NONMEM Users Network Archive

Hosted by Cognigen

RE: Truncated Emax

From: Elassaiss - Schaap, J. <jeroen.elassaiss-schaap>
Date: Tue, 20 Dec 2011 09:23:22 +0100

Hi Martin,

Thank you for pointing this out. I actually do agree with you! I certainly =
did not imply that deleting censored data is a guarantee for unbiased resul=
ts.

But please keep in mind that especially with pain censoring is not arbitrar=
y. It is actually a meaningful border, for example unbearable pain or perha=
ps safety of currents in this case. And as I referred to, I have compared m=
odels for pain with deletion or with M3 but could not find any difference i=
n results even with a high amount of censoring. My finding surprised me at =
first, but when I discussed this with our residential senior statistician h=
e told me that this, unbiased results after deleting of censored data, was =
common experience.

I would be curious about experience from others on this list! Please do sha=
re in if you have seen results one way or the other.

Best regards,
Jeroen

________________________________
From: Martin Bergstrand [mailto:martin.bergstrand
Sent: Tuesday, December 20, 2011 08:27
To: 'Francois Gaudreault'; nmusers
Cc: 'Waqas Sadiq'; Elassaiss - Schaap, J. (Jeroen)
Subject: RE: [NMusers] Truncated Emax

Dear François,

I do not agree with Jeroen that less than ~1/3 of total data censored is a =
guarantee for that these observations can be ignored without substantial bi=
as. I think this is highly dependent on the nature of the model (system), t=
he limit of quantification in relationship to Emax etc. To make a statement=
 on what percentage of censored data (out of the total) that will result in=
 negligible bias is never a good idea since it might be that only a small p=
ortion of the total data speaks to a specific parameter. If a substantial a=
mount of that small portion of data is censored it can have important impli=
cations while it is still just a minor percentage that is missing out of th=
e total. But importantly you do not need to take anyone's word for this si=
nce you can test it you self with simulation based diagnostics and/or simul=
ation and re-estimation with the applied censoring.

The way that I would go about this issue is that I would take into account =
also the censored observations. The below code is just a slight alteration =
of the M3 method suggested by S. Beal for the handling of observations belo=
w the limit o detection (BQL)[1]. More detail on how this is best implement=
ed in NONMEM is given in a paper by Anh et.al [2]. Me and others have also =
several times discussed how to best diagnose models in the presence of cens=
ored observations (see NMusers archive).

;;; ---------------------------------------------------------
$ERROR
W = THETA(.) ; Residual error model (in this example simple =
additive)
ULOQ = 10 ; Upper limit of detection (10mA)
IPRED = PT ; Individual prediction of perception threshold=
 according to your desired model
DUM = (IPRED-ULOQ)/SIG
CUMD = PHI(DUM)

; Flag variable CENS in dataset. CENS=1 => observation >ULOQ
IF(CENS.EQ.0) THEN ; <ULOQ
     F_FLAG = 0
     Y = IPRED+SIG*ERR(1)
ENDIF
IF(ALQ.EQ.1) THEN ; >ULOQ
     F_FLAG = 1
     Y = CUMD
ENDIF
;;; ---------------------------------------------------------

Obs. When applying this code the SIGMA variance is fixed to 1 ($SIGMA 1 FIX=
) and the Lapalcian estimation option needs to be utilized (or possibly SAE=
M etc.) [2].

This type of coding have previously been successfully applied by my colleag=
ue Waqas Sadiq. A manuscript on this project is currently in preparation an=
d might be referenced once published (look out).

[1] Beal SL. Ways to fit a PK model with some data below the quantification=
 limit. J Pharmacokinet Pharmacodyn. 2001 Oct;28(5):481-504.

[2] Ahn JE, Karlsson MO, Dunne A, Ludden TM. Likelihood based approaches to=
 handling data below the quantification limit using NONMEM VI. J Pharmacoki=
net Pharmacodyn. 2008 Aug 7.


Kind regards,

Martin Bergstrand, PhD
Pharmacometrics Research Group
Dept of Pharmaceutical Biosciences
Uppsala University, Sweden
martin.bergstrand

Visiting scientist:
Mahidol-Oxford Tropical Medicine Research Unit,
Bangkok, Thailand
Phone: +66 8 9796 7611


From: owner-nmusers
 Behalf Of Elassaiss - Schaap, J. (Jeroen)
Sent: Tuesday, December 20, 2011 4:12 AM
To: Francois Gaudreault; nmusers
Subject: RE: [NMusers] Truncated Emax

Hi Francois,

For pain measurements it is not uncommon to analyze data with a upper limit=
 of quantitation. You can follow the literature on BQL, only reversing from=
 a lower limit to an upper limIt. In my experience just deleting censored d=
ata works fine, certainly as a first attempt, as long as censoring stays be=
low ~1/3 of total data.

Best regards,
Jeroen

J. Elassaiss-Schaap
Scientist PK/PD
MSD
PO Box 20, 5340 BH Oss, Netherlands
Phone: + 31 412 66 9320
Fax: + 31 412 66 2506
e-mail: jeroen.elassaiss

________________________________
From: owner-nmusers
ilto:owner-nmusers
Sent: Monday, December 19, 2011 21:40
To: nmusers
Subject: [NMusers] Truncated Emax
Dear NM users

I am currently developing a PK PD model for local anesthetics using a seque=
ntial approach with ADVAN6. The PD model is a sigmoid Emax with an effect c=
ompartment (Ce).

The intensity and duration of nerve blockade are monitored throughout the p=
erioperative period in patients using a quantitative pharmacodynamic endpoi=
nt, i.e, the current perception threshold (CPT) REF: Can. J. Anesth, 57 (S1=
) 2010). Briefly, CPT is evaluated before and after the administration of t=
he local anesthectic. Data are normalized by baseline using the following e=
quation :
(observed-baseline) / (max-baseline) *100 (%)

Here is the problem. The device only goes to a maximum of 10 mA. In some pa=
tients, the real Emax is much higher. Any ideas on how handle a truncated E=
max ?

Thanks in advance



--
François Gaudreault, Ph.D. Candidate
Pharmacométrie / Pharmacometrics
Charger de cours / Lecturer
Faculté de pharmacie / Faculty of Pharmacy
Université de Montréal


Notice: This e-mail message, together with any attachments, contains
information of Merck & Co., Inc. (One Merck Drive, Whitehouse Station,
New Jersey, USA 08889), and/or its affiliates Direct contact information
for affiliates is available at
http://www.merck.com/contact/contacts.html) that may be confidential,
proprietary copyrighted and/or legally privileged. It is intended solely
for the use of the individual or entity named on this message. If you are
not the intended recipient, and have received this message in error,
please notify us immediately by reply e-mail and then delete it from
your system.

Received on Tue Dec 20 2011 - 03:23:22 EST

The NONMEM Users Network is maintained by ICON plc. Requests to subscribe to the network should be sent to: nmusers-request@iconplc.com.

Once subscribed, you may contribute to the discussion by emailing: nmusers@globomaxnm.com.