"Greg Woolridge" <greg.woolridge@TAP.COM> wrote in message news:OF6477FF5C.887C2297-ON86256D2F.00652D01@production.tappharma.com... > Roland wrote (in part): > > > This is a problem that I struggle with almost daily. If you take very > > literal translation of the definition, then because our programs > > "...create,modify..." data in one way or another, then yes they are > subject > > to Part 11 and should be validated. > > Ah, but they don't "create, modify" data. The programs accept the data as > it > is. Where I have worked then it is almost unheard of for a program to > change > any input data. Even if you know it is wrong. Data Management have to make > the change going through their own procedures and have all the > documentation > or electronic timestamps to back it up and then give us fresh data. And of > course you do not create "data" at any stage as per definition "thing known > or granted". At least I have never done so. You don't invent subjects and > give them measures. > > > I agree with your point, but the QA person I have to deal with would argue > that if you calcuate any statistics on the data, such as mean, max, min, > you have "created" data. And if you use a format to change the appearance > in the report, you have "modified" it. He is a pain in the "you know > where".

A statistic is an inference, surely. It is not a thing observed.

> Greg M. Woolridge > Manager, Study Programming > TAP Pharmaceutical Products Inc. > e-mail: greg.woolridge@tap.com > phone: 847-582-2332 > fax: 847-582-2403