Dear group,
I have a baseline as covariate in the model, and it is the same measurement as the modeled variable. To me, it is reasonable to believe that there is measurement error for this covariate. Is there a way to incorporate some kind of error term into this baseline covariate? Thanks a lot.
LI,HONG
Graduate student
University of Florida
School of Pharmacy
Department of Pharmaceutics
Office: P4-10
Phone: (352)273-7865
adding error term to covariate
Dear Li,
You are right in thinking that your baseline better not be treated as an
ordinary covariate (where we pretend that the covariate values are
measured without error). Unless you want to be very restrictive in how
to use the model (e.g. change from baseline in a study with similar
design and patients with same baseline distribution), it is best to
estimate the baseline. This also allows you to investigate relations (on
the individual level) between baseline and drug effect/disease
progression, etc.
I am sure you will find many old threads on this topic if you search the
nmusers archive. Additionally, here is a recent article:
Related Articles, LinksDansirikul C, Silber HE, Karlsson MO.Approaches
to handling pharmacodynamic baseline responses.J Pharmacokinet
Pharmacodyn. 2008 Jun;35(3):269-83. Epub 2008 Apr 30.
Cheers
Jakob
Quoted reply history
-----Original Message-----
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]
On Behalf Of LI,HONG
Sent: 29 July 2008 16:06
To: [email protected]
Subject: [NMusers] adding error term to covariate
Dear group,
I have a baseline as covariate in the model, and it is the same
measurement as the modeled variable. To me, it is reasonable to
believe that there is measurement error for this covariate. Is
there a way to incorporate some kind of error term into this
baseline covariate? Thanks a lot.
LI,HONG
Graduate student
University of Florida
School of Pharmacy
Department of Pharmaceutics
Office: P4-10
Phone: (352)273-7865
Hello,
I was bit by this problem once in the past. My quick fix was to put etas on
Smax & baseline (both parameters were estimated) and then put a block between
these etas. It has a few limitations a) A fixed effect is modeled as a random
effect; b) post-hoc plots of individual Smax and baseline were needed to
visualize how the baseline influenced Smax; c) This forces a linear correlation
when in fact the correlation can be curvilinear.
Despite these limitations, the quick fix worked very well and I could meet my
tight deadline. Can others point out other pros and cons of the above
methodology.
Thanks in advance for your feedback...Mahesh
Quoted reply history
-----Original Message-----
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] Behalf Of Ribbing, Jakob
Sent: Tuesday, July 29, 2008 4:52 PM
To: LI,HONG; [email protected]
Subject: RE: [NMusers] adding error term to covariate
Dear Li,
You are right in thinking that your baseline better not be treated as an
ordinary covariate (where we pretend that the covariate values are
measured without error). Unless you want to be very restrictive in how
to use the model (e.g. change from baseline in a study with similar
design and patients with same baseline distribution), it is best to
estimate the baseline. This also allows you to investigate relations (on
the individual level) between baseline and drug effect/disease
progression, etc.
I am sure you will find many old threads on this topic if you search the
nmusers archive. Additionally, here is a recent article:
Related Articles, LinksDansirikul C, Silber HE, Karlsson MO.Approaches
to handling pharmacodynamic baseline responses.J Pharmacokinet
Pharmacodyn. 2008 Jun;35(3):269-83. Epub 2008 Apr 30.
Cheers
Jakob
-----Original Message-----
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]
On Behalf Of LI,HONG
Sent: 29 July 2008 16:06
To: [email protected]
Subject: [NMusers] adding error term to covariate
Dear group,
I have a baseline as covariate in the model, and it is the same
measurement as the modeled variable. To me, it is reasonable to
believe that there is measurement error for this covariate. Is
there a way to incorporate some kind of error term into this
baseline covariate? Thanks a lot.
LI,HONG
Graduate student
University of Florida
School of Pharmacy
Department of Pharmaceutics
Office: P4-10
Phone: (352)273-7865
Dear group,
Thanks for your suggestions and the reference paper.
But I still have question that what we do expect from
incorprating error term to covariate considering the issue of model fitting.
I incorprated error term to covariate using method B2 in the paper but there is not any significant improvement in model fitting and parameter estimate. Dose that mean in my case residual error in covaraite is not really important? Or because residual error is small (< 10%), it is not necessary to worry about.
Thanks
Hong
Quoted reply history
On Wed Jul 30 15:03:01 EDT 2008, "Samtani, Mahesh [PRDUS]" <[EMAIL PROTECTED]> wrote:
> Hello,
>
> I was bit by this problem once in the past. My quick fix was to put etas on Smax & baseline (both parameters were estimated) and then put a block between these etas. It has a few limitations a) A fixed effect is modeled as a random effect; b) post-hoc plots of individual Smax and baseline were needed to visualize how the baseline influenced Smax; c) This forces a linear correlation when in fact the correlation can be curvilinear.
>
> Despite these limitations, the quick fix worked very well and I could meet my tight deadline. Can others point out other pros and cons of the above methodology.
>
> Thanks in advance for your feedback...Mahesh
>
> -----Original Message-----
> From: [EMAIL PROTECTED]
> [mailto:[EMAIL PROTECTED] Behalf Of Ribbing, Jakob
> Sent: Tuesday, July 29, 2008 4:52 PM
> To: LI,HONG; [email protected]
> Subject: RE: [NMusers] adding error term to covariate
>
> Dear Li,
>
> You are right in thinking that your baseline better not be treated as an ordinary covariate (where we pretend that the covariate values are measured without error). Unless you want to be very restrictive in how to use the model (e.g. change from baseline in a study with similar design and patients with same baseline distribution), it is best to estimate the baseline. This also allows you to investigate relations (on
>
> the individual level) between baseline and drug effect/disease
> progression, etc.
>
> I am sure you will find many old threads on this topic if you search the
>
> nmusers archive. Additionally, here is a recent article:
>
> Related Articles, LinksDansirikul C, Silber HE, Karlsson MO.Approaches
>
> to handling pharmacodynamic baseline responses.J Pharmacokinet
> Pharmacodyn. 2008 Jun;35(3):269-83. Epub 2008 Apr 30.
>
> Cheers
>
> Jakob
>
> -----Original Message-----
>
> From: [EMAIL PROTECTED] [ mailto:[EMAIL PROTECTED]
>
> On Behalf Of LI,HONG
> Sent: 29 July 2008 16:06
> To: [email protected]
> Subject: [NMusers] adding error term to covariate
>
> Dear group,
>
> I have a baseline as covariate in the model, and it is the same measurement as the modeled variable. To me, it is reasonable to believe that there is measurement error for this covariate. Is there a way to incorporate some kind of error term into this baseline covariate? Thanks a lot.
>
> LI,HONG
> Graduate student
> University of Florida
> School of Pharmacy
> Department of Pharmaceutics
> Office: P4-10
> Phone: (352)273-7865
LI,HONG
Graduate student
University of Florida
School of Pharmacy
Department of Pharmaceutics
My office: P4-10
Phone: (352)273-7865