I was introduced to the open access movement in medical publishing in 1999/2000, during a stint as an in-house editor at the Canadian Medical Association Journal (quick plug: CMAJ on-line was from the start an open access journal). At that time, Harold Varmus (head of the NIH) had proposed eBioMed, an electronic archive and pre-print server, in the style of the successful high energy physics archive arXiv, which had since its inception in 1991 become the archive of record for that particular highly specialized area. BioMed Central’s Freedom of Information Conference 2000 captures the debate that attended that proposal, as did a debate (that does not seem to be archived) on HMS Beagle, an electronic science newsmagazine published by the late, lamented BioMedNet. Robert Kling, Joanna Fortune, and Adam King trace The Remarkable Transformation of eBioMed into PubMedCentral (working draft; the final paper was subsequently published with the usual access restrictions), which stripped the proposal of its most controversial aspect, the preprint server, and made it an open-access repository of material voluntarily submitted by traditional publishers. To quote from Kling et al:

The differences between the May 5 E-Biomed proposal and the August 30 PubMed Central proposal (see Table 1) can be described as a shift from one that emphasized authors’ and readers’ interests to one that favored the interests of societies’ and publishers’.  Though both NIH proposals were framed in terms of the interests of the “scientific community,” the composition of the “scientific community” is defined differently in the two proposals. The May 5 E-Biomed proposal defines the scientific community as authors and readers, excluding scientific societies:

“a mechanism for governance (the E-Biomed Governing Board) that involves all of the parties concerned-the scientific community (readers and authors), editors, computer specialists, and funding agencies.”

This view of the scientific community is typical of electronic publishing enthusiasts in the scholarly communication literature (Brody, 1996; Odlyzko, 1996), and seemed to be accepted uncritically by many supporters of the original proposal. Scientific societies, on the other hand, were careful to stress their role in the scientific community, and tended to define the scientific community in terms of journal readership and society membership. For example, the officers of the American Society for Pharmacology and Experimental Therapeutics stated:

“The scientific community is diverse and should not necessarily speak with a single voice. Individual scientific society disciplines and their respective journals should be celebrated and encouraged to succeed. (ASPET 1999)”

Thus, the perception of whether the E-Biomed archive would be a boon or a bane for the scientific community depended to a great degree on just how the term “scientific community” was being defined. Supporters and  opponents of the proposal  tended to characterize the “scientific community” so as to bolster their particular positions.

After an initially bumpy start, with criticism from disappointed supporters of e-Biomed and skeptical supporters of commercial publishing alike, PubMed Central has grown steadily. Varmus and others advocates of open source went on to found the Public Library of Science, and then to start two journals, PLOS Biology and PLOS Medicine, refereed, open-access, and published under a creative commons license. More are to come, including one on clinical trials. Recently, the Wellcome Trust in the UK and the National Institutes of Health in the US have both announced their support of open access, the NIH by requiring submission of the final version of any manuscript accepted for publication to PubMed Central. And in a critique of the relationship between the pharmaceutical industry and the medical journals, former BMJ editor Richard Smith returns to the model put forward for e-Biomed as a means of reducing bias: an independent open-access archive of trial results, with an edifice of commentary and criticism built around it. Certainly we have the tools to do so: the required elements in reporting clinical trials could be wrapped in XML and repackaged in any way required. Standardization to a well-worked standard would make reanalysis and systematic reviews possible, and the annotation tools and methods developed for bioinformatics could be deployed.