Skip the Abstract
July 21, 2006 on 10:32 am | In Critical Appraisal Tips |
Print
|
Clinical Papers are written up in a standard scientific style with an introduction of the paper topic, details of the method used in the study, the study results and a discussion of these results. Often clinical papers will also have a summary at the beginning called an abstract.
The British Medical Journal has added a study to Online First ahead of publication in the paper journal that examined the content of the abstract compared to the main paper.
This study found that the p-values quoted in abstracts are not evenly distributed but are skewed towards values representing statistical significance. The study went on to check the accuracy of the p-values where the data was available. Of the 27 studies that were checked 5 studies were found to be quoting incorrect p-values with an additional 9 studies that used the wrong statistical test or altered the data before performing the test. A table summarises these results.
The conclusion of the paper is “Significant results in abstracts are common but should generally be disbelieved”. So, when reading a clinical paper, skip the abstract!
Bookmark in del.icio.us