Fifty-Year Fate and Impact of General Medical Journals

Background Influential medical journals shape medical science and practice and their prestige is usually appraised by citation impact metrics, such as the journal impact factor. However, how permanent are medical journals and how stable is their impact over time? Methods and Results We evaluated what happened to general medical journals that were publishing papers half a century ago, in 1959. Data were retrieved from ISI Web of Science for citations and PubMed (Journals function) for journal history. Of 27 eligible journals publishing in 1959, 4 have stopped circulation (including two of the most prestigious journals in 1959) and another 7 changed name between 1959 and 2009. Only 6 of these 27 journals have been published continuously with their initial name since they started circulation. The citation impact of papers published in 1959 gives a very different picture from the current journal impact factor; the correlation between the two is non-significant and very close to zero. Only 13 of the 5,223 papers published in 1959 received at least 5 citations in 2009. Conclusions Journals are more permanent entities than single papers, but they are also subject to major change and their relative prominence can change markedly over time.


Introduction
Medical journals shape clinical practice, health policy, public health, and biomedical research. Many people think of journals as stable values and don't contemplate that many influential journals may become non-influential or cease to exist in the future. One needs to examine the history of journals to understand these possibilities of change and decay. General medical journals emerged in the Age of Enlightenment. For example Lancet started in 1823 and NEJM in 1812, changing names several times (New England Journal of Medicine and Surgery, New England Medical Review and Journal, Boston Medical and Surgical Journal) before taking its current name in 1928. In the last decades, the number of journals has grown rapidly, including an increasing list of specialty venues [1]. Just as new journals appear, existing journals may change, or may stop their circulation. Moreover, the relative impact of specific journals compared to others changes over time. Journals' impact is traditionally measured by citation metrics, such as the Journal Impact Factor (JIF) [2] that takes into account the number of citations received by very recent papers. How does this compare with the impact of papers published by these same journals long ago?
Here we examine the long-term fate and impact of general medical journals. We have taken a snap shot of journals publishing papers 50 years ago, in 1959. We have examined the fate of these journals in the subsequent half-century and assessed the impact of the papers they published in 1959 as compared with the current JIF.

Methods
We identified all journals that had published articles indexed in the General and Internal Medicine subject category by Thomson ISI Web of Science [3] for 1959. With appropriate subscription, it is possible to refine the searches in ISI so as to include specific subject categories. Then we searched the Journal function in PubMed to see if these venues continued or stopped publication, and if so, when. We also recorded whether journals merged with other journals, and if they changed names. We also traced previous names (before 1959) of each journal and when the journal (or its predecessors) had first been published.
We retrieved all items published in the eligible journals in 1959 and retained only those categorized by ISI as articles or reviews. The definition of what constitutes an article, a review or other item has been a contentious issue and some journals may try to increase their impact factor by publishing items that are not categorized as papers counted in the denominator of impact factor calculations. We have tried to avoid adding another layer of subjective tagging by re-characterizing the category of each of the published items ourselves, and thus we adopted the tagging provided already by ISI. However, this caveat should not be dismissed as it may affect the impact factor calculations of some journals.
We recorded from ISI the total citations received until the end of 2009 only for the articles and reviews of each journal (excluding citations to other items). We estimated the average 50-year citation impact per paper (article or review) for each journal as the ratio of the total citations (during 1959-2009) to articles and reviews published in 1959 divided by the number of such papers. Similarly, we estimated the average 2-year citation impact per paper as the ratio of the total citations during 1960-1961 to articles and reviews published in 1959 divided by the number of such papers. We assessed the Pearson correlation coefficient between the 50-and 2-year citation impact per paper and the latest JIF (Journal Citation Reports, 2008 edition). JIF uses in its calculation two years of recent publications and one year of citations (e.g. citations in 2009 for papers published in 2008 and 2007) and counts in the nominator also citations to items other than articles and reviews [2]. Therefore, we also estimated the 2year citation impact per paper for papers published in 2007 for symmetry of definition to the respective 1959 metric and sensitivity analyses were also performed with the traditional definition of JIF with similar results (not shown in detail).
Furthermore, we have examined how many of the 20 general or internal medicine journals with the highest impact factors currently (per Journal Citation Reports 2008) were not published 50 years ago (neither with the same nor different name). Given that journals may be seen as businesses, for comparison, we also examined how many of the current top businesses worldwide (based on the Fortune 500 global edition, 2010), were not yet incorporated 50 years ago with information obtained from wikipedia.
Finally, we also assessed which articles published in 1959 received at least 5 citations in 2009.

Fate of journals
We identified 27 eligible journals in 1959 (Table 1). Four journals were no longer published by 2009. Two of them, the Transactions of the Associations of American Physicians and the Bulletin of the Johns Hopkins Hospital were among the oldest general journals and remained highly influential in medical research for a century (they had been launched in 1886 and 1889, respectively) before stopping circulation. Another 7 journals had changed names (two of those had also merged with other journals), but they still continued publication currently. One of them, the Annales de l'Institut Pasteur was considered as a journal covering the General and Internal Medicine category (in addition to Microbiology) in 1959, but lost this character in its subsequent transformation. Sixteen journals continued publication currently with the same name as in 1959.
Eleven of the 27 journals had already succeeded a predecessor with a different name before 1959 (Table 1) Of the 27 journals, only 6 were published continuously with the same name since their first circulation. Of these 6, only Lancet preceded the 20 th century.

Citation impact of journals
In the 50-year frame, Lancet received the highest number of citations for the articles it published in 1959, followed at a distance by NEJM, BMJ, JAMA, and American Journal of Medicine that received approximately the same number of citations among them ( Table 2).
Medicine (Baltimore) had the highest 50-year citation impact per paper. Each of the 15 papers that it published in 1959 received an average of 118 citations in 1959-2009. It was followed by the QJM (91 per paper), the Bulletin of the Johns Hopkins Hospital (59 per paper), the American Journal of Medicine (59 per paper), and Lancet (35 per paper), NEJM (30 per paper) and Annals of Internal Medicine (28 per paper). The 2-year impact per paper (1960)(1961) was highly correlated with the 50-year impact per paper (r = 0.87, p,0.001).
The impact of the papers published in 1959 gives a very different picture compared with the current JIF of these journals.  (Table 2).

Discussion
Our evaluation shows that most of the influential journals of 50years ago have survived to-date, but many have changed names and at least 4 have stopped circulation. The 50-year citation impact gives a very different picture about the relative influence of the general medical journal compared to the current JIF. If journals are seen as businesses, and one accepts that businesses come and go, then the changes in names and the lack of stability for journals is not much different to what in seen for top business corporations. Finally, less than 1 in 400 papers get 5 or more citations per year after 50 years have lapsed.
Few medical journals have kept their original name throughout their history. Sometimes, a change in name may be just a trivial modification, but in other occasions it may signify a change in course, focus, or perception about the mission of a journal, its audience and its content. Usually names of brands (not only journals) do not change unless there is a major reason, since the name of a brand is tied to the recognition and prestige of its products. Keeping abreast of developments in biomedical sciences is a perpetual challenge. Old disciplines disappear and new ones emerge. As an example of a disappearing discipline, until the advent of penicillin, the study and management of syphilis occupied a large specialty with many practitioners and scientists.  Note that the nominator includes citations to all items published by a journal regardless of whether this is an original article, review or other type of item (e.g. editorial, essay, etc), while the denominator includes only original articles (articles and proceedings papers, a category that did not exist in 1959) and reviews. Moreover, the journal impact factor considers papers published during two years and the citations that they received in a single year (one or two years after their publication year, respectively). Therefore, the definition is slightly different from the 2-year citation impact metric that we used for papers While general journals are more stable than specialty journals by virtue of wider circulation or wider outreach and by their ability to accommodate material from whatever are the thriving or emerging specialties and disciplines du jour, they still have to struggle to survive themselves in a changing practice and research environment. Change and even decay are common. In fact, the first English general medical journal, the Medicina Curiosa that started in 1684, ceased its publication after only two issues [17]. The second one, the Medical and Philosophical Commentaries, which was launched in 1773 was very influential for almost two centuries [18], but changed names several times. It stopped circulation as Edinburgh Medical Journal in 1954.
JIF was first calculated [19] by Eugene Garfield in 1972 and has had an increasingly pervasive influence on appraising journals since then [2]. We should acknowledge that physicians and clinician-investigators did not depend on JIF or any other citation metrics to appraise the prestige of journals back in 1959 and it is difficult to say what they thought exactly about the relative ranking of journals back then. However, even as a retrospective exercise, the calculation of the impact metrics for papers published in 1959 shows the influence these papers had in the subsequent literature. Our analysis shows that the difference between the 50-year impact and current impact is not due to the short-term nature of calculations in estimating JIF. The relative impact of journals for their 1959-published papers was similar regardless of whether we examined 2-or 50-year citations. Apparently, sleeping beauties (articles that don't get cited initially, but receive many citations after several years) are rare [20][21][22].
Our data suggest that most influential articles were probably sent to different journals in 1959 than they would have gone to in 2009. Medicine, QJM and American Journal of Medicine were considered more prestigious than NEJM in 1959. Modest changes in the JIF ranking of journals have been seen in an analysis of 7 general journals covering a 12-year period (1994-2005) [23], but changes are more striking over half a century. It is difficult to see how and why perceptions about specific journals' ranking changed over time. Most likely this has been a very complex process and each journal has its own story to tell. General journals in these 50 years have had to compete for the coverage of an increasing number of specialties and sub-specialties that gradually became independent with their own stand-alone journals [24]. The longtail distribution principle (few papers get a lot of citations, most papers get few citations) probably has operated during the whole length of this half century [25]. However, at a time when JIF was not yet proposed or at least not influential, leading papers were probably sent to diverse journals and a change in the relative citation impact of different journals could have happened both by wise editorial choices but also even by chance, given that it has been difficult to know with perfect certainty which papers will eventually be most influential [26], perhaps with the exception of some large landmark studies that are often collaborative and for which a large number of scientists awaits there results.
In the current environment where JIF has reached its apotheosis [27], it may be more difficult for a journal with substantially lower JIF to outperform one with much higher JIF in citation counts. Even if the lower-JIF journal publishes better papers, the community may feel obliged to cite papers published in journals with higher JIF [28], regardless of their merit. This preference leads to a spurious centralization of science to a few journals [29,30], even though the extent of the centralization has been debated or even refuted [31]. It is unclear whether this situation will also continue in the future, as there is increasing interest to adopt additional, different metrics of impact [27,32].
As for single papers, the vast majority of them have a citation life of anywhere between a few years to a couple of decades [2]. Few papers survive 50 years in the citation game. The citation decay is even more rapid in other scientific fields, such as molecular genetics, while some fields with slower turn-over (e.g. mathematics) may have more papers that continue to be cited for many years. Papers that are no longer cited have not necessarily been refuted and proven to be wrong in their inferences. Possibly most of them stop being cited simply because their field makes progress and investigators are inclined to cite papers that are more recent.

Conclusion
Overall, our evaluation shows that single papers have a very transient presence in citations and with few exceptions they are rarely cited half a century after their publication. Journals are somewhat more permanent than single papers, but even most influential journals cannot avoid change and decay. The club of influential journals changes membership and ranking over time and in the long-term changes may become impressive. The very name of journal, derived from the old French jurnal and from the Latin diurnus denotes something which has a daily character, something that pertains to a single day, lasts for a day or is important for a day -only. The Greek equivalent word for journal is EWGMERIS and not surprisingly it offers the root for what is ephemeral. As Heraclitus said, TA PANTA REI, everything changes.