Documenting Manual Integrations

[Room 101]

In doing a lot more government work, we are now required to document ALL manual integrations. This requires a "before" and "after" snapshot of the integrated peak, a reason for the manual integration, date, time and analyst initials.

Any advice on how to handle this documentation requirement? Or how to better minimize the number of manual integrations.

We are using Chemstation and Star software.

Running standards is no problem, but when an actual sample is analyzed, "dirty" matrices can cause all types of integration issues.

Thanks!

[Peter Apps]

I know that it is a counsel of perfection, but the best way to minimise manual integration is to improve the chromatography so that peaks are reproducibly separated well enough to let the integrator find them reproducibly.

If the chromatography is as good as it should be, then you need to look carfully at integration parameters, and setting them to appropriate values. NB these do not have to remain the same thoughout the run time.

Peter

[Room 101]

I agree.

We are running close to 100 compounds in 20 minutes. Trying to get integration parameters to work for every compound at every possible level is a toughy. Factor in the huge non-target analyte inteferences makes it more complicated.

Thanks!

[JGK]

It's basically a safety system should anyone question a reintegrated result. In GLP bioanalytical work for example, if we wanted to reintegrate a peak we had to document a "valid reason" and have the departmental manager sign off on it before doing the reintegeration. CFR compliant software and audit trails on the projects kept copies of all chromatograms (deleting any from the system was not possible) and logged all actions by named users. Controlling reintegrations like this removes any chance of the analyst arbitrarily reworkng data that just "doesn't look right".

[Room 101]

I agree with that as well.

We are trying to find a fast and easy way to accomplish this while keeping up with tight production schedules.

Perhaps we need different software?

[tphase]

sounds like you do need different software. Check out gcwerks - http://www.gcwerks.com/. It was designed for atmospheric monitoring so it may not be quite what you need but it will read chemstation chromatograms and do a much better integration job.

Gerry

[WillyOne]

Can you post one chromatogram?

[varossf]

Hello room 101.
Try to call Agilent and upgrade your software to the OpenLAB CDS EZChrom Edition (NOT the ChemStation one).
In the EZChrom software you work project oriented. Per project you can set the audit trail, together with how to sign etc. Furthermore you can set the project to store all result versions, creating a version control on the results and allowing to go back to previous integrations. Audit trail the logs who has done what when and why.
This gives you the possibility to electronically create documentation on what was done with the results. This would speed up the process without being hampered by all kind of external logging. In addition you can use the electronic signatures to approve/reject the results.

Note: this is only for non-MS instruments.

Regards,

[Room 101]

Thanks varossf. Unfortunatley I need for MS.

Does anyone know if the audit trails do before/after snapshots of the integrated peak? This is my main issue. Our clients need a "picture" of the peak before integration and a "picture" after integration. Simply flagging the data with an "m" in not sufficient.

Right now, we are generating these snapshots to PDF Factory.

[carl.nott]

I've been working on this, a bit, for electronic CLP package generation. At the moment the system involves setting up the default printer on the chemstation computer to be a .pdf printer (one that works with chemstation, bioPDF does (with ours) but not all do) and set up the method to automatically print a detailed report after each run. bioPDF allows you to set the save name parameters and can be set up to just dump the file, so when the analyst runs all of the data gets a detailed .pdf created. Then after data review the analyst prints another detailed .pdf that reflects the changes. We've done this for a few CLP-like packages that they just wanted .pdfs of (no hard copy).

Eventually I want to pull the 'm' flag (denoting a manual integration) for each analyte when the data is parsed into our LIMS system and then used those m flags to sort, clip, and compile a .pdf for each data file that has the before/after along with the summary for the data (veryPDF Split-Merge should work via command line interface). But since we don't do very many CLP jobs that project is on the back burner.