Access to scientific publications: the scientist's perspective

Abstract

Background: Scientific publishing is undergoing significant changes due to the growth of online publications, increases in the number of open access journals, and policies of funders and universities requiring authors to ensure that their publications become publicly accessible. Most studies of the impact of these changes have focused on the growth of articles available through open access or the number of open-access journals. Here, we investigated access to publications at a number of institutes and universities around the world, focusing on publications in HIV vaccine research--an area of biomedical research with special importance to the developing world.

Methods and findings: We selected research papers in HIV vaccine research field, creating: 1) a first set of 50 most recently published papers with keywords "HIV vaccine" and 2) a second set of 200 articles randomly selected from those cited in the first set. Access to the majority (80%) of the recently published articles required subscription, while cited literature was much more accessible (67% freely available online). Subscriptions at a number of institutions around the world were assessed for providing access to subscription-only articles from the two sets. The access levels varied widely, ranging among institutions from 20% to 90%. Through the WHO-supported HINARI program, institutes in low-income countries had access comparable to that of institutes in the North. Finally, we examined the response rates for reprint requests sent to corresponding authors, a method commonly used before internet access became widespread. Contacting corresponding authors with requests for electronic copies of articles by email resulted in a 55-60% success rate, although in some cases it took up to 1.5 months to get a response.

Conclusions: While research articles are increasingly available on the internet in open access format, institutional subscriptions continue to play an important role. However, subscriptions do not provide access to the full range of HIV vaccine research literature. Access to papers through subscriptions is complemented by a variety of other means, including emailing corresponding authors, joint affiliations, use of someone else's login information and posting requests on message boards. This complex picture makes it difficult to assess the real ability of scientists to access literature, but the observed differences in access levels between institutions suggest an unlevel playing field, in which some researchers have to spend more efforts than others to obtain the same information.

Figure 1. Availability of HIV vaccine research articles online.

Two randomly selected sets of papers were created and checked for availability online. Most of the recently published papers required subscription for access, while the “cited” set of papers (randomly selected from citations within the “recent” set) were more accessible, presumably due to journals moving articles from subscription-restricted content into open access. A small proportion of cited articles were not available online.

Figure 2. Subscription-based availability of HIV vaccine research articles at a number of institutes and universities around the world.

The “recent” and “cited” are the two sets of papers described in the legend to Fig. 1. Only papers which were not available through open access sources, but were available online, were included in this analysis.

Figure 3. Outcomes of contacting authors with an email request for a copy of the paper.

The “recent” and “cited” are the two sets of papers described in the legend to Fig. 1. Emails were obtained from journal's website, NCBI, by web searches and in institution's directories. “Positive responses” include responses with a pdf or Word document of the requested paper attached to the email. “No response” includes bounced emails (due to email address no longer being active), lack of reply and, in one case, refusal to send a copy due to publisher holding the copyright. “Email not found” includes articles for which the email of the corresponding author could not be located.