Chromatography Forum

Is it true that Empower comes validated straight away? Whereas Excel you have to set up validated spreadsheets and test then yourself?

Empower also releases each Build and goes into detail about exactly what was changes and fixes etc. Microsoft don't seem to do this.

I was told by a Waters employee that Empower and Waters have the whole industry pretty much sewn up, and when I think of it, that's pretty accurate. What exactly is Waters competition? There might be a few companies that offer HPLC software but in the main, Waters rule. Thoughts?

The "big three" would be Waters, Agilent, & Thermo. I would contact Agilent & Thermo to get their response as well as Waters.

The instrument companies make their source code available to the accrediting authorities to examine, so that they can sell instruments. Microsoft doesn't make it's money from analytical labs so it has nothing to gain from getting its software accredited.

Peter

Agreeing w/ Peter - to the extent you document and lock down Excel spreadsheets and comply with your internal SOPs, you can "validate" them. Absent source code, however, you can never be truly certain that you will pass muster with some auditors, especially if you allow MS upgrades to Office without further testing.

That said, CFs written in Empower and similar programs also need to be tested as it is as possible to create junk formulas in them as in Excel.

The instrument companies make their source code available to the accrediting authorities to examine, so that they can sell instruments. Microsoft doesn't make it's money from analytical labs so it has nothing to gain from getting its software accredited.

Peter

They don't need and ISO 17025 accredited labs don't need either:
Note from ISO 17025:
NOTE Commercial off-the-shelf software (e.g. wordprocessing, database and statistical programmes) in general use within their designed application range may be considered to be sufficiently validated. However, laboratory software configuration/modifications should be validated as in 5.4.7.2 a).

The instrument companies make their source code available to the accrediting authorities to examine, so that they can sell instruments. Microsoft doesn't make it's money from analytical labs so it has nothing to gain from getting its software accredited.

Peter

They don't need and ISO 17025 accredited labs don't need either:
Note from ISO 17025:
NOTE Commercial off-the-shelf software (e.g. wordprocessing, database and statistical programmes) in general use within their designed application range may be considered to be sufficiently validated. However, laboratory software configuration/modifications should be validated as in 5.4.7.2 a).

Catch an auditor on a bad day and you will get written up for using office software for a laboratory application i.e. outside it's designed application range

Peter

Hi EmpowersBane,

Waters chief competition? That'd be Thermo Dionex' Chromeleon. Waters has been working on Unifyso long now...Empower 3 FR2 seems to me to be the end of the road. I do not work for Waters (as you know) just my opinion.

I think that IQ/OQ mostly satisfies the validation requirements for CDS ware, Custom Fields (I'm still behind the 8-ball with these, sadly) may need to be validated per case (not hard), just to ensure the math is correct.

Excel sheets? Must validate every time...folks addressed this well above.

Agreed, Chromeleon has been nibbling at Waters' lunch for some time now. Waters is just not very agile when it comes to their development cycles. Their advantage still has a fair bit to do with their hardware market share and the lag time involved with CDS "A" getting working drivers for company "B's" hardware. This goes both ways, but Waters and Agilent both have a pretty significant hardware advantage over the competition with the possible exception of Thermo/Dionex stuff which also works well. If you want the latest kit with good drivers, you get your CDS and hardware from the same place, or you wait a year (or more if you're trying to run Thermo/Dionex stuff on Empower).

Any custom fields we create in work have to be tested as per SOP. But the SOP is incredibly vague and along the lines of "ensure it works as required"!!

I normally test the CF on at least 3 peaks to ensure I get the right result and I use a calculator to test it against.

We have Excel sheets in work but for most of them, they are not required. Something in Empower or the report method could do the work and once a cf is validated it doesn't need to be questioned or checked after that. It will work whereas Excel is prone to user error with entering in figures etc.

The report forms exported from Agilent GC-FID and GC-MS systems all use a very early version of Excel.

MS Excel 2007 before SR1 had division by 2 error in range 65536 and greater.
Somebody in MS decided to optimize division by binary shifting and made off-by-one error.

I'm using Excel 2003 to generate my reports by macro refreshed links with multiple spreadsheets (instrument data file, COC table, analysis request table, client info table). I link the data file and Ctrl-X links in lookups from the other table files to generate my reports. It beats the heck out of sneaker net to the front office and transcribing data into reports. It may be whistling past the graveyard. I can't afford a commercial data handling package for my single chemist lab. I'll have to look into whether there are errata in the 2003 release.