The vast majority of freely-available biomedical articles were published by societies using traditional subscription models, a new study reports.
The article, Status of Open Access in the Biomedical Field in 2005, was published in the January issue of the Journal of the Medical Library Association and is freely available. The first author, Mamiko Matsubayashi, is a lecturer in the Graduate School of Library, Information, and Media Studies at the University of Tsukuba, Japan.
The authors randomly sampled nearly five thousand articles published in 2005 and manually searched, one year later, several indexes (including Google Scholar and OAIster) to determine free fulltext copies. 
The researchers report that more than 70% of the free articles were found on publisher websites. Personal websites and institutional repositories contributed only 5.9% and 4.8%, respectively, to the total.
The authors write:
Although many OA advocates have considered self-archiving, or the green-road, as a feasible means of advancing OA, this method did not contribute substantially to OA in the biomedical field in 2005
On the other hand, 88% of the biomedical articles found in institutional repositories could not be found freely available anywhere else on the web, suggesting that institutional repositories are providing a unique — not redundant — service. In contrast, the authors report that 92% of the free articles found in PubMed Central were also available on journal websites. The authors conclude:
OA in the biomedical field in 2005 was achieved under an umbrella of existing scholarly communication systems, the majority of which still use traditional paid-access journals.
While this article does not focus on details, its main contribution is clear: society publishers have lead the effort in promoting free access to the biomedical literature. Efforts like HighWire’s ‘Free Online Full-text articles’ doesn’t call itself “Open Access” but remains, by far, the largest source of free scientific articles on earth.
Those who have used the subscription model in unison with free-access ventures appear to have hit a sweet-spot in a wide spectrum of extremes. While not perfect, this may be, as the German philosopher Leibniz coined, “the best of all possible worlds.”
 Editor’s Note: Sampling Bias? The sampling method used in this study included only articles with pagination numbers between 11 and 19. Considering that large journals publish hundreds or thousands of articles per volume, this sampling technique leads to a strong sample bias toward smaller journals. The result of this study may have under-reported the proportion of articles published in society and high-impact journals (which tend to be larger), and over-reported the effect of smaller, and lower-impact journals, including those published by BioMed Central.