A methodological survey of the analysis, reporting and interpretation of Absolute Risk ReductiOn in systematic revieWs (ARROW): a study protocol

作者：P. Alonso-Coello, A. Carrasco-Labra, R. Brignardello-Petersen, I. Neumann, E. Akl, Xin Sun, B. Johnston, M. Briel, J. Busse, D. Glujovsky, C. Granados, A. Iorio, Affan Irfan, L. García, R. Mustafa, Anggie Ramírez-Morera, I. Solà, K. Tikkinen, S. Ebrahim, P. Vandvik, Yu-Qing Zhang, A. Selva, A. Sanabria, Oscar E. Zazueta, R. Vernooij, H. Schunemann, G. Guyatt · 发表于：Systematic Reviews · 年份：2013 · DOI：10.1186/2046-4053-2-113 · 被引用次数：26 · 研究领域：Medicine

BackgroundClinicians, providers and guideline panels use absolute effects to weigh the advantages and downsides of treatment alternatives. Relative measures have the potential to mislead readers. However, little is known about the reporting of absolute measures in systematic reviews. The objectives of our study are to determine the proportion of systematic reviews that report absolute measures of effect for the most important outcomes, and ascertain how they are analyzed, reported and interpreted.Methods/designWe will conduct a methodological survey of systematic reviews published in 2010. We will conduct a 1:1 stratified random sampling of Cochrane vs. non-Cochrane systematic reviews. We will calculate the proportion of systematic reviews reporting at least one absolute estimate of effect for the most patient-important outcome for the comparison of interest. We will conduct multivariable logistic regression analyses with the reporting of an absolute estimate of effect as the dependent variable and pre-specified study characteristics as the independent variables. For systematic reviews reporting an absolute estimate of effect, we will document the methods used for the analysis, reporting and interpretation of the absolute estimate.DiscussionOur methodological survey will inform current practices regarding reporting of absolute estimates in systematic reviews. Our findings may influence recommendations on reporting, conduct and interpretation of absolute estimates. Our results...