The trusted source for
healthcare information and
abstract & commentary
Synopsis: An analysis of abstracts in prominent medical journals found that these journals frequently failed to accurately reflect the data presented in the articles themselves.
Source: Pitkin RM, et al. Accuracy of data in abstracts of published research articles. JAMA 1999;281:1110-1111.
Pitkin and colleagues assessed the information in the abstracts of research articles published in six medical journals to determine whether they accurately reflected the information in the articles themselves. The journals evaluated were: Annals of Internal Medicine, British Medical Journal, Journal of the American Medical Association, Lancet, New England Journal of Medicine, and the Canadian Medical Association Journal (CMAJ). All 44 qualifying articles in the CMAJ over a 15-month period were evaluated, as were random samples of 44 articles from each of the other five published over a 12-month period. Abstracts were considered deficient if they contained data absent or inconsistent with those found in the article.
By these criteria, the proportion of abstracts in each journal that were incorrect ranged from 18-68%. Inconsistencies were more commonly encountered than were omissions, but both types of errors occurred together in 24% of the total. Discrepancies ranged from minor to important.
Lacking the time necessary for careful analysis of a medical article, physicians commonly rely upon the accompanying abstract. Thus, the presence of inaccuracies in those abstracts is of great concern.
The journals selected for examination are all well-known and respected and the editors of those journals, as pointed out by Pitkin et al, were all founding members of the International Committee of Medical Journal Editors, a group that sets editorial standards. Thus, the results reviewed here may seem startling to many, but not to those of us who labor in the vineyards of Infectious Disease Alert (IDA).
While each of the IDA editors works somewhat differently, there is a common theme that I believe is represented by the way in which I approach an article. In each case, I ignore the published abstract and prepare my own summary of pertinent points from the text, tables, and figures of the article itself. This bypasses the problem of the inaccuracies and/or "spin" of the published abstracts and provides the reader of IDA with an independently developed summary and commentary. It also overcomes a potential problem not examined by Pitkin et al—the possible lack of inclusion in the abstract of important information contained in the article. Furthermore, our commentary section provides us with an opportunity to reflect upon the overall results of the article as well as upon the statements made by the authors in their discussion and to place them in context.
As Pitkin et al indicate, the use of structured abstracts is unlikely to overcome this problem. Furthermore, such structure often reduces the readability of the abstract. An accompanying editorial indicates that JAMA implemented a set of abstract quality criteria in January 1998, and that a preliminary analysis indicates that this may have improved the accuracy of its abstracts. However, as Pitkin et al state: "As part of the copy editing process, the abstract needs to be scrutinized painstakingly on a line-by-line or even word-by-word basis and each bit of information verified individually and specifically." While you wait for this utopian vision to arrive, I suggest that you and your colleagues continue to rely on IDA to save you time and invigorate your mind.
1. Winker MA. JAMA 1999;281:1129-1130.