Hi satimis,
Any version shall be fine. R can be installed on Windows/FreeBSD/Linux/Mac. You see many R running on Windows because many end users use Windows, but that does not mean Windows is the best OS to run R.
It does not really matter. R is a CLI based statistical environment and this is similar to SAS. Although R does have some additional GUI package that helps user get used to some function (e.g. R Commander, TinR, Jagger, etc), in the long run command line is the way to unleash its potential.
In R you will need to write/edit your code. Whereas there are many text editors around, I STRONGLY recommend Emacs (you can find in the ports: editors/emacs) + ESS ("Emacs Speaks Statistics") even though the learning curve of Emacs is somewhat steep. But once you get familiar with Emacs and ESS, your efficiency will be largely improved. (Besides, you can also use Emacs+ESS to write/edit SAS code).
This really depends on what exactly you are going to do. Further detailed info would help me give you some advice.
PubMed or Medline is usually the first choice. Whilst most database/journals need subscription, there exist some "Open Access" databases/journals that are open to the public to access with no cost (e.g. BMC series). Your medical company should have some info regarding the database/journals associated with your field. (But I wonder why you are looking for these articles since the list was just to show you that many research works done with R have been published in big-name pharmaceutical/medical journals.)
Since your company have also been using SAS, the following books might be of some interest to you:
- Muenchen, Robert A. (2008) R for SAS and SPSS Users. Springer; 1st ed. ISBN-10: 0387094172
- Kleinman, Ken and Horton, Nicholas J. (2009) SAS and R: Data Management, Statistical Analysis, and Graphics. Chapman & Hall; 1st ed. ISBN-10: 1420070576