282396_en-US

liliconcom

Passing-Bablok & weighted-Deming

Hello,

 

According to CLSI EP09-A3 various regression techniques are available to assess the linear relationship between 2 analytical methods. Deming (D) is implemented within JMP as Orthogonal Fit. What is missing are 2 methods which are more appropriate for most (if not all) chemical or immunological assays (oncology, virology, cardiovascular, inflammatory medical devices) i.e. weighted-Deming (wD) which take care of the heteroscedasticity of the measurement (that's always the case for immuno-assays) and Passing-Bablok (PB) which in addition is robust to outliers (that's often the case). The later (PB) being the best in class and most commonly used in the field.

For that reason Analyze-it (EXCEL add-in) is still used for regulatory submission (FDA, cFDA, CE marking) rather than JMP for linear regression assesment.

 

Emmanuel (former  Beckman-Coulter employee)

Statistics and AnalyticsRe: Passing-Bablok & weighted-Deming

Yes indeed, and I was using it at Beckman-Coulter. That being said it is not so practical and cannot be used for submission purpose unless validated which is not worth the effort (at least within Beckman-Coulter).

 

Re: Passing-Bablok & weighted-Deming

I know that Roche has build their own deming regression add-in as it was not available in JMP.

Re: Passing-Bablok & weighted-Deming

Hi Emmanuel,

Thanks for bringing the topic up again and I agree that it is time to get these methods into JMP.  I for one will add my vote!!!  I spoke with Russ Wolfinger because he had done some work with another customer getting the Method Comparison addin (https://community.jmp.com/t5/JMP-Add-Ins/Method-Comparison/ta-p/21520) working using Passing or Bablok.  Notice there has been a fair amount of discussion on the topic.  There are regulatory authorities using this addin as well.  I will discuss with others here at JMP that I know will want to see this, including maybe Russ Wolfinger, Don McCormack and Mark Bailey?

Re: Passing-Bablok & weighted-Deming - Status changed to: Open for Voting

I can report that there has been work in this area for JMP 17. Passing Bablock analysis is now available in Fit Y by X, as an option for Bivariate regression analysis. JMP 17 will be available externally in fall of 2022.