I've added my own version of %left and %trim and test packs for them. This is so that people can maybe ditch the macros supplied by the SAS Institute if they don't use them. %left and %trim might well be used even though %sysfunc calls make them redundant. I don't know if people use any of those supplied sasautos other than %left and %trim.

"Roland" <roland@rashleigh-berry.fsnet.co.uk> wrote in message news:banijt$5s6$1@news7.svr.pol.co.uk... > Hopefully, most of you who work in the field of Clinical Trials reporting > with SAS will know I have written a large number of macros that support that > field of work. They live here: > http://www.datasavantconsulting.com/roland/ > However, they have disclaimers on. Although I am confident that they are > fully correct and have tested them and they work to my own satisfaction, it > obviously in no way "validates" them. For a start-off it is improper QC > because I am both the author and the tester. Hence the disclaimer. But it > seems a shame to waste the test packs. I spent far more time, on average, > testing the macros rather than writing them. There is no doubt in my mind > that they all work correctly. But they can only be validated by the users > and so I am going to make my test packs available. It is a start only. The > user can alter what is in the test pack to test the macros. Ultimately it is > their responsibility and only they can "validate" them. The owner of my web > space, Gregg Snell, has kindly given me ftp access to that area so that I > can do the updates myself. I haven't decided the best way to lay it all out > but I think I'll put links to the test packs alongside links to the macros > themselves. I might have a blitz on them on Monday since it is a bank > holiday over here in the UK. > > I feel these macros might have a role to play in the future of Clinical > reporting. I think soon that the FDA rules for CFR Part 11 will finally > become clear in people's minds. Maybe I am just old-fashioned but I feel > that education wasn't what it was 30 years ago or more.It seems to me that > there are a lot of young people in this field with very good paper > qualifications but they lack the ability to read and understand English and > lack the ability to approach and solve problems. Also the use of English has > become too fluid so that words have lost their meaning. What is a "datum", > for example? If you calculate a mean then is that a "datum" as well? If you > produce a bar chart in SAS then is that an "electronic record"? I think that > what has happened in the industry is that the FDA rules for the handling of > data has been incorrectly assumed to apply to the statistical analysis of > data. However, the safety reporting of data might be done with no QC checks > because it will be assumed that the output is correct. If that is the case > then the pharmas should be obliged to fully validate that system and to > control it in such a way that errors could not enter the system. There > shouldn't be anyone diving in and fixing the macro code because they spotted > something that looked wrong. It should be a highly regulated environment > with a full audit trail and thorough testing and documentation of that > testing with archived versions of software. It should have restricted > access. But this will only be a subset of the analysis. I think the rest of > it will be assumed not to fall under CFR Part 11 and things will go on as > normal. You have only got to stop and think logically. The FDA make extra > requests for analysis and expect and extremely short turnaround. This would > not be possible in a highly regulated environment. So I think CFR Part 11 > will have its scope stretched outside of Data Management activities to cover > only those computer systems that cover safety reporting and then only if > that system is assumed to be correct by the users and the output therefore > not QC'ed. And I have to wonder whether this would be cost effective and > that it might be better to QC the output in any case. > > What I hope to provide with my macros is a set of tools that are robust and > realiable that can be easily validated by the users and that the set covers > the usual needs of safety reporting. If they could become standard across > the industry then not only would I feel rather pleased with myself but it > would mean that other programmers could reach for them when they need them > when they move from job to job and see them there in identical form and > knowing the functionality it could give them. And they are all public > domain, BTW, so they are free. > > I'll aim to have these test packs in place by the end of Monday. > > Roland > -- > "Today's mighty oak is just yesterday's nut, > that held its ground." - David Icke > >