Raw Data - a definition?

[mk65]

Does anyone have a definition for raw data? Ideally to cover electronic data and paper trails.

From the definition what would be expected to be stored and backed up?

Regards

mk

[Mary Carson]

From 21CFR58:
(k) Raw data means any laboratory worksheets, records, memoranda,
notes, or exact copies thereof, that are the result of original observations and activities of a nonclinical laboratory study and are necessary for the reconstruction and evaluation of the report of that study. In the event that exact transcripts of raw data have been prepared (e.g., tapes which have been transcribed verbatim, dated, and verified accurate by signature), the exact copy or exact transcript may be substituted for the original source as raw data. Raw data may include photographs, microfilm or microfiche copies, computer printouts, magnetic media, including dictated observations, and recorded data from automated instruments.

[mk65]

Thanks!

How about for chromatographic data? Obviously the first data obtained electronically is the raw data but what about any subsequent re-integrations or reprocessing of that data? Does this then also class as raw data? We have a data sysem that tracks any changes made with a full audit trail before the reported version is printed for use, is the printed chromatogram also then classed as raw data?

MK

[Tim]

It all depends on what you are trying to comply with. If it's 21CFR Part 11, then you need the original raw data, which as you say is the original signal from the source, PLUS you need the meta data used to generate that data in the first place (i.e. the instrument and data acquisition parameters) and the meta data used to process that raw data into the final result. You also need an audit trail of all the steps used in getting from one to the other, i.e. versions from each reprocessing of the data which gives a record of the decision making process you went through to get your final result.

You should also consider what is used to determine if whatever you are analysing is the correct strength, etc. If you are using the paper copies for that then, Yes, they should also be counted as raw data.

However, if you can get all the information given on the printout on a screen, I would get away from printing all together - saves a few trees in the process too! Ask yourself why you spent all the time and effort putting an Electronic Records system in only to have to print a copy of every chromatogram which then needs to be filed away (the excuses "it's in the procedure to print it out" or "we like paper copies" or "we have it in case the computer dies" are not viable )

[jbennett]

On a related topic......

My laboratory (an environmental lab) currently is totally paper based, but we are in the process of getting a new LIMS that will (hopefully) enable us to reinvent our processes. I want to scan all the paper that we generate in the course of analyzing samples, and get rid of the paper. I was wondering if anyone had experience with regulators accepting scans as "original" data or with lawyers in terms of defensibility.

Thanks,

Jack

[Tim]

I don't believe they would accept that as an original, since the signature/s on the page would not be original and you could have used whiteout to obliterate an erroneous result, write in the new one and on the scanned copy there would be no evidence you had done that (when GxP would require the entry to be crossed through so still legible, initialled/dated and a reason added if necessary). Even a colour scan may not show that up.

When faced with an auditor requesting to see a notebook, for example, we would generally copy the relevant pages at the time of the audit, stamp them up as "uncontrolled" and pass that through. That way they only get the information they requested and are not tempted to flick through the book to find other potential issues. Of course, should they then ask for the original we can give that to them.

If space is a concern, look to using deed boxes that can be locked and stored by an audited external company, with just recent documents kept on site. With the right company you should be able to get the documents within 24 hours at the most, though in an audit situation plan ahead and get the documents you think they might ask for out before they arrive, e.g. validation documentation.

What you could also consider is reinventing your paperwork to minimise the quantity you generate when testing each sample - only document critical parameters and use common notebooks to document reagent preparation that can be used in testing multiple batches then just have a reference on the worksheet.

Final consideration is ditch the paperwork all together and use something like electronic notebooks or record everything directly in LIMS with connection to lab equipment to permit electronic transfer of weights, chromatographic results, etc. If you need to record batch numbers from reagents, etc., look at barcodes to reduce risk of error in data entry.

Of course this all puts up the cost for implementation and validation in the first place, but could be a saving compared to errors introduced by incorrect data transcription.

[jbennett]

I never considered that whiteout might not show up on a scan.......

Since 21CFR58 specifically indicates that microfilm or microfiche copies are to be considered raw data, I do not understand why a scan would not be considered equivalent to them. I realize that the environmental field is not covered under 21CFR, but the underlying concerns are the same in both areas (i.e. is this a valid representation of what actually occurred).

We are planning to eliminate as much paper as possible by doing many of the things that you suggest. However you cant eliminate everything....field chains of custody, signed copies of final reports, ect. It is those types of things that I want to be able to store electronically.

Thanks for your input.

Jack

[Tim]

For chain of custoday, I assume you are refering to taking the sample and how it progesses through the lab? A LIMS should be able to handle that - the person generating the sample can log it onto the LIMS with all relevant details that also indicates the testing required, it gets received into the lab, you can enter dates to indicate when testing started and then through to final approval on the LIMS.

With respect to final reportsand need for signatures, if you have a LIMS that is validated for electronic signatures, you don't need a physically signed report. Everything can be done electronically up to and including the approval. Should you wish/need to generate a report (for example to send to a "customer" who does not have access to your LIMS), the fact that the sample was released electronically can be indicated on the report.

We use such a system for tracking samples through the lab and batch release of pharmaceutical bulk tablets and packed stocked.