Dear NONMEM Users,
I'm dealing with some longitudinal count data (number of AEs observed in fixed
time interval over certain time period). One feature with the data is the delay
in PK/PD relationship because it takes time to develop the AE (i.e. there is
lag time between the start of the AE formation process and the time you can
observe the AE). The onset of the process of developing an AE is clearly PK
dependent, but the exact time (or PK level) it onsets can never be observed.
In addition, the exact time that the AE first appears is also not available, we
only know the interval during which it first appears. What is a good approach
to handle this kind of delay? Is there any published paper for similar data
type?
Many thanks in advance for you suggestions!
Yaming Hang, Ph.D.
Pharmacometrics
Biogen Idec
14 Cambridge Center
Cambridge, MA 02142
Office: 781-464-1741
Fax: 617-679-2804
Email: [email protected]<mailto:[email protected]>
How to model delayed count data due to time needed to form an event
Dear Yaming,
My 2 cents ...
You may need a joint model for time to first AE with left and right
censoring (see tutorial on TTE from Holford and Lavieille last year at
PAGE) and a model for your count data.
This paper may help, although it does not fully cover your case:
Ito K, Hutmacher M, Liu J, Qiu R, Frame B, Miller R. Exposure-response
analysis for spontaneously reported dizziness in pregabalin-treated
patient with generalized anxiety disorder. Clin Pharmacol Ther. 2008
Jul;84(1):127-35.
Kind regards / Mit freundlichen Grüßen / Bien cordialement
Pascal Girard, PhD
[email protected]
Head of Modeling & Simulation - Oncology
Global Exploratory Medicine
Merck Serono S.A. · Geneva
Tel: +41.22.414.3549
Cell: +41.79.508.7898
Quoted reply history
From: Yaming Hang <[email protected]>
To: "[email protected]" <[email protected]>
Date: 18/06/2012 17:42
Subject: [NMusers] How to model delayed count data due to time
needed to form an event
Sent by: [email protected]
Dear NONMEM Users,
I’m dealing with some longitudinal count data (number of AEs observed in
fixed time interval over certain time period). One feature with the data
is the delay in PK/PD relationship because it takes time to develop the AE
(i.e. there is lag time between the start of the AE formation process and
the time you can observe the AE). The onset of the process of developing
an AE is clearly PK dependent, but the exact time (or PK level) it onsets
can never be observed. In addition, the exact time that the AE first
appears is also not available, we only know the interval during which it
first appears. What is a good approach to handle this kind of delay? Is
there any published paper for similar data type?
Many thanks in advance for you suggestions!
Yaming Hang, Ph.D.
Pharmacometrics
Biogen Idec
14 Cambridge Center
Cambridge, MA 02142
Office: 781-464-1741
Fax: 617-679-2804
Email: [email protected]
This message and any attachment are confidential and may be privileged or
otherwise protected from disclosure. If you are not the intended recipient, you
must not copy this message or attachment or disclose the contents to any other
person. If you have received this transmission in error, please notify the
sender immediately and delete the message and any attachment from your system.
Merck KGaA, Darmstadt, Germany and any of its subsidiaries do not accept
liability for any omissions or errors in this message which may arise as a
result of E-Mail-transmission or for damages resulting from any unauthorized
changes of the content of this message and any attachment thereto. Merck KGaA,
Darmstadt, Germany and any of its subsidiaries do not guarantee that this
message is free of viruses and does not accept liability for any damages caused
by any virus transmitted therewith.
Click http://www.merckgroup.com/disclaimer to access the German, French,
Spanish and Portuguese versions of this disclaimer.
Yaning, A trick Nick Holford showed me many years ago that works nicely: Set up two parallel systems. The "real" system has your "real" pk (doses, observations etc). The other system is the delay system. Doses are the same into the two systems. No pk observations in the delay system (probably non-pk observations of AEs in the delay system, dependent on the delay pk). But, key is that the same parameters (estimated from the "real" system) are also applied to the delayed system. So, the delayed system is just like the real system, except delayed. You can then use the concentrations in the delayed system to drive the AE's. I've even done this with two delay systems, works very nicely. So, Data looks like: ID TIME CMT AMT DV 1 0 1 100 . ; CMT 1 is "real" system 1 0 2 100 . ; CMT 2 is delayed system 1 1 1 . 100 ; real observation 1 2 1 . 50 ; real observation 1 3 1 . 25 ; real observation control file is: $MODEL COMP = (REAL) COMP = (DELAY) $PK K10 = THETA(1) K20 = K10 S1 = THETA(2) S2 = S1 ALAG2 = THETA(3) $DES DADT(1) = -K10*A(1) DADT(2) = -K20*A(2) This will give you a concentration in compartment 2 that is identical to compartment 1, except delayed by ALAG2. PK parameters are drive only by observation in compartment 1. Mark Sale MD President, Next Level Solutions, LLC www.NextLevelSolns.com 919-846-9185 A carbon-neutral company See our real time solar energy production at: http://enlighten.enphaseenergy.com/public/systems/aSDz2458
Quoted reply history
-------- Original Message --------
Subject: [NMusers] How to model delayed count data due to time needed
to form an event
From: Yaming Hang < [email protected] >
Date: Mon, June 18, 2012 11:08 am
To: " [email protected] " < [email protected] >
Dear NONMEM Users, I’m dealing with some longitudinal count data (number of AEs observed in fixed time interval over certain time period). One feature with the data is the delay in PK/PD relationship because it takes time to develop the AE (i.e. there is lag time between the start of the AE formation process and the time you can observe the AE). The onset of the process of developing an AE is clearly PK dependent, but the exact time (or PK level) it onsets can never be observed. In addition, the exact time that the AE first appears is also not available, we only know the interval during which it first appears. What is a good approach to handle this kind of delay? Is there any published paper for similar data type? Many thanks in advance for you suggestions! Yaming Hang, Ph.D. Pharmacometrics Biogen Idec 14 Cambridge Center Cambridge, MA 02142 Office: 781-464-1741 Fax: 617-679-2804 Email: [email protected]