Correlation Ranges

10 followers
0 Likes

How does your facility establish an acceptable range of correlation for glucometers. From analyzer to analyzer, and lab to analyzer. We have Roche Accuchecks. We have been going based off the 12.5mg/dL or +/-12.5% whichever is higher from our CAP surveys. Just wanting to see if this was accurate. 

11 Replies

I use the DI total allowable error table. They share the requirements from various sources. For glucose, we allow 6mg/dL or 10%.

https://datainnovations.com/allowable-total-error-table

We use the same as Tianna Stein, 
Total Allowable Error Table - Data Innovations    

Attachment.
image.png

Thank you, that is a great resource. It does list a whole blood and serum option though. 
Attachment.
image.png
Attachment.
image.png

Yes, ultimately, it is a conversation between you, your admin director, and your Medical Director, on which value you want to use. When CLIA is listed as a source, I generally stear towards using their standards/requirements.

Thank you!

Since glucose testing is waived, are correlations required. I am trying to no longer perform them due the time it takes me

Hi Tamara - they are not required if you are not using them off-label. As far as I know, the Nova glucometers are the only ones that have FDA approval for use in critical care settings. If you use a different meter in critical care areas, the testing is considered off-label and jumps to a modified laboratory test and are then subject to non-waived testing equipment requirements such as linearity and correlations. We use the Roche meters currently, so for our ICU areas (SICU, PICU, CVICU, NICU), the meters are treated as non-waived with moderate complexity operator requirements.

We use nova glucometers and was aware of that.  Was just doing for good laboratory practice but it's become time consuming.  Thank u

It really is good lab practice. My plan for 2024 is to use the CAP Quality Cross Check products. Twice per year you get 30 sets of 5 levels of controls. Serves as instrument comparison and alternative proficiency, can compare up to 30 instruments.

We use the quality cross checks, but I do not do lab correlations. 

We use Nova Statstrip.  Our criteria is If Gluc <75 15% or 15 mg/dL; if Gluc >/- 75 then 20% 
Set by our medical director.

Reply
Subgroup Membership is required to post Replies
Join POCT Listserv now
Jennifer Toncray
almost 2 years ago
11
Replies
0
Likes
10
Followers
834
Views
Liked By:
Suggested Posts
TopicRepliesLikesViewsParticipantsLast Reply
HMS Hepline kit
Cassie Laranio
about 14 hours ago
1052
Miranda Cotter
about 11 hours ago
Critical Documentation on the GEM 5000
Samantha Draper
about 17 hours ago
80136
Miranda Cotter
about 11 hours ago
Method Correlation Total Allowable Error vs Slope and Correlation Coefficient
Jo Win
about 18 hours ago
10166
Carl Bayaca
about 13 hours ago