I. Introduction

Pharmacological Reviews has been publishing high-quality review articles since 1949. Over those 60, years the pharmacological sciences have changed dramatically; accompanying this change in the science has been the development and use of information technologies. These changes have undoubtedly led to transformations in the users, purposes, and accessibility of review articles. Insights into these changes should inform decision making by authors of review articles and the editorial policies of journals that publish them.

II. The Traditional Role of Review Journals

Why did review journals like Pharmacological Reviews develop? A reasonable hypothesis would be that in the era before simple electronic literature searches, the most readily accessible source of information that compiled published literature in a specific topic or field were “definitive reviews,” such as those published by this journal. In 1949, the number of journals that published scientific research was substantially less than it is today. On the other hand, identifying and accessing these articles was considerably more challenging for scientists and for the public than it is today. Index Medicus provided a compendium of published articles. However, it was not universally accessible. Furthermore, listings in Index Medicus were monthly and annual, rather than cumulative, and used a limited number of search terms for scientific topics. As a result, identifying the literature relevant to a research question was challenging and time-consuming. In such an environment, review articles provided a forum for recapitulation of the findings in a field and a comprehensive repository for references in the area of interest. This is reflected in the style, length, and references cited in early Pharmacological Reviews articles. In 1950, articles in the journal typically cited 150 to more than 500 publications, despite the limited number of journals and publications in that era (Table 1). As citable references increased, so did the length of articles and the number of citations (Table 1). By 1975, articles cited up to 700 references.

Those reference lists, which took enormous effort to compile, provided a valuable resource to scientists, particularly to those new to the reviewed field. It is noteworthy that this approach to review article construction has largely remained intact up to the present. Although the number of references in contemporary articles in Pharmacological Reviews varies widely, the number typically ranges from 200 to 500. The number of articles cited in these reviews has remained relatively stable, despite the enormous increase in the universe of articles published and the much easier electronic means with which to access cited articles. This probably reflects a more selective use of references by authors of reviews. Alternatively, this stability in cited references might reflect the development of reviews in more focused topic areas; such reviews would focus on a smaller segment of a broader scientific field surveyed by the previous generation of more all-encompassing reviews.

III. How Has Improved Access to the Scientific Literature Affected the Importance of Review Journals? Do They Remain Relevant?

In contrast to the conditions in 1949, today every scientist and much of the public have ready access to and the capability to perform comprehensive electronic literature searches. These searches can be customized, reviewed, and refined using iterative approaches that allow identification of published works of interest and specific needs or questions to be addressed. Thus, a comprehensive list of works cited in a review article, even in electronic format with direct links to cited articles, is much less important as a resource than it was in 1949. In fact, this aspect of a review article may have less utility today than electronic bibliographic resources such as PubMed or Google Scholar. This is due to the superior abilities of these electronic resources to identify specific articles based on the use of multiple search terms. Thus a major role of the traditional review article (i.e., as a compendium of all known knowledge in a field) is no longer as important in the current scientific environment.

The other major feature of “traditional” reviews has been their encyclopedic nature, a characteristic widely associated with Pharmacological Review articles. How well has that form of communication stood the test of time? In some respects, the answer has to be “remarkably well.” Publications that are dedicated to review articles, including Pharmacological Reviews, remain highly cited in the biomedical literature (Tables 1 and 2, Fig. 1). Beyond traditional measures of “impact,” electronic access provides new metrics for assessing review article utilization. Electronic access also permits a more immediate measure of interest in an article or topic. Our reviews, including newly published ones, are still widely used, even by those newer standards. For example, an article on “The Role of Incretins in Glucose Homeostasis and Diabetes Treatment” published in December 2008 (Kim and Egan, 2008) was accessed 808 times in January 2009 and had the second highest access count among Pharmacological Reviews articles that month. For Pharmacological Reviews, these “e-data” support the hypothesis that reviews are widely used as reference material beyond the impact measurable by their citation numbers (Table 2). It is noteworthy that electronic access of full articles far exceeds the number of citations for the same articles. Such on-line access suggests a much broader readership for our articles than is simply reflected by citation rates.

However, closer examination of impact factors and electronic access rates suggest shortcomings in applying these approaches to assess the “worth” of traditional reviews. Most scientists would agree that the principal purpose of any biomedical scientific communication is to transmit information important either 1) in the generation of new knowledge (i.e., in facilitating other scientists/readers to put their own work in perspective or to stimulate new studies) or 2) as a stimulus to clinical application of that knowledge. In this context, how well do parameters such as impact factors or electronic “hits” hold up as indices for assessing the importance of review articles? We would suggest that both of these indices may overestimate the impact of review articles. With regard to the high internet access rates that Pharmacological Reviews articles enjoy, in some cases the disproportionate number of electronic viewings may represent curious inquiries by nonscientists for certain topics such as “penile erection” and “cocaine.” This concept is supported by disproportionate numbers of electronic accessions (compared with the number of citations) for some articles (Table 2). “Pharmacology of Penile Erection” (Andersson, 2001) has remained the most accessed Pharmacological Reviews article for the past several years (accessed over 1300 times in January 2009 alone). However, more than 90% of electronic viewings for this article were of the HTML full text version rather than the PDF, whereas for most articles, the number is closer to 50%. In contrast, although the article on “The Glutamate Receptor Ion Channels” (Dingledine et al., 1999) had only half the number of electronic accesses, it has been cited almost 10 times more often than Dr. Andersson's review. This analysis suggests that some articles that might be viewed as “high impact” based on electronic access are being used in a more cursory and perhaps less-than-“academic” manner (and as a result, are less likely to affect new knowledge generation or clinical application). Without completely dismissing the utility of publishing review articles for these other uses, such usage has not been traditionally viewed as the principal purpose for their publication.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1.

Impact factors for Pharmacological Reviews (Pharm Rev), Annual Review of Pharmacology and Toxicology (Ann Rev), and Trends in Pharmacological Sciences (TIPS). The impact factor reflects citations from the reporting year of those articles published in the preceding 2 years. For example, the 2008 impact factor for a journal is the average number of citations for articles published in that journal in 2006 and 2007.

What about the use of impact factors as the “gold standard” for assessing the “worth” of a review article in changing scientific thought/direction? Even here, we would suggest that the remarkably high impact factors scored by Pharmacological Reviews and other related journals may also not readily translate to “high impact”—at least as we have defined it—with regard to influencing scientific thought/direction. In a nonsystematic survey, we found that review articles were frequently cited in the Introduction or in the introductory paragraph of the Discussion section of an article. Furthermore, references to reviews were most often used to support general statements about the field of an original article and to provide a resource for interested readers. Thus, the centrality of standard reviews to the works that cite them (or the impact of those reviews in stimulating the design and/or execution of studies described in those articles) can be questioned.

Beyond such quantitative measures of usage and utility, there is a growing sense that encyclopedic review articles and the journals that publish them are becoming less important as resources for learning. Reviews have traditionally provided a means by which scientists could become conversant with a particular area of biomedical science. However, it is questionable whether reading 50 or more pages of dense journal text meets the learning needs and styles of the contemporary scientist. Shorter, more focused reviews may be of higher utility. Consistent with this, many respected venues for original research articles, including The Journal of Biological Chemistry, Molecular Pharmacology, and The Journal of Clinical Investigation, as well the journals of the American Heart Association (among many others), now publish extended perspectives and “minireviews” in areas perceived to be of high current interest. These reviews typically allow readers to familiarize themselves with a much more circumscribed research area than those covered in more traditional reviews. Furthermore, in the best examples of this type, such reviews provide readers with a succinct understanding of the field's significance and potential future directions. The impact of such perspectives and minireviews has generally been positive (at least as judged by impact factor trends for many of the specialty journals). However, the ultimate “value,” as judged by how effectively the articles “move” scientific thinking, remains to be established.

Thus, despite some measures demonstrating the persistent utility and impact of review articles, such as those that published in Pharmacological Reviews, one might reasonably question whether encyclopedic reviews and review journals, like ours, remain relevant to 21st-century scientists.

IV. Is There a Future for Review Journals? If So, What Should They Look Like?

We believe that review formats will continue to be important forms of scientific communication, despite their shortcomings and despite the dramatic changes in how and why scientists access information. However, our sense is that it is important to have a clear perspective on what scientists require when they read reviews and for editors to use such information to optimally shape editorial policy. We suggest that there are several important functions that review articles could subserve and that may have significant impact on reshaping the format of review journals like ours.

V. Closing Thoughts

During our term as Editors for Pharmacological Reviews, we have tried to understand the new dynamics of the biomedical review article and to allow the journal's evolution to meet these trends. With the assistance of the Associate Editors, we have worked to identify topics that we thought were of high current interest, and we encouraged the development of shorter, more focused reviews prepared with more aggressive time lines. It is noteworthy that we have asked authors of reviews to be more integrative and synthetic in their approach and to be more forward looking, as opposed to being exclusively retrospective. At the same time, we have acknowledged the value of the more encyclopedic review as a repository for the work in a field and have encouraged such reviews in some areas. Whether these efforts have been either meritorious or successful we leave to the pharmacology community to decide.