Method Validation Matters

Posted by Sten Westgard, MS

https://www.evaluate.com/vantage/articles/news/snippets/healthineers-sinks-atellica-letdown

I'm late to the party on this story, but I think this illustrates an atrophy in our core laboratory skills. When a multibillion-dollar diagnostic manufacturer loses profitability due to the "longer than expected installations times" of its instruments, that's a telling sign we are losing our ability to validate and verify our methods.

For too long, labs have been out-sourcing these skills to the manufacturers, making that validation and verification phase part of the purchase of a new instrument. Or if they're not out-sourcing it, they are not training enough staff internally to handle the big instrument validation projects - where hundreds of tests need to come online in a very short period of time.

Our neglect of statistical training means we don't have enough skilled staff, either at the manufacturer or in laboratories, to complete the necessary validation / verification studies. Then it takes longer to get these instruments up and running. Certainly bad for profit, but possibly even worse for patients.

[An even worse scenario would be if the instruments are actually failing validation or verification... but we don't know that is happening here.]

It was pointed out that there isn't a pure causal relationship established in this article - that I'm being speculative about the reason installation is taking longer than expected.

There are other possibilities that could delay profitability, but most of them are less flattering to the manufacturer:

• The complexities of installing systems these days are much higher. But if that's true, isn't it foreseeable? Isn't it therefore possible for the company to either hire more staff for installation, or warn analysts that profitability will be delayed while installation is performed? If the company didn't know installation was going to be so complex, that's a failure of execution and implementation. If the company knew but pretended it wasn't going to happen, that's a failure of leadership and financial disclosure.
• A further follow-up, if the company is having problems installing 340-odd systems, what's going to happen when they scale up to the 1500+ systems that they claim they will achieve this year?

My speculation places the blame closer to the lab, for lacking the skills to be able to efficiently validate new instruments. That's something that's out of the control of the manufacturers, and less predictable. I'm trying to give the manufacturer more benefit of the doubt here. But if you prefer, we can go dark - and blame the manufacturer for all the problems.