Back to the Top
Dear Forum,
Is there any relationship between the intercept of the linear equation
and LLOQ area ratio values.
Thank you for sharing your experiences.
Regards
Back to the Top
The following message was posted to: PharmPK
Hi,
Interept gives an idea how high is your blank
backgroud, which usually has something to do with the
interference. By my understanding, the LLOQ signal to
the signal of intercept needs to be above 5.
Xiaodong
Back to the Top
The following message was posted to: PharmPK
Hello,
In my view, the intercept in a regression line derived from response
(peak are ratio) against concentration is related to the "noise" in the
method made up of the various contributing sources of uncertainty.
Mathematically, is the response at zero input, or "background" of the
system. Now, the type of statistical weight used during the regression
analysis will have an impact on the intercept and on the accuracy of the
calculated concentration at the LLOQ, so once can see a relationship
between the intercept and the LLOQ. I do not think there is anything
else beyond this.
Luis
Back to the Top
One must be cautious assigning meaning to a y-intercept by regression.
First, it may be observed but not statistically significant. It is
then ignored.
It may be statistically significant but not of any practical
significance because we do not know if the regression curve fits at x
= 0. The response vs. conc relationship in the intercept region may be
different from what is observed over the calibrators or standards.
If it is practically significant, it is probably not a correct reading
of the constant error term. That is gotten as the Total Youden Blank
which is an independent regression.
Regards,
Stan Alekman
Stanley L. Alekman PhD
S.L.Alekman Associates Inc.
Back to the Top
I am not sure if I have folowed the complete thread of this issue, but
if using a statistical algorithm with the calibration data, e.g. OLS,
you can get a (+) or (-) intercept without any "interference" actually
being present at all. This is by virtue of the fact that if you are
using OLS regression the LS criteria minimizes the SS of the
deviations between the data and the fitted points - This criterion is
actually a variance term in disguise, and thus the larger
concentrations have more "pull" in the process of getting the best fit
at the expense of the smaller values, therefore the apperarance of an
intercept term in the calibration. This of course is why various
weighting schemes are used, e.g. 1/X or 1/X**2, the latter probably
being the first to try as it more closely resembles the SS.
Cheers
BC
Back to the Top
The following message was posted to: PharmPK
The intercept is an extrapolated or theoretical value.
The LLOQ is a fixed point- based on concentration- it is not
extrapolated or theoretical. It is proven by having met performance
acceptance criteria.
The general relationship is intercept < LLOQ where the response is
directly proportional to the concentration and intercept > LLOQ where
the response is inversely proportional to the concentration.
--
Ed O'Connor, Ph.D.
Laboratory Director
Matrix BioAnalytical Laboratories
25 Science Park at Yale
New Haven, CT 06511
Web: www.matrixbioanalytical.com
Email: eoconnor.-at-.matrixbioanalytical.com
Want to post a follow-up message on this topic?
If this link does not work with your browser send a follow-up message to PharmPK@boomer.org with "Intercept and LLOQ" as the subject | Support PharmPK by using the |
Copyright 1995-2011 David W. A. Bourne (david@boomer.org)