Best Practice Diffusion Rates-How fast do good ideas spread?

I wonder how fast good ideas get to those that could benefit from them. I once read that it can take ~15 years for best practices to permeate primary care practice in medicine but that was in a book probably 4 years old citing even older research. I just read an article where a researcher, who published a filtering list for use in drug screening, was contacted by pharmaceutical companies 48 hours after his research was published in a journal.

How did these companies find his research? Did they have a subscription to the journal or did they use some service like Google Reader that compiles a number of RSS feeds. Do they rely mostly on a single third party news service? How efficient are each of these methods. I use Google Reader to view feeds from sites like physorg.com and sciencedaily.com. Physorg.com says they have 7 full time staff and 6 contributing authors and produce about 100 news articles a day about the latest research. Many times these articles are comprised mostly of the press release of a research center. These sites also have no specialty, they cover any science and technology related research. 13 full time people avails 520 hours a 40hr week. How many scholarly articles are published a week?

Barriers to diffusion

When I survey journal articles I often don't read past the title. I look for things that seem to be significant to my inexpert eye. If it seems significant I'll read the abstract and sometimes stop there because it doesn't meet the significance threshold (and often because I don't want to pay $30). Sometimes it appears the most significant idea in a published article is the research procedures used and developed as opposed to the specific focus of the research. I wonder how often procedural developments aren't appropriately highlighted in the title, abstract, or entire article to reflect their significance and what this costs in terms of the diffusion of such best practices and advancing the field. Furthermore the perusing of on-topic research articles or the use of on-topic keyword searches could omit relevant developments in other fields. What is the cost of this?

Language is another barrier to the diffusion of information. According to "Science and Engineering Indicators:2010" by NSF the US comprises 33% of total global spending on R&D with Japan and China in second and third place. Citing the OECD the same report also shows china, the US, and the EU as all having about the same number of researchers. A UK government report indicated that China moved into second place in total number of published papers, displacing the UK and raising their percent of total papers published to something around 11 percent however these stats aren't indicative of the number of papers published in non-english languages which I wasn't able to find in my brief attempt (highlighting a failure in information diffusion as it seems likely such information exists..I just can't find it). These stats do however indicate that the potential for language barriers in best practice diffusion is worth evaluating

Additionally of course the myth of free information that the internet let's us indulge in must be put in it's place. The internet reduces the cost of information diffusion to near zero but not information necessarily information creation. All the arguments for copyrights and patents apply. Thus funding and motivating research by charging for access to the results is a reasonable model. i think it best to consider this barrier to information diffusion a separate question and ponder simply the diffusion of information that is made freely available by it's authors

Diffusion mechanisms

Of course any thought on information propagation can't omit consideration of search engines in general. Suffice it to say it seems were are in the infancy of search engine development. That term presumes a natural development with time but i suppose it's possible, that to go much beyond keyword frequency and other present algorithms, it could essentially require the development of human level AI..so we may be in for a wait.

What do you think? This is another top of the head rambling so correct and guide.