Open Access Compliance..?

"As I head towards the E2E workshop on Open Access Issues and Possible Solution which will major on metadata requirements for monitoring, reporting on and evidencing compliance with various Open Access mandates, I am reminded of a recent discussion I had with a senior official from the US NIH (National Institutes of Health).  The NIH are one of the largest funders of medical and health related research in the United States of America and also fund outside the US. They could perhaps be considered to be the equivalent of the MRC, Wellcome Trust and AMRC all rolled into one… well not really but you get the idea. Anyway, back to Open Access, or as they call it in the US, Public Access.  There has been a lot happening since the February 22, 2013 memorandum by the Office of Science and Technology Policy (OSTP) which includes access to data as well as articles.  As might be imagined the NIH use PubMed and have a long tradition of supporting (and paying for) Open Access.  As in the UK they require that research outputs (mainly journal articles) produced from research that they fund is made freely available, immediately where possible, and if not then after the usual sort of embargo period.  Like the next-REF requirements for the UK (and perhaps further afield if the franchise migrates!?) the clock starts ticking at the point of acceptance (unlike the RCUK mandate which is based on the date of publication). Being a large federal funder with a lot of clout that has being doing this for a while you might imagine that compliance rates would be fairly high.  The reported level is now about 85% (if I recall correctly, up slightly from the 82% in the next link) which considering the Wellcome compliance levels of about 69% is quite startling.  When I quizzed him further on compliance criteria I was further amazed that in order to be compliant the article has to be OA immediately on acceptance, none of this 3 month window for sorting things out.  85% compliance immediately on acceptance – surely that is not right?  Well further probing confirmed this… whilst articles are not OA compliant if they are not immediately OA on acceptance, if, at any point in the future they become freely available, then at that point they become OA compliant.  Yes you heard it right, yet another take on what OA compliance means! ..."