Connections between open access publishing and access to gray literature

The potential of open access publication to increase accessibility to peer-reviewed literature is cause for celebration. As we celebrate, we should not lose sight of the longstanding challenge of providing better access to the gray literature that provides an essential complement to peer-reviewed findings. We do not need to launch an open access movement to obtain this material, due to its lack of commercial significance. Instead, the challenge is to develop bibliographic resources of comparable depth as those available for the peer-reviewed literature.

As of September 2003, BioMed Central (BMC) had 322 institutional members from thirty-three countries [1]. An additional approach began in October 2003 with the launch of PLoS Biology by the Public Library of Science (PLoS), followed by PLoS Medicine in 2004. BMC has begun the work of motivating tenure-driven scholars to publish in open access journals; the advent of PLoS represents an effort to increase movement in this direction. One positive effect of PLoS, even before it began publishing, has been to draw increased media attention to the open access movement. The Washington Post published a front-page story about PLoS in early August [2], and, a few days later, an editorial in the New York Times lauded this effort [3].

Although stalled in Congress at the time of this writing, the recently introduced Public Access to Science Act is another sign of changes ahead. It seeks an exemption from copyright protection for the results of federally funded scientific research [4]. This act would apply to a high majority of medical research in the United States, which is funded through the National Institutes of Heath (NIH). Research findings would become part of the public domain and thus available without charge.

Although scholars in all disciplines would benefit from barrier-free access to the latest research, this benefit is especially true in the arena of medical publishing. Information in medical journals is often, quite literally, of life and death importance. Moral logic argues that such information should be freely available; market logic has turned it into a valuable commodity. In consequence, medical libraries consistently face exorbitant subscription costs [5].

The library community has responded vigorously to this challenge. One example is the Association of Research Libraries' Scholarly Publishing and Academic Resources Coalition (SPARC), which is a broad-based effort to increase competition in the scholarly publishing market. A more localized approach is the effort of librarians at the University of Virginia's Claude Moore Health Sciences Library (CMHSL) to encourage faculty to consider publishing in a BMC journal [6].

Not surprisingly, these efforts face resistance from publishing companies. For example, Elsevier argues that open access publishing is an untenable business model that threatens the viability of niche journals [7]. The ongoing tussle between libraries and publishing companies foreshadows a difficult road ahead.

In addition to for-profit publishers, scholarly societies are currently hesitant to endorse open access publishing. In contrast to for-profit publishers, these societies often support the principle of open access, however, their publishing revenues underwrite other activities of their societies [8]. It is important for librarians to be sensitive to these realities; challenging the motives of all publishers may inadvertently alienate the scholars who should be natural allies.

Despite the challenges ahead, open access will inevitably become the norm for scholarly communication. In the print-only era, publishers provided the indispensable function of distribution. In an electronic age, this indispensability is no longer true. Once a critical mass of scholars publishes in open access journals, their colleagues will follow. This is the time when viable business models for open access publishing will emerge.

As open access publishing becomes more prominent, it is likely to be well integrated into the bibliographic infrastructure already in place for locating peer-reviewed literature. For example, as part of their campaign to encourage faculty to consider open access publishing, the librarians at CMHSL highlighted the fact that articles published in BMC received immediate citation in PubMed [9]. In contrast, the tools for locating gray literature are much less developed. This material is available to anyone today, if only it can be found.

One definition of gray literature is, “that which is produced on all levels of government, academics, business and industry in print and electronic formats, but which is not controlled by commercial publishers” [10]. In general, gray literature does not undergo peer review. A major benefit of reading gray literature, which seems counter-intuitive, is that it is more likely to report studies with non-significant results than peer-reviewed literature. This likelihood may balance against the tendency of authors to publish only statistically significant findings, which inadvertently inflates the perceived value of these findings [11]. Other benefits are that gray literature is more likely to report studies that ceased prematurely, as well as innovative pilot projects [12]. In short, gray literature provides invaluable context for understanding and critiquing the peer-reviewed work found in MEDLINE.

Because there is no MEDLINE for gray literature, health sciences librarians must devote extensive time to identifying these documents. The New York Academy of Medicine's Grey Literature Report is a useful resource as are email alerting services from various foundations [13]. The dexterity required to locate gray literature highlights the value of having citations in one place. Even after gray literature is identified, the documents can be expensive to obtain [14], not due to exorbitant subscription costs but rather due to poor distribution mechanisms. The unsettling result is that much valuable research becomes relatively inaccessible [15].

Accessing gray literature is a challenge in all fields of scholarship. In the medical arena, it is a particular challenge to locate results of population-based interventions. The NIH budget continues to prioritize clinical interventions [16], which benefit individuals more than populations. Because the researchers who depend upon NIH grants are more likely to publish in mainstream journals [17], obtaining information about clinical interventions is easier than obtaining information about population-based interventions. This is true despite strong evidence that cultural and social factors are major contributors to premature death and disease [18].

Although it is tempting to pit clinical and population-based approaches against each other, the more productive course is to view them as complementary tools for improving health. (This is also the best way to understand the relationship between peer-reviewed and gray literature.) With this in mind, the challenge for health sciences librarians is to determine how to funnel more population-based work into the system that has served clinical literature well.

There are several indications that now is the time for strong movement in this direction. One is the increased societal attention to the public health infrastructure following the September 11, 2001, terrorist attacks. This attention has focused on how to respond to incidents of bioterrorism, and both librarians and members of the public health workforce have done excellent work in improving access to relevant resources.

In 2003, the National Library of Medicine's (NLM's) Medical Subject Headings (MeSH) Section incorporated several terms related to public health into the vocabulary, including the identification of “public health informatics” as a distinct type of medical informatics [19]. As MeSH has traditionally been much stronger in the clinical domain, this is an exciting harbinger of improved indexing of public health literature.

Along the same vein, the Robert Wood Johnson Foundation has funded a groundbreaking study at the Center for Natural Language Processing (CNLP) of Syracuse University. Currently in its second year, this project has identified key elements from the gray literature in public health [20]. These elements—which are extracted and synthesized into automatically generated abstracts—include the health issue addressed, a description of the intervention, the document type, the target population, and the geographic location. The principal investigators are currently conducting focus groups and surveys to gauge the sufficiency of these summaries. As one of the investigators, Anne Turner of Oregon Health & Science University, says, “It would be great if our automatically generated abstracts could be used in an indexing scheme to better organize this broad and diverse literature” [21].

Just as open access to clinical literature is only possible in an online era, the CNLP's research highlights the power of computers to improve access to gray literature. Health sciences librarians should perceive these challenges as opposite sides of the same coin: open access removes economic barriers, and improved indexing of gray literature removes bibliographic barriers. We have made great strides in the open access movement, which should continue. Now is the time to renew our attention to improving access to population-based gray literature.