Is R overtaking SAS at universities and colleges?
in [SAS]
|
From: rjf2 on 9 Jul 2008 12:10 tangential comment: My division is coming up to speed on understanding how to work with grantees and other non-fed employees on managing projects using open-source software. Ronald J. Fehd, IT Spec team lead: Statistical Software and Knowledge Worker Support
From: NordlDJ on 9 Jul 2008 12:57 > -----Original Message----- > From: SAS(r) Discussion [mailto:SAS-L(a)LISTSERV.UGA.EDU] On > Behalf Of Mary > Sent: Wednesday, July 09, 2008 8:55 AM > To: SAS-L(a)LISTSERV.UGA.EDU > Subject: Re: Is R overtaking SAS at universities and colleges? > > Stuart (I can't help noticing the "NIH" in your name), > > Perhaps you could ask at the NIH for us! It would seem that > acceptance would come at the NIH before the FDA, since many > NIH grants are to universities (I'm on an NIH grant), whereas > the FDA more deals with pharmaceutical firms. > > Certainly, if people on NIH grants start publishing papers > that have statistical analysis done in R, then that would > "open the flood gates", as it were to more people using R to > do their analysis. I would think that the FDA would be more > conservative in requiring the use of SAS for longer than the > NIH would. Many statistical analysts just coming out of > college may get their first jobs on NIH grants at > universities, and would naturally use R if that is what > they've been trained in unless there's an NIH requirement > prohibiting it; even if there is, they may still use R for > exploratory work and switch over to SAS for formal work. > > Thus, Stuart, do you know anything about the NIH's feelings > towards using R for statistical analysis? > > -Mary There have been a number of discussions on the R-help list. If you Google "R-help fda 21.CFR.11" you will find all sorts of information. Here is just one of many URLs. There is no problem with using R. http://134.148.236.121/R/help/04/12/9405.html Hope this is helpful, Dan Daniel J. Nordlund Washington State Department of Social and Health Services Planning, Performance, and Accountability Research and Data Analysis Division Olympia, WA 98504-5204
From: Marc Schwartz on 9 Jul 2008 13:55 long3(a)NIEHS.NIH.GOV ("Long, Stuart (NIH/NIEHS) [C]") writes: > What does the FDA (or any federal department that hands out grants) > think of R? That might be an important question to consider if R is > currently a threat to replace SAS. The primary analyses for the Avandia safety review performed internally by the FDA were done using R. There is a demonstrable increase in the use of R by the FDA internally. Is R a "threat" to SAS? Not by intent, but I would suggest that the increasing frequency of R related discussions in this forum and others is an indication that the normally glacial marketplace is taking increasing notice. Is this going to be an overnight phenomenon? No, but we are slowly coming upon the "Crossing of the Chasm". Mary made note in this thread of the academic facilities using and contributing financially to R via the R Foundation. One might want to take note of the pharma companies and clinical research institutions also listed on that same page: http://www.r-project.org/foundation/memberlist.html If you would like an admittedly biased perspective, you might want to review the presentations at last year's useR! meeting at Iowa State University. There was a panel session on the use of R in this domain. Go here: http://user2007.org/program/ and scroll down about half way or so to the entry entitled: "Panel Session: The Use of R in Clinical Trials and Industry-Sponsored Medical Research" where you can download copies of the presentations made, including one by an FDA statistical reviewer, who made verbal note during his presentation of the use of R for the Avandia safety review. Finally, for those who may have been at JSM 2006, Sue Bell et al from the FDA offered this presentation, entitled: Times “R” A Changing: FDA Perspectives on Use of “Open Source" http://www.fda.gov/Cder/Offices/Biostatistics/Bell.pdf As Tony Rossini and some others have noted in this forum previously, there is a document on the R web site at: http://www.r-project.org/doc/R-FDA.pdf which covers many of the regulatory framework issues that are germane here and there is an important pending update to that document which we are hoping to have posted at that link very soon. Regards, Marc Schwartz
From: Stephen McDaniel on 9 Jul 2008 13:52 Officially, FDA doesn't endorse any product like SAS. They just put out the quality assurance regs and put the official onus on the pharma's to comply. SAS is simply accepted as a de-facto stable platform but I see no reason why they wouldn't accept R submissions if fully documented and QA'd just like a SAS submission. That said, the switching costs for most pharma's are enormous and would take years. When you look at big and even medium sized pharma's, the cost of SAS is a hiccup in their budget. Smaller biotechs might find R appealing, but will likely have a problem getting the staff to do not just the stats in R, but all the reporting and analyses to support the stats R is touted for being so good at... My two cents... Stephen On Wed, Jul 9, 2008 at 9:57 AM, Nordlund, Dan (DSHS/RDA) < NordlDJ(a)dshs.wa.gov> wrote: > > -----Original Message----- > > From: SAS(r) Discussion [mailto:SAS-L(a)LISTSERV.UGA.EDU] On > > Behalf Of Mary > > Sent: Wednesday, July 09, 2008 8:55 AM > > To: SAS-L(a)LISTSERV.UGA.EDU > > Subject: Re: Is R overtaking SAS at universities and colleges? > > > > Stuart (I can't help noticing the "NIH" in your name), > > > > Perhaps you could ask at the NIH for us! It would seem that > > acceptance would come at the NIH before the FDA, since many > > NIH grants are to universities (I'm on an NIH grant), whereas > > the FDA more deals with pharmaceutical firms. > > > > Certainly, if people on NIH grants start publishing papers > > that have statistical analysis done in R, then that would > > "open the flood gates", as it were to more people using R to > > do their analysis. I would think that the FDA would be more > > conservative in requiring the use of SAS for longer than the > > NIH would. Many statistical analysts just coming out of > > college may get their first jobs on NIH grants at > > universities, and would naturally use R if that is what > > they've been trained in unless there's an NIH requirement > > prohibiting it; even if there is, they may still use R for > > exploratory work and switch over to SAS for formal work. > > > > Thus, Stuart, do you know anything about the NIH's feelings > > towards using R for statistical analysis? > > > > -Mary > > There have been a number of discussions on the R-help list. If you Google > "R-help fda 21.CFR.11" you will find all sorts of information. Here is just > one of many URLs. There is no problem with using R. > > http://134.148.236.121/R/help/04/12/9405.html > > Hope this is helpful, > > Dan > > Daniel J. Nordlund > Washington State Department of Social and Health Services > Planning, Performance, and Accountability > Research and Data Analysis Division > Olympia, WA 98504-5204 >
From: Mikko J Virtanen on 9 Jul 2008 16:03
Angel <norcalangel(a)gmail.com> writes: > I tend to restructure/rearrange/organize only the output i > want to display in SAS and then ODS it and put the finishing touches > on it personally using Excel. You should definitely try to outlearn using Excel for "puttin the finishing touches on it personally". After you graduate, you'll end up in sutuations, where you need the finishing touches but do not have to do them personally. Only way to avoid this is to learn to do them with ODS. You could equally learn to do them with Sweave or something else, as long as you think of the process in terms of BI. This will make you more productive, and as an important side effect, it will make your research reproducible. Where I work, we consider "knowing Excel" a big minus for a (bio)statistician. Knowing ODS is a huge plus even when the project does not use SAS. MJ; -- ..signature necesse est |