[NMusers] RE: IIV on res error

From: Eleveld, DJ <d.j.eleveld_at_umcg.nl>
Date: Fri, 16 Jan 2015 19:27:44 +0000

Hi Yuma,

My experience is that some model modifications can greatly reduce objfn but=
 make prediction actually worse. I like to use repeated 2-fold cross-valida=
tion since I am usually interested in accurate predictions for out-of-sampl=
e data. This may or may not be what you want your model to do. Once you hav=
e decided what you actually want your model to do then test for whatever th=
at thing is along with objective function, accepting into your model what i=
mproves both measures.

Look closely as to why some individuals get higher residual error. You can =
put it into or omit it from your model but you should have a good reason wh=
y. Do you trust the doses? Are there outlier data? Are all the covariates c=
orrect? Did people simply write down incorrect things? Look at the individu=
als who get assigned high residual error. Are the data reasonable? Did some=
body write down a wrong weight or age or height or BMI?

One danger is that you mask model misspecification with IIV on residual err=
or. If residual error correlates with say, obesity and your model works poo=
rly in the obese then you get improved model fit to the data by effectively=
 reducing the impact of obese on the model fit by assigning them higher res=
idual error. You dont want to mathematically reduce the impact of those ind=
ividuals that demonstrate real shortcomings of the structural model.

Warm regards,
Douglas Eleveld

From: owner-nmusers_at_globomaxnm.com [owner-nmusers_at_globomaxnm.com] on behalf=
 of Y.A. Bijleveld [y.a.bijleveld_at_amc.uva.nl]
Sent: Friday, January 16, 2015 3:09 PM
To: nmusers_at_globomaxnm.com
Subject: [NMusers] IIV on res error

Dear all,

I am modeling multi-center log-transformed neonatal data and have construct=
ed a 2-compartment model with ETA’s on Cl, V1 and V2. However, when intro=
ducing interindividual variability on the residual error the MOFV drops >15=
0 points, while previously significant relationships between clearance and =
covariates disappear. I find it strange that the introduction of the IIV ha=
s such an impact and don't fully understand. I have already checked the dat=
a for (extreme) outliers.

Can anyone shed some light?

Thank you so much.

Yuma Bijleveld.

F1=(BIO1**FDS12) * (BIO2**FDS34)

IF(F.GT.0)W = SQRT(THETA(3)**2)
(0, 75.7) ;1 V1
(0, 2.09) ;2 CL
(0, 0.376) ;3 add
(0, 3) ;4 Q
(0, 31.8) ;5 V2
(0, 3.3) ;6 GA
0.167 ;1 V1
0.0824 0.12 ;2 Cl
0.1 ;3 V2
0.1 ;4 RES


AMC Disclaimer : http://www.amc.nl/disclaimer


 De inhoud van dit bericht is vertrouwelijk en alleen bestemd voor de geadr=
esseerde(n). Anderen dan de geadresseerde(n) mogen geen gebruik maken van d=
it bericht, het niet openbaar maken of op enige wijze verspreiden of vermen=
igvuldigen. Het UMCG kan niet aansprakelijk gesteld worden voor een incompl=
ete aankomst of vertraging van dit verzonden bericht.

The contents of this message are confidential and only intended for the eye=
s of the addressee(s). Others than the addressee(s) are not allowed to use =
this message, to make it public or to distribute or multiply this message i=
n any way. The UMCG cannot be held responsible for incomplete reception or =
delay of this transferred message.
Received on Fri Jan 16 2015 - 14:27:44 EST

This archive was generated by hypermail 2.3.0 : Fri Sep 27 2019 - 16:42:44 EDT