Real-time updates of meta-analyses of HIV treatments supported by a biomedical ontology.

Meta-analyses involve a systematic statistical explanation of available evidence and are a commonly used systematic reviewing strategy for addressing health related scientific research. Publication biases often delay reporting of meta-analyses and subsequent policy making. Ontologies have the capacity to provide real time meta-analyses that can reduce publication delay. The purpose of this study was to describe the meta-analytic process of a recently revised ontology. Three recently published meta-analyses were replicated using an ontology that extracts appropriate statistical "tags" from the originally reviewed published manuscripts. The ontology was able to closely represent the findings of the original meta-analysis including funnel, forest, and summary statistics. The use of an ontology could improve the speed in which meta-analytic publications occur as well as potentially standardizing the required elements within a randomized controlled trial. The findings in this study demonstrate similar ontological output as used in the Trial Bank Project with decreased time requirements of the manuscript author and the software programmers. The use of an ontology may improve the reporting of meta-analysis and, when combined with an open access website, should assist in timely policy-making decisions by healthcare providers.

Duke Scholars

Author Chad E. Cook Orthopaedic Surgery, Physical Therapy