Back to the Top
Dear all,
I am learning linear mixed effect model by myself recently. I am a beginner in this area. My
textbook is the well-known "Pharmacokinetic-pharmacodynamic modeling and simulation 1st Ed" by Peter
L Bonate.There is an example of modeling tumor growth in the "Charpter 6: Linear mixed effect
models" in that book (Page 197). Because the book did not tell how to set up the model in details,
it seems I have a lot of questions regarding this example.
1. I do not have the PROC MIXED software as the author used. I only has the Phoenix 6.3 in hand. I
know Phoenix has a module "linear mixed effects". I wonder whether Phoenix can fully do this work as
the book showed.
2. for this tumor growth modeling example, how to correctly write the "model specification" in the
"fixed effect" tab? In the setup, I mapped "group", "subject" and "day" as the "classfication", TV
(tumor volume) as the "Dependent" and "BW (body weight)" as the regressor (I guess BW is a
covariate).
3. Which one should be the "random effect" in the "variance structure" tab?
4. how to get the model predicted TV values in Phoenix? I saw the plots of observed time-dependent
TV values against model predicted values of each subjects in the book (Page 202, Figure 6.10). But
do not know how to do it in Phoenix.
Many thanks in advance.
Gary
Back to the Top
Dear Gary,
I would suggest you also ask this question on the Phoenix user forum; www.pharsight.com/extranet
(you will need to register but this is free)
There are several existing threads around tumour growth models and some example projects
However I have to point out that it sounds like you're picking up the wrong operational object to
do what you are asking, you need to select "Phoenix Model" and then you will see the NLME engine and
all the different models, algorithms etc. available there. The LME option is a statistical tool that
you might use for e.g. BE.
I think working through some of the basic examples in the NLME examples guide
(C:\Program Files (x86)\Pharsight\Phoenix\docs) and looking at the forum would really help you
Best regards,
Simon.
--
Simon.Davis.aaa.certara.com
Back to the Top
Dear Gary,
Yes, Phoenix can run this linear mixed effects model.
2. for this tumor growth modeling example, how to correctly write the "model specification" in the
"fixed effect" tab? In the setup, I mapped "group", "subject" and "day" as the "classfication", TV
(tumor volume) as the "Dependent" and "BW (body weight)" as the regressor (I guess BW is a
covariate).
Answer:
# Once you have mapped the variables, you should see them appear under the fixed effects Tab. I
# assume you are treating "day" as categorical as well. Based on how you have mapped, you should be
# seeing, group, subject and day under classification window, bodyweight under regressor. Then now
# you can specify the fixed effects model, by dragging you model terms on to the "Model
# specification window" and specify the fixed effects model using the operator buttons (+, x etc..).
For ex:
DV = day+group+bodyweight+group x day.
# Then you have to specify the random effects.
Click on Variance structure, make sure Random1 tab is active. You would see your model variables
under regressors and classification tab as you saw in the fixed effects window. You can then
specify your random effects model, inside the random effects tab...
For ex: u1+u2 x day, if day is one of your random effects variable. (This is an example of random
intercept and random slope model). Make sure to check the random intercept tab.
Hope this helps...
But as Simon suggested, it would be good if you post the question on the Pharsight Extranet, where
we can attach screen shots or share the phoenix project.
Thanks
Mathangi
Mathangi Gopalakrishnan, MS, PhD
Research Assistant Professor |Center for Translational Medicine
Email: mgopalakrishnan.-a-.rx.umaryland.edu
http://www.pharmacy.umaryland.edu/centers/ctm/
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@lists.ucdenver.edu with "Linear mixed effect model for tumor growth_Phoenix application" as the subject |
Copyright 1995-2014 David W. A. Bourne (david@boomer.org)