The impact of review articles.

The expansion of the scientific literature has produced a concomitant increase in the number of review articles. One may posit that the sheer number of review articles belies their function. This study examines the growth of the review literature, what types of journals publish these papers, and provides data on the citation rate of the review literature. Focus is given to the pathology literature, defined as papers that have the word 'pathology' or its derivatives in the title, abstract or as a key word. The pathology literature is proliferating at a rapid rate; from 1991 to 2006, the total number of original articles increased 2.3-fold, while the number of reviews increased 5.6-fold. Furthermore, in that same time frame, approximately 90% of pathology articles and reviews were not published in pathology journals. An examination of the 538 review articles that were published in pathology journals in 2005 reveals that only 21% of them have been cited more than 10 times since their publication. The impact factors of 12 pathology journals were compared with and without review articles for the period 2000-2006, including The American Journal of Pathology (AJP), The Journal of Pathology (JP) and Laboratory Investigation (LI). Inclusion of reviews increased the impact factor for JP by 0.610+/-0.153 U (+/-s.d.), which was significantly greater than that for AJP (0.109+/-0.086) and LI (0.147+/-0.088). However, for all three journals the total impact factor was largely a reflection of the citations of original articles. The motivations of authors and editors who produce review articles are considered, such as career progress and increasing journal visibility, respectively. The fact that many review articles are poorly cited raises concern about the harm that poor review articles can cause, first by making it more difficult to find the good reviews, and in the worst case by propagating scientific error through lack of critical appraisal of original research. The attributes of the best reviews that serve to shape the future of science are described. These data are presented with the hope that authors and editors will carefully consider their respective roles in ensuring that the body of review literature will be of maximum benefit to the scientific and biomedical community.