Highly Cited Articles in Environmental and Occupational Health, 1919–1960

ABSTRACT

Although numerous lists of “citation classics” have been compiled across a variety of scientific fields, few have included articles from environmental and occupational health (EOH). This investigation sought to identify and analyze the most highly cited articles ever published in the Journal of Industrial Hygiene (1919–1935), the Journal of Industrial Hygiene and Toxicology (1936–1949), the Archives of Industrial Hygiene and Occupational Medicine (1950), the American Medical Association (A.M.A.) Archives of Industrial Hygiene and Occupational Medicine (1950–1954), and the A.M.A. Archives of Industrial Health (1955–1960). Regularly cited topics included metal fume fever and various studies of beryllium, whereas the most highly cited article of all was a 1957 paper describing the control of heat casualties at military training centers. Interestingly, the most highly cited articles were not the oldest, and nor were they written as literature reviews. Overall, this study suggests that although some citation patterns in EOH reflect those of other disciplines, the trend is not uniform and EOH itself appears to have some distinctive bibliometric characteristics.

KEYWORDS:

• citation classics
• environmental health
• highly cited articles
• occupational health
• publishing

Rapid advances in computing power, increased communication via the Internet, and a contemporary academic obsession with bibliometric indicators have all ensured that the field of citation indexing has expanded dramatically in recent years. Counting the number of times that an individual article, author, or journal has been cited by others, and who has been cited more than who, now represents one of the largest subdisciplines of bibliometrics. ¹ Increasing academic interest, and perhaps general curiosity, has also been focused on papers in a particular journal or field that have attracted an unusually high proportion of all citations, the so called highly cited articles. Another term commonly used to describe the same thing, citation classics, was coined by Eugene Garfield in the late 1970s. ² Garfield had earlier proposed the concept of an impact factor in 1955, ³ and founded the Institute for Scientific Information (ISI) in 1958. ⁴ One of its core products was a database of scientific articles known as the Science Citation Index® (SCI®), which was first published in 1961. ⁵ Having such a large database of published articles and their reference lists made it possible to recognise various citation-based phenomena that were occurring in the journal literature.

By the late 1960s, computers had made it easier to search through large databases such as those created by Garfield, and therefore, to establish which scientific papers were being highly cited. In one of the first studies published in 1971, Garfield compiled a list of the 50 most-cited articles from 1967. ⁶ This was followed by a list of the 50 most-cited articles from 1961 to 1972, ⁷ a list of classic papers from the late 19th and early 20th centuries, ⁸ and a paper describing the most highly cited articles, ever. ⁹ Being founder and director of the ISI allowed unlimited access to the company databases, and as such, much of the early work on highly cited papers during this period continued to be written by Garfield himself. ¹⁰ Lists of citation classics focused on individual journals such as the Lancet, ¹¹ the New England Journal of Medicine, ¹² Annals of Internal Medicine, ¹³ and the Journal of the American Medical Association, ¹⁴ as well as in broad scientific fields ranging from “anthropology to urban studies.” ¹⁵ Academic curiosity was evidently stimulated by Garfield's articles, and by the 1980s, other authors had started publishing lists of highly cited articles in various fields. ¹⁶ There has also been a certain amount of overlap between disciplines. Casey, ¹⁷ for example, investigated the influence of American and British journals on medical progress. Interestingly, his analysis did include an environmental and occupational health (EOH) journal, the Journal of Industrial Hygiene. A comprehensive historical review of early citation analysis in the broader sphere of “health science” was published some time later. ¹⁸

By the 1980s, investigations by scholars other than Garfield began to emerge. One of the earliest appears to have focused on the journal Fertility and Sterility, which included a list of citation classics in 1987 ¹⁹ and later in 2006. ²⁰ In 1988, a list of citation classics was published for the Archives of Physical Medicine and Rehabilitation, ²¹ followed by the American Journal of Hospital Pharmacy in 1989 ²² and the Journal of Investigative Dermatology in the same year. ²³ In 1990, citation classics in the Journal of Laboratory and Clinical Medicine were also identified. ²⁴ Possibly due to the fact that much of the early work was published by Garfield himself, most of the early compilations of citation classics utilized the ISI database. Even so, then as now, the identification and publication of citation classics within a particular journal or speciality relies heavily on the motivations and interests of authors who write for it. Certain fields have attracted more attention from the “citation counters” than others, however, with dermatology for example, having many such lists. ^{25 , 26 , 27 , 28 , 29 , 30} General medicine has not been neglected in recent years either, with citation classics and highly cited authors in general medical journals, ³¹ surgical journals, ³² as well as speciality fields such as peritoneal dialysis. ³³ Journals in newly emerging speciality areas have also published lists of classics in recent years, including the Journal of Analytical Toxicology, ³⁴ the Journal of Pediatric Orthopaedics, ³⁵ and Plastic and Reconstructive Surgery, ³⁶ to name a few.

Despite this seeming proliferation of articles describing their highly cited cousins, relatively few authors have specifically looked at top-cited papers in the journals of environmental and occupational health (EOH). This is despite over 100 years of EOH journals being published across a variety of countries. ³⁷ As readers may know, the Archives of Environmental & Occupational Health (AEOH) has a long and distinguished history that has so far involved a total of 7 different names since 1919. ³⁸ Briefly, the AEOH's earliest ancestors can be traced to foundation of the Journal of Industrial Hygiene (JIH) by the US Industrial Medical Association (IMA) in 1919. The JIH would publish issues until 1935, after which time it was continued as a succession of periodicals known as the Journal of Industrial Hygiene and Toxicology (JIHT) (published between 1936 and 1949), the Archives of Industrial Hygiene and Occupational Medicine (AIHOM) (published in 1950), the American Medical Association (A.M.A.) Archives of Industrial Hygiene and Occupational Medicine (AMA-AIHOM) (published between 1950 and 1954), the A.M.A. Archives of Industrial Health (AMA-AIH) (published between 1955 and 1960), and perhaps its most famous incarnation, the Archives of Environmental Health (published between 1960 and 2004). Since 2005, the journal has been known as the Archives of Environmental & Occupational Health. Due to the breadth of material published by this journal and its predecessors over the past 90 years, the current article focuses on highly cited papers published in its first 5 incarnations between 1919 and 1960. A detailed citation analysis of the sixth and longest running version of the journal, the Archives of Environmental Health (AEH), is described elsewhere. ³⁹

Table 1 Highly Cited Articles From the Journal of Industrial Hygiene (1919–1935)*

Rank	Author(s)	Title of Article	Year	Volume	Pages
1	Drinker P, Thomson RM, Finn JL. ^51	Metal fume fever: II. Resistance acquired by inhalation of zinc oxide on two successive days.	1927	9	98–105
2	Edsall DL, Wilbur FP, Drinker CK. ^52	The occurrence, course and prevention of chronic manganese poisoning.	1919	1	183–193
3	Haggard HW. ^53	The toxicology of hydrogen sulphide.	1925	7	113–121
4	Bulmer FM, Mackenzie EA. ^54	Studies in the control and treatment of “Nickel Rash.”	1926	8	517–527
5	Drinker P, Thomson RM, Finn JL. ^55	Metal fume fever: IV. Threshold doses of zinc oxide, preventive measures, and the chronic effects of repeated exposures.	1927	9	331–345
*As identified by Scopus.

METHODS

Establishing citation counts for articles up to 90 years old incurs numerous technical challenges, not the least of which is the fact that it involves a time before digital archives, personal computers, and citation indexing existed. As such, it is often difficult to find electronic records, and sometimes even printed versions, of articles published in the earliest journals. Searching for citations themselves, particularly older ones, often leads to the discovery of citation errors and misquoting, which can seriously thwart the ability of search engines to properly count the true number of citations received. Citations themselves are not always accurate, ⁴⁰ and incomplete citations are known to represent a major source of bibliographic chaos. ⁴¹ When preparing for the current analysis, numerous errors were found relating to journal names. Some of the older incarnations of the journal do not appear to be listed in the usual bibliometric databases, with many citations having to be extracted from the reference lists of published articles. When searching in this manner, it was also ascertained that many reference lists compiled by authors have evidently used incorrect or incomplete journal names. It is worth noting that uniform requirements for manuscripts submitted to biomedical journals were not introduced until the late 1970s. ⁴²

Another source of confusion for authors may relate to the fact that although the journal changed its name many times, the names themselves remained somewhat similar. For example, in 1936 the only change was the addition of the words “and Toxicology,” whereas in 1950 the “AMA” prefix was added. The term AMA itself has also appeared in reference lists in many different variations, such as “AMA” (without periods), “A.M.A.” (with periods), or even the full words: “American Medical Association.” Similarly, misspelled author names also added another layer of complexity when searching for citations. At least one highly cited article had been cited in different publications with different first authors. Although citations errors such as this can be reduced to some extent, ⁴³ there is still work to be done to prevent them altogether. ⁴⁴

Table 2 Highly Cited Articles from the Journal of Industrial Hygiene and Toxicology (1936–1949)*

Rank	Author(s)	Title of Article	Year	Volume	Pages
1	Hardy HL, Tabershaw IR. ^59	Delayed chemical pneumonitis occurring in workers exposed to beryllium compounds.	1946	28	197–211
2	Hueper WC, Wiley FH, Wolfe HD. ^60	Experimental production of bladder tumors in dogs by administration of beta-naphthylamine.	1938	20	46–84
3	Deichmann WB, LeBlanc TJ. ^61	Determination of the approximate lethal dose with about six animals.	1943	25	415–417
4	Levine L, Fahy JP. ^62	Evaluation of urinary lead concentrations. I. The significance of the specific gravity.	1945	27	217–223
5	Eisenbud M, Wanta RC, Dustan C, et al. ^63	Non-occupational berylliosis.	1949	31	282–294
*As identified by Scopus.

Table 3 Highly Cited Articles from the Archives of Industrial Hygiene and Occupational Medicine (1950)*

Rank	Author(s)	Title of Article	Year	Volume	Pages
1	Landahl HD, Herrmann RG. ^69	Retention of vapors and gases in the human nose and lung.	1950	1	36–45
2	Srbova J, Teisinger J, Skramovsky S. ^70	Absorption and elimination of inhaled benzene in man.	1950	2	1–8
3	Cochran KW, Doull J, Mazur M, et al. ^71	Acute toxicity of zirconium, columbium, strontium, lanthanum, cesium, tantalum and yttrium.	1950	1	637–650
4	Hall RH, Scott JK, Laskin S, et al. ^72	Acute toxicity of inhaled beryllium. III. Observations correlating toxicity with the physicochemical properties of beryllium oxide dust.	1950	2	25–48
5	Lester D, Greenberg LA. ^73	The toxicity of sulfur hexafluoride.	1950	2	348–349
*As identified by Scopus.

With these limitations in mind, the next issue is to decide which search engine one should actually use. Possibly due to the fact that all of the early citation classics identified by Garfield had been extracted from the ISI databases, much of the later work published by other authors also appears to have utilized this system (nowadays known as the Thomson Reuters Web of Science® database). ⁴⁵ Competitors have emerged in recent years, however, with Google Scholar and Scopus being 2 examples, each with its own strengths and weaknesses. ^{46 , 47} How these intrinsic differences in search engines actually affect the end result when searching for citations depends heavily on what one will define as a “valid” citation. That is, was the article in question cited by an international academic journal or a well-known book, or, was it cited in a conference abstract, a thesis, an unpublished report, or elsewhere? The different search engines mentioned above all detect citations in different ways, and according to Bakkalbasi and colleagues, ⁴⁶ none of them in isolation is the answer to all citation tracking needs, although all three can certainly retrieve some interesting material.

Nevertheless, there are various reasons for using the Scopus database exclusively in the current study. Firstly, it has been shown that Scopus can be used as the sole data source for citation-based research in certain fields, ⁴⁸ including EOH. ⁴⁹ Secondly, aside from having a relatively large database of source journals, Scopus also tracks articles in the reference lists of others. This makes it very useful for identifying very old references from old journals that are not included on the ISI journal list. Thirdly, although other Internet-wide search engines such as Google Scholar may be useful for identifying “grey literature” (ie, older, nonlisted journals), this method is known to be very time-consuming when compared to some of the more scientifically orientated databases. ⁵⁰ As a result, the current study was undertaken using the Scopus database. To keep the results to a manageable level, however, only the top 5 most highly cited articles from each periodical title were sought.

The main searches were undertaken on Scopus during 2009 using the “Source Title” selection in a “Basic Search.” Multiple searches were then performed to account for each potential name variation for each incarnation of the journal. Searches targeted, for example, the “Journal of Industrial Hygiene,” “J Industrial Hygiene,” “J Ind Hygiene,” “J Industrial Hyg,” “J Ind Hyg,” and so on. Later version of the journal which included the “AMA” prefix were searched in similar ways, using variations such as “AMA” (without periods), “A.M.A.” (with periods), or even the full words: “American Medical Association.” All searches were manually examined and refined by limiting output to the specific journals of interest. This was achieved by using the “Refine Results” feature for “Source Title” (to establish if it was a journal of interest) and “Year” (was the journal of interest actually published in this particular year). Following extensive refinement in this manner, a list of the top 5 most highly cited articles for each journal was obtained.

RESULTS

Journal of Industrial Hygiene (1919–1935)

Highly cited articles from the JIH are displayed in Table 1. ^{51 , 52 , 53 , 54 , 55} Two of the most highly cited articles in this list described metal fume fever. The first, by Drinker, Thomson, and Finn ⁵¹ focused on the resistance acquired by inhalation of zinc oxide on 2 successive days. In their article, the authors describe how pronounced leukocytosis accompanied metal fume fever among 2 healthy, male subjects, although by the day after their second exposure, both were free from symptoms and felt themselves to be in normal health. ⁵¹ These articles were evidently part of a series on metal fume fever, with Part I being published by Sturgis and Thompson in the JIH during 1927, ⁵⁶ and Parts III and V being published by Drinker and colleagues in the same journal during 1927 ⁵⁷ and 1928, ⁵⁸ respectively. The second article on this topic, also by Drinker, Thomson, and Finn, ⁵⁵ describes threshold doses of zinc oxide, preventive measures, and the chronic effects of repeated exposures, with regard to metal fume fever.

Journal of Industrial Hygiene and Toxicology (1936–1949)

Highly cited articles from the second incarnation, the JIHT, are displayed in Table 2. ^{59 , 60 , 61 , 62 , 63} Similar to the JIH, 2 of the top 5 most highly cited articles from the JIHT focused on similar topics. In the first, Hardy and Tabershaw ⁵⁹ described delayed chemical pneumonitis occurring among 17 workers exposed to beryllium compounds in a fluorescent lamp manufacturing company. The disease was unusual due to its delayed onset, marked weight loss, intense dyspnea, and poor prognosis. ⁵⁹ Hardy and colleagues would later go on to publish numerous other articles on beryllium in the JIHT's successor journals the AMA-AIHOM, the AMA-AIH, and the AEH. ^{64 , 65 , 66 , 67 , 68} Another highly cited paper in the JIHT also focused on beryllium, this time with Eisenbud, Wanta, and Dustan, ⁶³ describing nonoccupational berylliosis.

Archives of Industrial Hygiene and Occupational Medicine (1950)

Highly cited articles from the third incarnation, the AIHOM, are displayed in Table 3. ^{69 , 70 , 71 , 72 , 73} The most highly cited article from the AIHOM was a paper by Landahl and Herrmann, ⁶⁹ which described the retention of vapors and gases in the human nose and lung. Their experiments were undertaken as part of a program to help evaluate factors involved in human gas exchange, as well as extending the investigation of particulate retention. Results obtained were intended to help assess the hazards of inhaled gases. ⁶⁹ Landahl and Herrmann had earlier published their report on the retention of airborne particulates in the human lungs in the JIHT during 1948. ⁷⁴ Parts II and III of these studies would eventually appear in the AMA-AIHOM. ^{75 , 76}

Table 4 Highly Cited Articles from the A.M.A. Archives of Industrial Hygiene and Occupational Medicine (1950–1954)*

Rank	Author(s)	Title of Article	Year	Volume	Pages
1	Smyth HF Jr, Carpenter CP, Weil CS. ^77	Range-finding toxicity data: List IV.	1951	4	119–122
2	Carpenter CP, Weil CS, Smyth HF Jr. ^78	Chronic oral toxicity of di-(2-ethylhexyl) phthalate for rats, guinea pigs, and dogs.	1953	8	219–226
3	Silverman L, Lee G, Plotkin T, et al. ^79	Air flow measurements on human subjects with and without respiratory resistance at several work rates.	1951	3	461–478
4	Vorwald AJ, Durkan TM, Pratt PC. ^80	Experimental studies of asbestosis.	1951	3	1–43
5	Smyth HF Jr, Carpenter CP, Weil CS, et al. ^81	Range-finding toxicity data: List V.	1954	10	61–68
*As identified by Scopus.

Table 5 Highly Cited Articles from the A.M.A. Archives of Industrial Health (1955–1960)*

Rank	Author(s)	Title of Article	Year	Volume	Pages
1	Yaglou CP, Minard D. ^86	Control of heat casualties at military training centers.	1957	16	302–316
2	Carpenter CP, Pozzani UC, Weil CS, et al. ^87	The toxicity of butyl cellosolve solvent.	1956	14	114–131
3	MacKenzie RD, Byerrum RU, Decker CF, et al. ^88	Chronic toxicity studies. II. Hexavalent and trivalent chromium administered in drinking water to rats.	1958	18	232–234
4	Wolf MA, Rowe VK, McCollister DD, et al. ^89	Toxicological studies of certain alkylated benzenes and benzene; experiments on laboratory animals.	1956	14	387–398
5	Curtis GH. ^90	The diagnosis of beryllium disease, with special reference to the patch test.	1959	19	150–153
*As identified by Scopus.

A.M.A. Archives of Industrial Hygiene and Occupational Medicine (1950–1954)

Highly cited articles from the fourth incarnation, the AMA-AIHOM, are displayed in Table 4. ^{77 , 78 , 79 , 80 , 81} Two articles from the list had focused on the same topic: “range-finding toxicity data.” The first most highly cited article comprised List IV of Smyth and colleagues’ ⁷⁷ series range-finding toxicity data. The fifth most highly cited article was again a range-finding toxicity data paper by Smyth and colleagues, ⁸¹ although this time it was “List V.” These articles were evidently part of a popular series written by the same authors who had earlier published List III in the JIHT during 1949. ⁸² In later years, Lists VI, VII, and 8 (written as numerical “8” by the authors, rather than VIII) would subsequently appear in the American Industrial Hygiene Association Journal and Toxicology and Applied Pharmacology. ^{83 , 84 , 85}

A.M.A. Archives of Industrial Health (1955–1960)

Highly cited articles from the fifth incarnation, the AMA-AIH, are displayed in Table 5. ^{86 , 87 , 88 , 89 , 90} The most highly cited article in this particular journal was a paper by Yaglou and Minard ⁸⁶ describing the control of heat casualties at military training centers. Aside from being the top-cited article in the AMA-AIH, their article was also the most highly cited article across all 5 versions of the journal between 1919 and 1960. Refer to Table 6. The purpose of Yaglou and Minard's article was to describe a study that helped define conditions under which heat injury may occur in the military, as well as helping to develop safe limits for physical exertion in hot environments. Yaglou had earlier published an article on the effective temperature index in the JIH during 1927, ⁹¹ and would later publish other classic papers relating to heat and thermal comfort in the AMA-AIH's successor journal, the AEH. ^{92 , 93} Constantin P. Yaglou (1897–1960) was himself a pioneering figure in thermal environment research, having developed the Effective Temperature Scale, a measure of thermal comfort. ⁹⁴ Since 1965, the Northern Point of Belding Island near Antarctica (Yaglou Point) has been named in his honor. ⁹⁵ Aside from Yaglou's thermal environment research, beryllium also made another appearance on a highly cited list, this time in an article by Curtis ⁹⁰ describing the diagnosis of beryllium disease, with special reference to the patch test.

Table 6 Summary of the Most Highly Cited Article in EOH, 1919–1960*

Authors: Constantin P. Yaglou and David Minard

Title: Control of heat casualties at military training centers ^86

Journal: American Medical Association Archives of Industrial Health

Date: October 1957

Citation: Volume 16, Number 4, Pages 302—316

Aim: To help define conditions under which heat injury may occur in the military and to develop safe limits for physical exertion in hot environments.

Methods: Investigation of heat-related illness in three Marine Corps Training Centers during the summer of 1954.

Results: Information on the analysis of heat loads and heat stress, heat casualties, tolerance limits for work in the heat, effectiveness of preventive measures, and the epidemiology of heat illness is provided.

Conclusion: The prevention of heat casualties in military training needs to be addressed via a combination of strategies including changes in the training schedule and consideration of the thermal environment (particularly the “wet bulb-globe temperature”).

*As identified by Scopus.

Category of Highly Cited Articles by Journal

Figure 1 indicates the proportion of highly cited articles by category and journal. As a combined group, over one third (68%) were original articles, 12% reviews, 12% case series, and 8% data lists. Only 2 journals (the JIH and AMA-AIHOM) contained highly cited articles that were reviews. All 5 citation classics from the AIHOM were original research articles, whereas the JIH, JIHT, and AMA-AIH all included highly cited case series.

Fig. 1 Category of highly cited articles by journal. (As identified by Scopus.) OA = Original Article; RV = Review; CS = Case Series; DL = Data List.

Citations Received in the Past 10 Years

Figure 2 indicates the trend of total citations received by all 25 highly cited articles across the 5 different journal titles during the most recent 10-year period, 1999 to 2008. This graph suggests that although the journal may have changed titles over time, classic articles are still being cited in the contemporary literature at an average rate of approximately 5 to 10 citations per year.

Fig. 2 Citations to highly cited articles over the past 10 years (1999–2008). (As identified by Scopus.)

DISCUSSION

It has been suggested that no medical specialty is any better than its published literature. ⁹⁶ As such, the identification of highly cited articles in EOH highlights important citation-based trends for both the profession itself as well as the broader scientific community. From early on, it had been recognized that not all scientific articles were being cited at the same rate, and indeed, most citations were to a few key papers in each field. ⁹⁷ One of the earliest bibliometric investigations conducted in EOH was published by McCunney and Harzbecker ⁹⁸ in 1992, who analyzed citation rates and impact factors in the British Journal of Industrial Medicine/Occupational and Environmental Medicine, the Journal of Occupational Medicine/Journal of Occupational and Environmental Medicine, the American Journal of Industrial Medicine, the Scandinavian Journal of Work, Environment & Health, as well as the International Archives of Occupational and Environmental Health. Nevertheless, citation classics were not identified. In 2001 and 2004, respectively, Takahashi et al ⁹⁹ and Navarro and Martin ¹⁰⁰ also published articles describing the literature of EOH; although again, no citation classics were identified. In 2006, Przyluska ⁴⁹ investigated the International Journal of Occupational Medicine and Environmental Health by using the Scopus database, although it was not until 2007 that a comprehensive investigation of highly cited articles, or citation classics, across a range of occupational medicine journals was first published by Gehanno and colleagues. ¹⁰¹

In Gehanno and colleagues’ study from 2007, ¹⁰¹ the authors analyzed Web of Science® citation counts for articles published in the British Journal of Industrial Medicine/Occupational and Environmental Medicine, the Journal of Occupational Medicine/Journal of Occupational and Environmental Medicine, the American Journal of Industrial Medicine, the Scandinavian Journal of Work, Environment & Health, as well as the International Archives of Occupational and Environmental Health. One major finding was that relatively few journals had attracted the majority of all citations, with 70% of the top 20 articles having been published by the British Journal of Industrial Medicine (now Occupational and Environmental Medicine). Interestingly, the AEOH was not included in Gehanno et al's ¹⁰¹ analysis, although EOH-related citation classics in some of the smaller periodicals such as Occupational Medicine (Oxford) ⁴⁵ and Industrial Health ^{102 , 103} have been recently described. This may be part of an emerging trend, whereby citation-based analysis has become an increasingly popular topic in the occupational health field. ⁹⁷ It has recently been demonstrated, for example, that impact factors are definitely rising in many journals of occupational medicine. ¹⁰⁴

Despite a relatively limited number of studies being conducted in the field of EOH, the identification of highly cited papers in the current study has revealed some interesting trends. Perhaps the most important of all is the relatively old age of EOH citation classics and their proportionately low citation counts when compared to other fields. In Gehanno and colleagues’ study, for example, ¹⁰¹ it was shown that the most highly cited paper in occupational medicine was from 1960, although it had received fewer than 1000 citations by 2007, and only 85 articles in the field had actually been cited more than 100 times. This can be compared to a 1951 laboratory methods article by Oliver Lowry and colleagues ¹⁰⁵ that had received almost 300,000 citations by 2005. ¹⁰⁶ The relative old age of Gehanno and colleagues’ top citation classic has not been confirmed by similar investigations undertaken in the smaller periodicals. In the current study, for example, the most highly cited article across all 5 titles, a 1957 report in the AMA-AIH by Yaglou and Minard, ⁸⁶ was not the oldest. This is consistent with previous studies conducted at Occupational Medicine (Oxford), ⁴⁵ where the journal's most highly cited article of all time ¹⁰⁷ was only 5 years old. A similar situation was also demonstrated at Industrial Health, ¹⁰³ where the most highly cited article ¹⁰⁸ was only 8 years old.

Such findings in EOH need to be seen in context, however, as history has shown that very few articles ever get cited at high rates. Garfield, for example, previously reported that of the 38 million “citable” articles published between 1900 and 2005, less than 1% were cited more than 200 times and half were not cited at all. ⁵ There are a few reasons for this. Firstly, it takes time for an article to actually receive any citations, due to the time lag between publishing an article, having it recognized by interested authors, having it cited in their reference list, and then having these citations actually counted. ¹⁰⁹ Secondly, there is the category of article itself. Reviews and methods articles tend to attract a proportionately greater numbers of citations, for example, with the importance of reviews in attracting high citation counts being previously demonstrated in the smaller EOH journals of Japan ¹⁰³ and the UK, ⁴⁵ whereas the relative popularity of toxicology and laboratory-based research has been recently demonstrated. ¹⁰² Special issues or theme issues may also be useful for raising awareness on a particular topic or a particular journal. ¹¹⁰ Thirdly, most research fields have intrinsically different citation rates, some of which will tend to change over time. Garfield, for example, previously described how the criteria for a citation classic rose from about 250 citations in 1955 to about 1000 citations in 1995. ¹¹¹ The development of highly cited articles from an author's perspective often provides an interesting insight into how the process works, such as Douglas's reflection ¹¹² on the evolution of his highly cited paper in psychology, ¹¹³ or Small's ¹¹⁴ examination of why authors think their articles were highly cited.

Fourthly, it is important to note that the 5 journals investigated in the current study all had different publication histories and durations. The JIH, for example, ran from 1919 to 1935, whereas the AIHOM lasted less than a year, publishing only 2 volumes between January and September 1950. Although any article has the potential to be a highly cited, it is reasonable to assume that the longer a journal lasts, the more articles it will publish and the more likely that at least one of these articles might actually become a citation classic. As such, results from the current study need to be seen in context, as the relative “importance” of each journal's citation classics should be considered with regard to how long that title existed and how many articles it published.

Aside from considering the most highly cited articles in a given field, it is also important to acknowledge other articles that, although not highly cited, are nevertheless classics that have stood the test of time. Sabin's 1960 paper on the live polio vaccine, ¹¹⁵ although representing one of the most important discoveries in public health, was not destined to become a citation classic. ¹¹⁶ Regarding articles published in the AEOH's predecessor journals between 1919 and 1959, there are many such examples in various categories. Legge's 1920 article, for example, on the notification of industrial diseases in Great Britain between 1900 and 1918, ¹¹⁷ may be seen as a pioneering effort in the field of occupational epidemiology. Another classic in this field, Sappington's 1924 investigation of sickness-related absences in the Edison Electric Illuminating Company of Boston ¹¹⁸ also highlighted the importance of reproductive system disorders and their negative effect on female workers and work productivity. Interestingly, the issue of dysmenorrhea as an industrial problem had been raised a few years earlier in the JIH by Meaker ¹¹⁹ and Sturgis. ¹²⁰ Although reproductive system disorders and their effect on female workers have been achieving greater attention in the occupational health field during recent years, ^{121 , 122} these articles from the JIH suggest that the concept had, in fact, been recognized and quantified many years earlier, even before the Moos Menstrual Distress Questionnaire was first proposed in 1968. ¹²³

Some other classic articles from the JIH, this time in the field of what would now be known as occupational hygiene, include Katherine and Philip Drinkers’ 1928 article on metal fume fever, ⁵⁸ Drinker and colleagues’ 1928 article on dust and fume inhalation in man, ¹²⁴ and Hatch and colleagues’ 1932 modification of the Greenburg-Smith Impinger. ¹²⁵ In the field of thermal physiology, Yaglou and Drinker published a classic article in 1928 ¹²⁶ describing their study of the relationship between climate and type of underclothing worn. Four years later in the same journal, Vernon described the measurement of radiant heat in relation to human comfort. ¹²⁷ Toxicology was evidently another interest area for the JIH, with Prodan's 1932 article on Cadmium poisoning ¹²⁸ offering important early insights into this disease. Various classic articles on the topic of occupational cancers would later be published in the journal, such as Baetjer's 1950 series on pulmonary carcinoma in chromate workers, ^{129 , 130} and Barsotti and Vigliani's 1952 study on bladder lesions caused by aromatic amines. ¹³¹

Analysis of highly cited articles in the current study also revealed a few surprises. Firstly, it is interesting to note that over two thirds of the citation classics were originals and only 12% were reviews. This is contrary to contemporary knowledge that suggests that review articles tend to be cited more often than originals. ⁹⁷ It may be that papers identified in the current study represent some of the earliest documented scientific evidence on these particular topics as they emerged in the literature. As such, subsequent authors would be inclined to cite the original research and the earliest documented findings. Secondly, the work of some well-known scientists does not appear to have been highly cited. Although she was no doubt a pioneering figure in EOH and had even been on the editorial board of the JIH from its foundation until her retirement, ¹³² no articles from Alice Hamilton, for example, were destined to become citation classics. This phenomenon would appear to be similar to the aforementioned fate of Albert Sabin's classic article on polio vaccine, and suggests that high citation counts are not necessarily the only indicator of scientific or humanitarian greatness.

CONCLUSIONS

Although citations are being increasingly seen as the “currency” of modern science, ¹³³ and lists of citation classics have been compiled in various fields, few, if any, authors appear to have used any databases to identify citation classics in the AEOH or its predecessor journals. In the current study, topics such as metal fume fever and various studies of beryllium were shown to have been regularly cited, whereas the most highly cited article of all was a 1957 paper describing the control of heat casualties at military training centers. Citation trends in EOH may not necessarily mirror those of the basic sciences, however, given that the most highly cited article was by no means the oldest. Furthermore, unlike laboratory science and some other disciplines, ¹³⁴ the most highly cited articles in EOH were not found to be “methods” papers. Such findings reinforce the notion that EOH is a diverse “boundary” discipline, the nexus where many different fields meet. ¹³⁵ Overall, this study suggests that although some citation patterns in EOH reflect those of other disciplines, the trend is not uniform and EOH itself appears to have some distinctive bibliometric characteristics.

Acknowledgments

Derek R. Smith is Professor of Environmental and Occupational Health, and Director of the WorkCover NSW Research Centre of Excellence, University of Newcastle, Ourimbah, New South Wales, Australia.