Off topic: software validation [Power / Sample Size]
❝ From a QA perspective, "such things" (and I am just trying to speak broadly here) are often about qualification and/or validation. So, we now have another card on the table which is whether a function is documented or not.
I think the order (from the user’s perspective) is: documented ⇒ qualified ⇒ validated (and hopefully valid).
If we don’t have any documentation on how to use a SW’s function, how could we test it?
BTW, @Detlew: How did you discover the existence of Owen’s Q in SAS?
❝ I guess we need to find out what it implies to document, qualify and validate a function in some stats software.
A good starter is the PIC/S Guidance. But where do we end? The core routines of SAS and WinNonlin were written in the mid-1960s. I wouldn’t bet that all of them have been ‘touched’ ever since. Even fancy R uses compiled FORTRAN-libraries (of similar age?). Remember the noncentral t (based on a C translation of algo AS243 from 1989) mentioned in this post or – even worse – that one.
❝ Once we have that in place we can discuss if a script that uses this software needs to be validated, qualified (and please let's mess as much up as possible by introducing flavours of IQ, OQ, PQ) or just documented and how.
Yes, how much is possible and what are the efforts? Example: Phoenix’s validation suite costs more than the software itself. If you test all components (NCA, BE, IVIVC, PK, NLME/PopPK,…) estimated runtime is about 40 hours. ‘Validation’ should be done everytime to OS is altered (not only service packs). This would include Hotfixes/Patches which are distributed by M$ monthly. Your productivity would go down by 25%. In a larger company hotfixes are centrally distributed, which means that you can’t even retreat to another machine in the meantime. Anyway I would call everything you obtain from a SW vendor (how extensive it every might be) not validation but qualification. It’s up to us to find out the flaws.

❝ Anyone know of a regulator who knows the difference between a function and a script?
Cough.
(at least one)Dif-tor heh smusma 🖖🏼 Довге життя Україна!
![[image]](https://static.bebac.at/pics/Blue_and_yellow_ribbon_UA.png)
Helmut Schütz
![[image]](https://static.bebac.at/img/CC by.png)
The quality of responses received is directly proportional to the quality of the question asked. 🚮
Science Quotes
Complete thread:
- Sample size calculations: SAS vs PowerTOST/FARTSSIE Oiinkie 2011-12-06 14:30
- Sample size calculations: SAS vs PowerTOST/FARTSSIE Helmut 2011-12-06 14:49
- Sample size calculations: SAS vs PowerTOST/FARTSSIE Oiinkie 2011-12-07 15:57
- The power not to know vs PowerTOST/FARTSSIE/... d_labes 2011-12-12 15:38
- Paired designs not that uncommon Helmut 2011-12-12 17:05
- Paired designs not that uncommon ElMaestro 2011-12-13 17:04
- Off topic: software validationHelmut 2011-12-13 18:10
- Off topic: software validation ElMaestro 2011-12-13 20:30
- Off topic: software validation Helmut 2011-12-14 14:14
- Discovery of Owen's Q in SAS d_labes 2011-12-14 15:34
- Off topic: software validation ElMaestro 2011-12-13 20:30
- Off topic: software validationHelmut 2011-12-13 18:10
- Paired designs not that uncommon ElMaestro 2011-12-13 17:04
- Paired designs in PowerTOST 0.9-0 Helmut 2011-12-19 20:05
- Paired designs not that uncommon Helmut 2011-12-12 17:05
- The power not to know vs PowerTOST/FARTSSIE/... d_labes 2011-12-12 15:38
- Sample size calculations: SAS vs PowerTOST/FARTSSIE Oiinkie 2011-12-07 15:57
- Sample size calculations: SAS vs PowerTOST/FARTSSIE Helmut 2011-12-06 14:49
