Objectives: (1) To determine the trends of the number of journals that publish injury prevention and safety promotion (IPSP) articles and the number of articles published each year from 1900 through 2006; (2) to determine the coverage of key IPSP journals in widely used literature databases.
Methods: Journals were identified through a structured procedure that included hand-searching selected classification categories of two major listings of periodicals that publish four or more IPSP-relevant articles per year. Article relevance was assessed using the inclusion criteria for SafetyLit (a database of scholarly literature selected for its relevance to the IPSP field). Each identified journal was hand-searched from its first issue, and IPSP-relevant articles were added to the SafetyLit archive database. A MySQL database was used to perform basic queries and statistical summary analyses for journals published 1900–2006. The number of IPSP journals and journal articles was plotted for these years to identify publication trends. The publication year range of each IPSP journal was compared with the coverage, if any, of the journals in each of the five databases.
Results: Of the 17 839 journals assessed, 597 met the threshold for inclusion in this study. This amounts to thousands of IPSP-related articles published each year: 6100 in 2006. Only 160 (26.8%) of the journals are included in all five databases, and none have all publication years included in all of the databases.
Conclusions: Some key IPSP journals are completely excluded from some of the databases. Thus, a search in a single database may miss key information from one of the many disciplines that publish IPSP-relevant information.
Statistics from Altmetric.com
In 2003, Barry Pless, the then editor of this journal (Injury Prevention, an injury prevention and safety promotion (IPSP) journal with a public health focus) asked “Is this journal really needed?”.1 He used the contents of SafetyLit2 to identify 126 journals that publish articles related to injury prevention. This inspired questions about the nature and quantity of the IPSP content of these and other journals and about the availability of IPSP-relevant articles in SafetyLit and other commonly used literature databases.
The IPSP field draws knowledge from at least 30 professional disciplines: agriculture, anthropology, codes and standards development, consumer product testing and safety, demography, dentistry, economics, education, engineering specialties, ergonomics, fire suppression and prevention, geography, geology, industrial design, interior design, law, management and administration, marketing, media studies, medicine, meteorology, nursing, occupational safety and hygiene, oceanography, pharmacology, physiology, political science and policy, psychology, public health, public safety, social work, sociology, sports and kinematics, toxicology, transportation safety, urban planning, and other fields.3 Authors from these disciplines publish their work in hundreds of journals. The contents of these journals are included in many literature databases, but database coverage differs greatly depending on the specific focus of the particular database.
The objectives of this study were to determine: (1) the trends of the number of journals that publish IPSP articles and the number of articles published each year from 1900 through 2006; and (2) the coverage of key IPSP journals in widely used literature databases.
This study is an inventory including the following steps: (1) determination of a comprehensive list of IPSP journals; (2) creation of a comprehensive list of journals that publish four or more IPSP-relevant articles per year; (3) quantification of the of IPSP content of each of these journals based on the SafetyLit inclusion criteria; (4) enumeration of the years that each journal (identified in step 2) is included in SafetyLit and five other commonly used literature databases.
Database selection, journal coverage, and journal list
Although there are many available databases, a recent survey of SafetyLit users found that fewer than 10% of respondents reported using any database other than EMBASE,4 ERIC,5 PsycINFO,6 PubMed/Medline,7 and the Web of Science,8 during the past 2 years.9 Therefore, the databases discussed in this report were limited to these six sources.
The journal listings of each database were examined to determine if each journal that met the study criteria is included in the database and, if so, to identify the years of the journal’s publication that each database contains.
Pless only had access to the public SafetyLit website when he identified IPSP journals,1 but we have access to the full SafetyLit administrative database, enabling a more detailed search for IPSP journals than could be performed using the SafetyLit website alone. The list of journals that contain articles of relevance to IPSP was developed by: (1) examining the journal classification categories in each of the two major listings of serials, the Online Computer Library Center’s WorldCat10 and Ulrich’s Periodicals Directory11; (2) selecting those categories that appear to have relevance to IPSP; and (3) identifying journals in each category that potentially contain IPSP-relevant articles. The category listings, WorldCat (n = 290) and Ulrich’s (n = 118), are available as supplementary material online at http://ip.bmj.com/supplemental. Journals were also identified through a search of the reference lists of articles, books, and reports, particularly those published during the first half of the 20th century.
The contents of each potential journal were assessed to determine if it included peer-reviewed material (as defined by the journal). If this criterion was met, each journal issue that was published in the most recent 3 years was examined by hand. Journals that published 12 or more articles that met the IPSP-relevancy selection criteria during the 3-year interval—an average of four articles per year (one per issue of a journal published quarterly)—were added to the journal list. Journals with a publishing life of less than 3 years were included only if four or more relevant articles were published per year. Journals that began publishing in 2007 were excluded. Journals that have changed names but retained the volume number sequence were counted as one journal. The journal list was maintained in a MySQL12 database which is an integral part of the SafetyLit computing infrastructure.
The articles were regarded as relevant and included in the inventory if any of three reviewers determined that they conform to the SafetyLit inclusion criteria (box 1).13
Box 1: Article relevancy criteria for inclusion in the inventory based on the SafetyLit article inclusion criteria13
Articles, editorials, letters, and policy statements written in the English language (or, if in another language, published with an English abstract) were included if they concerned:
intentional or unintended injuries (defined as damage to the body caused by exposure to environmental energy—kinetic, thermal, chemical, electrical, or radiation—in amounts that exceed the human body’s resilience14 or by poisoning, drowning, suffocation, exposure to temperature extremes);
injury prevention—that is, they concerned any of the pre-event or event elements of the Haddon Matrix15 (host factors, vehicle factors, and physical and psychosocial environmental factors);
safety promotion—that is, promoting a shared set of beliefs, attitudes, values, and ways of behaving that support the prevention of injury16;
the epidemiology of injury or injury risk factors; or
the financial, personal, or societal costs or consequences of any intentional or unintended injury or injury risk factor.
Items on other topics that may help a reader to make decisions about research or prevention strategies and priorities were selectively included.
We excluded articles on non-acute injuries such as those from repetitive stress or chronic exposures, medical or surgical treatment for injuries, or complications of medical care except when the article also contained information on one of the inclusion criteria. Most book reviews, conference reports, news summaries, and the like were not included.
As IPSP-relevant articles were identified they were added to the SafetyLit archive database. Articles were identified by hand-examining the contents of each journal that had been selected on the basis of a minimum of 12 articles in 3 years, from its initial issue through December 2006. Journal contents were obtained from the publishers or through a page-by-page review of the actual printed journals or online versions. The total (all years aggregated) number of relevant articles was determined for each journal by using administrative tools in the SafetyLit archive database that allow basic queries and statistical summary analyses to be performed. The number of non-relevant articles—that is, articles that did not meet the SafetyLit article inclusion criteria—was recorded for each journal while they were being hand-examined.
The numbers of journals and journal articles were counted for each year 1900–2006 and plotted to highlight annual trends.
Calculation of the proportion of IPSP-relevant articles and mean number of IPSP-relevant articles per year
The listed journals differed in periodicity (weekly, monthly, quarterly, etc) as did the number of pages in each published issue. Thus, to better reflect the content of each journal, the proportion of the total journal items (all years aggregated) that were IPSP relevant was also calculated. Book reviews, conference reports, and news items were excluded from both the numerator and denominator. The mean number of relevant items per year in each journal was calculated. In view of a large upward trend in the annual number of IPSP articles that began in 1950, the lower year limit was set at that year so that relevant articles from long-lived journals (eg, BMJ, JAMA, and Lancet) that were published before 1950 would be omitted from calculation of the means. This was done so that the rarity of relevant articles during the last half of the 19th century and first half of the 20th century would not have a strong deflating effect on the item means and proportion of IPSP-related content of these journals.
Of the 17 839 journals assessed, 597 met the threshold for inclusion in this study. Figure 1 shows that the number of journals that published, on average, four or more IPSP-related articles per year increased throughout the study period. The number of published IPSP journal articles remained between 29 and 40 per year until 1950, when the number more than doubled and started an almost exponential climb through 2006. During the first half of the 20th century, as more journals began publishing IPSP articles, the mean number of articles per year decreased. However, by mid-century not only did the number of IPSP journals increase, the mean number of articles per journal also increased.
Table 1 shows the 16 journals that publish 40 or more IPSP-relevant articles a year, the range of years that the journal has been published, the mean number of items published per year, the proportion of total articles published that are IPSP relevant, and the range of years that the journal has been included in five literature databases. An expanded version of table 1 with journals that publish four or more IPSP-relevant articles per year is available as an online supplement at http://ip.bmj.com/supplemental. The expanded table shows that only 160 (26.8%) of the 597 journals are included in all databases and that none of the journals have all publication years included in all databases.
Table 1 excludes the ERIC database because our examination found that, for many journals, the range of years had gaps in coverage—only 2 or 3 isolated years are included for some journals. This stop–start–stop–start pattern of journal inclusion would make an already complex table far too cumbersome. There is, however, a more important reason to exclude ERIC. Just as SafetyLit excludes articles on chronic exposures and repetitive stress that appear in an otherwise rich source of IPSP articles such as Safety Science, ERIC systematically excludes some articles highly relevant to IPSP on the basis of the specific educational focus of the database. For example, ERIC excluded some Journal of School Health articles about youth suicide,1718 exposure to violent television shows,19 and weapon carrying20 even when other articles from the same issues of the journal were included.
The growth in the number of journals publishing IPSP-relevant articles (fig 1) and the publication start dates and database coverage (table 1) suggest that the need to search multiple databases has increased over time. Although it is possible that, for very highly specialized projects, an exhaustive search using multiple databases may not always be necessary, a broad search will probably always be useful. This is because each of the 30-plus professional disciplines have their own professional journals, and sometimes these journals publish “cross-over” material—that is, an engineering journal may publish an article that describes the effect of risk-taking by drivers and the implications this may have for the design of vehicles or roadways. A limited search is not likely to find these potentially important articles.
A recent survey found that the two literature databases most used by IPSP searchers were PubMed and PsycINFO, and that fewer than 10% of respondents used any other databases.21 That considered, and in view of the information in table 1, it is clear that searchers who use only one or two databases are going to miss material from key journals. Although SafetyLit has the most comprehensive journal coverage of the databases examined in this study, there are key journals or journal years missing from each database. Thus, a searcher must query at least two databases to examine all years of all relevant journals.
Information retrieval depends on many factors, including the topic of the query, the range of search techniques used, and the information sources used. The ability to find published articles relevant to IPSP is important from at least two reciprocal perspectives: (1) we, ourselves, must find the articles we need and (2) we want our own articles to be found by other searchers. There are many practical and ethical reasons for conducting a thorough search of the literature before beginning any project or forming policy, but this is seldom done even by a well-intentioned searcher.22–24 One reason may be that searchers limit their queries to databases with which they are familiar.25 The procedures for conducting systematic reviews for Cochrane and Community Preventive Services include consulting multiple literature databases to find articles that meet the authors’ inclusion criteria.2627Table 1 shows that many journals are not fully included in all of the commonly used databases. Thus, for a topic to be fully researched, it is necessary to follow the example of the Cochrane and Community Preventive Services authors and systematically search multiple databases when conducting a review of the literature.
What is already known on this topic
The number of published journal articles relevant to injury prevention and safety promotion (IPSP) has increased dramatically each year since 1950.
What this study adds
Each of the most-used literature databases omits key IPSP journals.
It is necessary to conduct queries of several databases to perform a comprehensive search of the IPSP literature.
Strengths and weaknesses of the study
This study provides a comprehensive listing of the journals that publish IPSP-relevant articles, with metrics that allow the journals to be compared with one another. It also provides current information about the databases that include material from the listed journals.
Its primary weakness is that the four articles per year threshold for inclusion in this report is an arbitrary number (based on an average of one article per issue of a journal that is published quarterly) to keep the table size manageable. The number of articles published per year and the proportion of IPSP articles compared with the total number of articles published provides only part of the picture. Some journals, such as BMJ and JAMA, may publish fewer articles, but they are very important ones. A quantitative assessment of the impact of the journals listed was not addressed. Inclusion of an impact metric such as the proprietary ISI Impact Factor ratings was rejected for several reasons: (1) although impact factor measures may be relevant for the journals that devote most of their content to IPSP, the overall impact factor for other journals may have little relevance to the impact of the IPSP-related articles published in them; (2) impact factor measurements are not available for many of the journals listed in this report; (3) the listed measures (mean number of articles published per year and the proportion of total articles that are IPSP relevant) are more stable and, perhaps, a more useful guide for making a decision about whether to subscribe to a journal.
The scope of this study is limited to journals that publish articles written in the English language or that have English language abstracts. The number of IPSP-relevant journals published in other languages is unknown. The scope was also limited to peer-reviewed scholarly journals. Although important work is published in the grey literature (conference proceedings, agency reports), those sources are not well represented in literature databases.2829 Although every effort was made to conduct a systematic and comprehensive search for journals that publish IPSP-related items, it is possible that some were overlooked.
The same process of identifying IPSP-relevant journal articles was applied throughout the range of years under study. The search was facilitated by access to four university-based libraries with extensive engineering, medicine, science, and social science collections and strong representation of material published in the early 20th century. The study protocol called for examination of the reference lists of books, journal articles, and reports to identify potentially relevant journals. Because there was a smaller total body of literature in the early 20th century and, with passing years, more opportunities for articles from that period to be cited, it is likely that fewer articles from that era were overlooked than in more recent years.
The five databases were selected on the basis of the results of a survey of SafetyLit users, a group that may not be representative of IPSP researchers and practitioners in general. However, it is likely that SafetyLit users represent a group that has more interest in closely following the research literature than non-users.
The database listings are a snapshot of the current (1 November 2007) journal inclusion status. Database coverage of journals is fluid. Although the year-to-year changes are usually minor, additional journals are added each year and some are dropped. For example, PsycINFO began indexing Injury Prevention in 2007, and PubMed discontinued indexing Blutalkohol (a journal that publishes articles concerning alcohol, drugs and driving) in 1996.
The quality of indexing of IPSP-related articles will differ among the databases. The articles in some databases—for example, SafetyLit and Web of Science—are not indexed with terms from a controlled vocabulary. This makes finding all appropriate articles much more difficult.28 An earlier report described the terms used to perform text word queries of the SafetyLit database and found that most textword queries were insufficiently thorough to find more than half of the available items on the most popular search topics.30 Even when articles are indexed, the descriptor terms may not be assigned as the searcher expects. A recent study of the accessibility of journal articles about alcohol use by automobile drivers reported on an author name search in PubMed. Of the 19 articles found that specifically mentioned automobile driving in the title or abstract, only 11 had been assigned the index term (MeSH) “automobile driving”.21 Work is underway to develop a suitable indexing vocabulary for SafetyLit.3
IMPLICATIONS FOR PREVENTION
Since the 3Es (education, engineering, environment) concept was developed almost 100 years ago,31 a multifaceted approach to injury prevention has become the standard. However, if searchers limit the information they gather to only that from one or two familiar sources, they may miss key information from one or more of the many parallel disciplines that publish relevant information. This could negatively affect the quality of their policy, program, or research decisions.32
A search of a single database may return a satisfyingly long list of articles. However, the nature and focus of each of the many available literature databases will confine what is returned to sources selected to support the database’s purpose. This report should raise awareness among those who search the scholarly literature for IPSP-related journal articles that it is necessary to search several databases to find all relevant material.
Nilam Patel assisted with the data gathering, maintaining the SafetyLit database, and reviewing the manuscript.
Competing interests: DL serves on the editorial board of Injury Prevention and the journal is among those listed in this report. He is the editor of SafetyLit. SafetyLit is currently supported through contracts with several government agencies of the State of California, USA. This research, however, was self-supported. It was not supported by those government contracts. The SafetyLit electronic mailing list and website contain no advertising. SafetyLit is a free service. LL served on the editorial board of Injury Prevention in the past.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.