Open Access

Supporters of Open Access to scientific literature often portray it as the definitive and inevitable model for scientific publishing, but it is far from being the last word on new modes of access. In reality, stakeholders in scientific publishing are in the midst of adjusting to the revolutionary new possibilities offered by the Web and the online journal article for scholarly communication.

In this Nature forum, a range of stakeholders in scientific publishing have made their cases at length, and often persuasively. Agreement in the industry on the best route forward remains distant, however, and the level of emotion behind the debate has served in some cases to obfuscate discussion. This article aims to provide an independent assessment of the key arguments, and to flag up areas where questions remain unanswered.

Proponents of a move to open access argue that this will benefit science and society in general. A report published by the UK Wellcome Trust assumes that “the benefits of research are derived principally from access to research results”, and therefore that “society as a whole is made worse off if access to scientific research results is restricted”.

Part of the remit of not-for-profit organisations such as the Wellcome Trust which fund research may be the full dissemination of results. But even where research is publicly-funded, taxes are generally not paid so that taxpayers can access research results, but rather so that society can benefit from the results of that research; in the form of new medical treatments, for example. Publishers claim that 90% of potential readers can access 90% of all available content through national or research libraries, and while this may not be as easy as accessing an article online directly it is certainly possible.

Funding for scientific research also comes from a variety of sources – in some countries such as Australia and New Zealand around 80% of R&D funding comes from the public purse, while in Japan and Switzerland only about 10% is government-funded. It is therefore not necessarily the case that taxpayers fund most scientific research.

Another criticism of open access is that payment for publication could create conflicts of interest and have a negative impact on the perceived neutrality of peer review, as there would be a financial incentive for journals to publish more articles. The importance of the role of peer review does not diminish under an Open Access model, and structures need to be in place to ensure that peer reviewers are not unduly influenced by the needs of their publishers.

In some ways though this argument can apply as much to the current subscription-based system as publishers often justify price increases on the grounds of an increase in the number of journal articles published. This suggests that there are financial advantages for both Open Access and subscription-based publishers in publishing more articles.

A solution to the serials crisis?

Open Access is also often seen as a solution to the 'serials crisis,' a situation where many libraries have been forced to cut journal subscriptions because of price increases. Subscription charges - and the volume of content published - have undisputedly risen more sharply than library budgets. But it is worth questioning to what extent the serials crisis is also due to the level of library budgets – typically around 2% of a university's budget. Some would argue that this level is insufficient in the information age in which we live.

Advocates of Open Access charge that commercial publishers using traditional business models are making unreasonably large profits. Profit levels in scientific publishing are significant, and have been so for many years. Publishers are in business to generate profit, or in the case of learned society publishers, a surplus to support the society’s other activities.

They cannot be blamed for running their businesses in such a way as to maximise profits. Indeed, in recent years the investment made in online and electronic information services could be viewed as balancing out high levels of profits in previous years. Elsevier, for example, has invested at least £45 million in its ScienceDirect service over the past five years, and has also developed the free-to-use science search engine Scirus.

Moreover, although journals can be accessed free of charge under Open Access, publishing these often involves charging authors or their institutions article processing charges. Institutions pay in particular under a scheme introduced last February by the UK Open Access publisher BioMed Central; here institutional members pay a flat fee covering the publication charges for all their scientists.

Payment for publication under any institutional membership scheme is likely to have to come from the library budget rather than from research funding. Moreover, these fees, under a “per article published” basis, will rise year-on-year should an institution's scientists embrace publishing in Open Access journals. Some institutions have taken out a membership to BMC as their way of supporting the open access movement, but memberships might fall off if the full cost to the library rises significantly and regularly.

" To the extent that an institution's faculty embrace the BMC journals, the institution is faced with the same dilemma that we're used to under the subscription model - where is the money to pay the ever-increasing fee going to come from?", T.Scott Plutchak, director of the Lister Hill Library at the University of Alabama at Birmingham, stated recently adding that: “I’m led to think that my best option for supporting the BMC journals would be to NOT try to find the funds for an institutional membership, but to continue education efforts on campus and encourage individual researchers to pay the processing charges out of their grants. I can’t think of a practical reason to continue the institutional membership”. 

If this view is widespread, open access publishers may have to spend more effort selling the benefits of open access to academics and research funders, as well as to their traditional librarian customers. As many information providers have found before them, this is a significant marketing exercise, and could, for Open Access publishers, mean a rise in article processing charges to cover these new expenses.

Another controversial aspect of who should pay for Open Access is the concern that research-intensive academic institutions will end up subsidising access for other organizations with lower research outputs, including commercial companies. Institutions which focus heavily on research, may find that their expenditure on journals increases under an Open Access system, while others focussing more on student education, may find that their spending on journals decreases under Open Access.

Commercial organisations such as pharmaceutical manufacturers use research to drive product development, but this research is normally done for in-house purposes and is not issued for publication, at least not until after a product has gone to market. It is expected, therefore, that commercial organisations’ spending on journals will also fall under Open Access. The same argument can also be applied to countries – the UK, according to the Elsevier response to the Science & Technology Committee enquiry, currently accounts for 3.3% of global spending on journal subscriptions, but contributes 5% of published material.

Open Access unsustainable?

Many critics are dubious of Open Access because they do not believe that the model is economically sustainable, and that, if relied upon, could damage the market as publishing businesses and learned societies experience difficulties due to reduced revenues.

Jan Velterop, in a recent contribution to this Nature forum, claimed that “the alleged ‘unsustainability’ of the ‘input-paid’ Open Access publishing model is … a second-order myth” which he believes has been countered. However, even after several years BMC has not moved into profit; this is not expected for another two or three years.

Publishers which are publicly quoted companies see this issue as a key concern – the company’s share price is dependent on its financial performance, and any suggestion of adopting a business model which has not yet proven that it can support a business profitably will deter support from City bankers and investors.

However, traditional publishers are working with their customers to experiment with business models. In June 2004 Elsevier re-defined its policy on post-prints and institutional repositories:

“An author may post his or her version of the final paper on personal web sites and on the institution’s web site (including its institutional repository). Each posting should include the article’s citation and a link to the journal’s home page (or the article’s DOI). The author does not need our permission to do this, but any other posting (i.e. to a repository elsewhere) would require our permission. By “his version” we are referring to a Word or Text file, not a PDF or HTML downloaded from Science Direct – but the author can update the version to reflect changes made during the referencing or editing processes. Elsevier will continue to be the single, definitive archive for the formal published version.”

This policy puts the onus back on to the author – all those wanting to make their published research freely available to the wider community now have the opportunity to do so. Springer has also recently made a move in this area with the announcement of its Open Choice program, which similarly puts the onus on authors by giving them the choice of whether they want to publish their article free in a traditional manner – where it is available to subscribers only - or to pay a fee - currently $3000 - to make the article available on an Open Access basis.

OUP is another major player to investigate the possibilities of Open Access, moving its flagship Nucleic Acids Research journal to a full open access model in January 2005 (at a rate of $1500 per paper) following an 18-month period of experimentation with a hybrid OA model. Nature Publishing Group and the European Molecular Biology Organization this month also announced their intention to launch an international open access journal, Molecular Systems Biology.

Open Access and capacity for publishing innovation

The April 2004 Wellcome Report found that “an appropriate and conservative estimate of the charge per article necessary for author-pays journals lies in the range $500-$2500, with these costs varying depending on the type of journal. However, while the report authors have allowed for costs of peer review, production and administration as well as profit, no mention of investment in online systems and service development has been made.

Open Access, argue critics, does not allow for sufficient investment in technology. Part of the reason for publishers generating profits is to ensure that investment for new product development can take place – in the digital era this has proved more crucial than before, with the high costs of developing online services. This kind of investment is not only beneficial to end users, but also provides a platform through which publishers can compete with one another.

Competition in the scientific publishing world has previously been constrained to a certain extent because highly-renowned journals are not substitutable. However, publishers and other interested parties have invested significantly in recent years to develop services such as Thomson’s Web of Knowledge or Elsevier’s Scopus which allow access through a single point to content from a range of publishers – it seems likely that, since many readers are unaware of the publisher of the journals which they use, that capturing the end user in this way rather than through individual journal titles will be an important guarantor of long-term business success.

The concern here with regard to Open Access is that if, as expected, OA services are not as profitable as traditional activities then less investment will be made in these systems, the publishers’ ability to compete will be affected, and end users will suffer as services stagnate and price rises continue.

Capacity for innovation is particularly important in this information age. Market demands from the academic and scientific community are changing and developing alongside the growth of the Internet, and journal publishers are adapting their views of how journal article content fits into the information spectrum. There are now examples of publishers offering more than just the journal content, or bypassing the publication of research results as articles but instead providing citable references within a database.

An example of this is the Cell Signaling Gateway, a primary data environment created by AfCS (Alliance for Cellular Signaling) and Nature Publishing. The Gateway contains a data repository combined with toolsets and analytical programs, detailed reports on experimental procedures and protocols, and reference datasets needed by bench scientists.

Alongside this, pages edited by Nature staff include structured data on 3500 proteins, with key information derived from other database sources supplemented by summaries from 1500 experts and author-entered data. The results are edited and collated by Nature, which also organises the peer review of this content and assigns DOIs (Digital Object Identifiers) to it so that it can be cited effectively in the research journals.

While start-up was supported by the US taxpayer through NIGMS, the service is run largely through sponsorship by Genentech and Eli Lilly; in addition some pay-per-view activity is being developed alongside links to articles and third party content. This service demonstrates that lateral thinking and a proper disrespect for the rules of the game can create high value services.

Impact on Learned societies

The fact that revenue generated by charging author fees is unlikely to generate the same levels of income that subscriptions have traditionally provided, also creates problems for learned societies, in particular.

Profit levels are a major issue for learned societies: a straw poll of ALPSP members in February 2004 revealed that the vast majority of learned societies (87.5% of respondents) generate a surplus from their publishing activities; this surplus is necessary in order to support these societies’ other, non-revenue generating or loss-making activities.

For the many societies which use journal publishing revenues to support their other activities, the foremost concern about Open Access is in identifying revenue sources to replace the income lost by abandoning the practice of selling subscriptions. This could include increasing membership fees - which would probably be unpopular move given that the journal would no longer form part of the membership package) - raising delegate rates for conferences and events, or selling on other non-related services - some societies offer cut-rate car insurance, for example, on which they receive a commission.

In the worst-case scenario, losing a significant amount of revenue from journal publishing could destabilise the society as a whole. Implementing a new business model will be something the majority of societies will, therefore, approach with care and not a little trepidation.

The majority of learned societies continue to publish their journals simultaneously in print and online and will in all likelihood need to continue charging for print editions, perhaps on a print-on-demand basis, even if access to the electronic edition has been made free of charge. At present, many learned society members continue to prefer to receive their edition of the journal in print. Making a move directly to open access publishing and charging extra to receive a print copy could alienate a large percentage of the membership base.

It seems reasonable to assume that any system which cannot demonstrate economic sustainability in the long term will not prove successful – at present the Open Access system in not in a position to provide that proof for the satisfaction of the majority.

The Internet has created a difficult dichotomy, summarised by Stuart Brand in his 1987 book “The Media Lab: Inventing the Future at MIT”. Brand said that, “Information wants to be free because it has become so cheap to distribute, copy, and recombine - too cheap to meter. It wants to be expensive because it can be immeasurably valuable to the recipient. That tension will not go away.”

The ease by which content can be accessed has created the view that it should be free – despite that fact that everyone knows that creating and distributing content can be an expensive process. We can see the physical results of that process in a book, and understand where the cost (or at least some if it) lies – this does not happen in the same way online. However, scientific publishing is a demonstrably valuable service and one which does not come cheap, particularly in this era of electronic development. Any emerging models will have to be grounded firmly in economic reality to have any chance of success.