Introduction: This study clarifies the trends observed in open access (OA) in the biomedical field between 2006 and 2010, and explores the possible explanations for the differences in OA rates revealed in recent surveys.
Methods: The study consists of a main survey and two supplementary surveys. In the main survey, a manual Google search was performed to investigate whether full-text versions of articles from PubMed were freely available. Target samples were articles published in 2005, 2007, and 2009; the searches were performed a year after publication in 2006, 2008, and 2010, respectively. Using the search results, we classified the OA provision methods into seven categories. The supplementary surveys calculated the OA rate using two search functions on PubMed: "LinkOut" and "Limits."
Results: The main survey concluded that the OA rate increased significantly between 2006 and 2010: the OA rate in 2010 (50.2%) was twice that in 2006 (26.3%). Furthermore, majority of OA articles were available from OA journal (OAJ) websites, indicating that OAJs have consistently been a significant contributor to OA throughout the period. OA availability through the PubMed Central (PMC) repository also increased significantly. OA rates obtained from two supplementary surveys were lower than those found in the main survey. "LinkOut" could find only 40% of OA articles in the main survey.
Discussion: OA articles in the biomedical field have more than a 50% share. OA has been achieved through OAJs. The reason why the OA rates in our surveys are different from those in recent surveys seems to be the difference in sampling methods and verification procedures.