guidelines for combining observational studies. The search was supplemented by additional citations from the reference lists of included articles and meta-analyses identified by investigators in the field. Abstraction forms were developed using the draft Cochrane Collaboration guidelines for synthesizing non-experimental studies and other guideline papers. Each meta-analysis article was abstracted in duplicate and assessed in four main areas: study identification and inclusion, data extraction, data analysis and presentation, and communication of results.
Results: There was sparse reporting of searching strategies: none reported having a research protocol, half reported the database(s) searched, and less than 60% reported a searching strategy. Similarly, although most reports included information about the exposure and outcome being measured, less than half specified their abstraction methods and less than a third collected information to assess bias. A wide variety of analytic techniques were used, about half of the studies attempted to assess heterogeneity and perform sensitivity analyses while 75% reported subgroup analyses. In communicating their results, almost 90% of the authors addressed alternative explanations for their results. Funding sources were disclosed in less than half of the articles.
Discussion: There is a discrepancy between what are thought to be important aspects of methodologically rigorous observational study meta-analyses by experts in the field, and what is actually published in the peer-reviewed Environmental and Occupational Health literature. There is some indication that the more recently published meta-analyses more closely follow the guidelines proposed by the experts but a gap still exists.
Methods: We identified relevant articles in MEDLINE using the keywords: meta-analysis, environmental exposure, occupational exposure, occupational disease, and textword: meta-analysis. All abstracts were reviewed and full articles were retrieved if they presented either a meta-analysis of environmental health or occupational health studies or were methodological papers that discussed guidelines for combining observational studies. The search was supplemented by additional citations from the reference lists of included articles and meta-analyses identified by investigators in the field. Abstraction forms were developed using the draft Cochrane Collaboration guidelines for synthesizing non-experimental studies and other guideline papers. Each meta-analysis article was abstracted in duplicate and assessed in four main areas: study identification and inclusion, data extraction, data analysis and presentation, and communication of results.
Results: There was sparse reporting of searching strategies: none reported having a research protocol, half reported the database(s) searched, and less than 60% reported a searching strategy. Similarly, although most reports included information about the exposure and outcome being measured, less than half specified their abstraction methods and less than a third collected information to assess bias. A wide variety of analytic techniques were used, about half of the studies attempted to assess heterogeneity and perform sensitivity analyses while 75% reported subgroup analyses. In communicating their results, almost 90% of the authors addressed alternative explanations for their results. Funding sources were disclosed in less than half of the articles.
Discussion: There is a discrepancy between what are thought to be important aspects of methodologically rigorous observational study meta-analyses by experts in the field, and what is actually published in the peer-reviewed Environmental and Occupational Health literature. There is some indication that the more recently published meta-analyses more closely follow the guidelines proposed by the experts but a gap still exists.