Wednesday, April 13, 2005

The argument against testing

Did you ever think you'd read about a validation/quality assurance person talking about talking people out of testing? Yeah, me either. But that's just what I've been doing lately.

My current employer is a relatively small and young company; as such, in the typical young company zeal to do the right thing, they have long held the attitude that everything computer-related must be validated in a full-blown validation effort: system master plan, requirements spec, test protocol(s), test scripts, traceability, and summary and conclusions and master reports aplenty. And I mean everything; we have a Microsoft Word template that we create as part of a service to our clients. All it does is apply font sizes and formats to existing content, and yet for several releases now, it's been validated in a full-scale validation effort.

I'm one of the strongest proponents of performing a strong testing effort, and doing it right the first time, but there's just a lot of needless effort going on here. That Word template, for instance. We develop it as part of a service to our clients, so I definitely believe that it needs to be tested to some extent. But does it really need to be validated, with all of the bells and whistles that validating something involves? I don't think so. My perspective is that if we qualify Microsoft Word as a part of our operating environment (and we do,) would we need to validate every document created by Word? No. So why would we need to validate a template that does nothing more than apply formatting to existing content (or format a segment of a document in preparation for the addition of content)? My feeling is that, frankly, we don't need to, and my only regret is that I didn't catch this sooner and make it easier on everyone nine months ago when I first got a look at this stuff.

21 CFR 820.70(i) states that "When computers or automated data processing systems are used as part of production or the quality system, the manufacturer shall validate computer software for its intended use according to an established protocol. All software changes shall be validated before approval and issuance. These validation activities and results shall be documented." In the particular case of the Word template I'm talking about, this template is neither part of pharmaceutical product production or of the quality system. This template is about as far removed from drug production as it gets; it's just a template for authoring reports. The template itself has zero impact on the content. That having been said, all of this is predicated on Word being a qualified part of the operating environment (and, again, it is here,) but once Word is verified, you really don't need to do more provided you're not using mechanisms within Word itself to derive, modify, or delete information that has regulatory impact.

So far, the person I've tried to talk out of validating his Word template has come back to me once to debate the decision. I'm open to the debate, but it does make me chuckle a little to think that, for once, I'm on the side that's arguing against fully documented testing.