Blood Gas Analyzer/GEM Validation Precision CV values and Internal vs External QC

19 followers
0 Likes

I am trying to validate the new GEM 7000 device but this applies to the GEM 5000 devices as well.

Our Werfen rep ran external QCs for a precision study with external controls and had an observed CV of 14.3%.

Ideally shouldn't the CV be <5% ?
Should it be based on manufacturer's insert?

What is everyone's policy?

Based on the manufacturer's precision study, didn't they also fail?

Calling Werfen technical support is not very helpful so I was hoping everyone here with our collective minds can solve this issue.

Thank you all!

From manufacturer IFU
Attachment.
image.png

FROM EP EVALUATOR - what we got

Attachment.
image.png


FROM COLA
Attachment.
image.png

6 Replies

It looks like you have all of the information you need at your fingertips. CV% can vary based on what your mean is and what 1SD is. Some institutions use a set acceptable %CV, like 5% or so, unless stated by the manufacturer (and FDA approved) at higher. You should meet the manufacturer's claim for Precision, at minimum. 14% for CV does NOT meet the claim of 6%. It appears that you have two data points that look suspicious when compared to the other data points. When I think about what might increase pO2 so wildly from the other data of the like, temperature of the vial and how long the vial was open comes to mind. Maybe the vial was held in the person's hand for a period of time, not well mixed or was open too long.. these would all lead to higher-than-expected pO2 values. Can you exclude any of them scientifically? What does your Precision look like over time, say over 2 to 5 days?

That is exactly what I would do too, repeat the study. I always keep in mind that it is not the responsibility of the manufacturer to ensure we are compliant with our policies and accreditation standards. It is ours. 10 samples for total Precision, within day/run and over time, is VERY small statistically speaking. I would do more testing and get your "N", used to calculate your mean, higher. My usual goal is to see 5 samples ran over 5 days, with 2 levels.

Thank you Erika, you are correct. It is just frustrating that the vendor doesn't raise the 14 percent CV as an issue to me, I had to read through the ep evaluator to bring it up to them. 

Have you reviewed the QC package insert? If 14% is within the CV specified in the QC package insert, you should be OK. However, you should also revise your validation/verification policy to accept the manufacturer's QC package insert CV, so that inspectors won't cite you in the future.  

The actual SD is shown as 1.9 with the specification: 5 SD (or CV%). Because the mean PO2 is very low, CV% will be very high even with an acceptable SD value. 1.9 mmHg PO2 is clinically irrelevant at a PO2 of 31.
Of course, the measurement of PO2 value in aqueous control solution is highly variable, influenced by sample size, humidity or residual rinse solution with sample path (with PO2 close to ambient, generally) as well as other factors such as equilibration temperature mentioned by others.

@Randy Byrd Good eye. You are right. The actual lab validation was 5.2 which is technically out from the 5 SD specification. I think I still would need to redo the precision. Thank you. 

Reply
Subgroup Membership is required to post Replies
Join POCT Listserv now
Jo Win
1 day ago
6
Replies
0
Likes
19
Followers
284
Views
Liked By:
Suggested Posts
TopicRepliesLikesViewsParticipantsLast Reply
Verifying Reportable Range using Controls and Calibrators
Jo Win
about 5 hours ago
2093
Jo Win
about 4 hours ago
Waived Testing Annual Recertification
Jamey Irvine
about 9 hours ago
70204
Peggy Mann
about 3 hours ago
Blood Gas Analyzer/GEM Validation Precision CV values and Internal vs External QC
Jo Win
1 day ago
60285
Jo Win
about 11 hours ago