Article type
Year
Abstract
Background: For various reasons, readers of the literature may only refer the abstract of a systematic review to judge the magnitude of effects of alternative patient management strategies. Abstracts of systematic reviews reporting absolute estimates of effect (e.g. risk difference, number needed to treat) on the most patient important outcomes may facilitate decision-making.
Objectives: To determine, in Cochrane and non-Cochrane systematic reviews reporting dichotomous outcomes: (a) the concordance between abstracts and full texts in the reporting of patient important outcomes, and (b) the extent to which absolute estimates of effect are reported in the abstract.
Methods: We searched OVID Medline and the Cochrane Database of Systematic Reviews to identify eligible systematic reviews published in the year 2010. Citations were exported into Distiller (http://systematic-review.net/) and screened using pre-specified eligibility criteria. We stratified studies into two groups: Cochrane and non-Cochrane systematic reviews. Within each group, 100 reviews were randomly selected for data abstraction. Teams of two reviewers are using pre-tested and standardized forms with accompanying instructions, to select and abstract a pair-wise comparison, the most patient important outcomes of benefit and harm respectively, as well as population, intervention and design characteristics from the full-text of each included review. We will select the beneficial and harmful outcomes of interest using a hierarchical approach that will prioritize the outcome of most patient importance (Table 1 below).
Results: Data abstraction is ongoing. Analyses of the results will be presented as a descriptive summary (proportion reporting on each of our objectives) in a tabular format.
Conclusions: This project will determine the extent to which abstracts of systematic reviews report data on the most patient important benefits and harms consistently with the full text, and the extent to which such data are reported in absolute terms.
Objectives: To determine, in Cochrane and non-Cochrane systematic reviews reporting dichotomous outcomes: (a) the concordance between abstracts and full texts in the reporting of patient important outcomes, and (b) the extent to which absolute estimates of effect are reported in the abstract.
Methods: We searched OVID Medline and the Cochrane Database of Systematic Reviews to identify eligible systematic reviews published in the year 2010. Citations were exported into Distiller (http://systematic-review.net/) and screened using pre-specified eligibility criteria. We stratified studies into two groups: Cochrane and non-Cochrane systematic reviews. Within each group, 100 reviews were randomly selected for data abstraction. Teams of two reviewers are using pre-tested and standardized forms with accompanying instructions, to select and abstract a pair-wise comparison, the most patient important outcomes of benefit and harm respectively, as well as population, intervention and design characteristics from the full-text of each included review. We will select the beneficial and harmful outcomes of interest using a hierarchical approach that will prioritize the outcome of most patient importance (Table 1 below).
Results: Data abstraction is ongoing. Analyses of the results will be presented as a descriptive summary (proportion reporting on each of our objectives) in a tabular format.
Conclusions: This project will determine the extent to which abstracts of systematic reviews report data on the most patient important benefits and harms consistently with the full text, and the extent to which such data are reported in absolute terms.
Images
