Ann,
This error message usually means try NONMEM thinks your OMEGA initial
estimates are weird. A leading suspect would be the OMEGA estimates for
the two lag time parameters. Try fixing both of them to zero and see if
that gets your model going. If that doesn't work then you could try
fixing all OMEGA values to zero then adding them in one at a time.
You don't need to use $DES for your linear model. $SUB ADVAN11 is a
closed form 3 cpt model with bolus/zero order input (or ADVAN12 with
first order absorption as well).
An efficiency tip -- there is no point calling the time consuming PHI()
function until you really need it and also no need for a FLG data item.
You can simplify your $ERROR code like this:
$ERROR
SD=F*THETA(9) ; any residual error model you like assuming $SIGMA 1 FIX
LLOQ = 0.4 ; nm/L
IF (DV.GE.LLOQ) THEN ; non BQL values
F_FLAG=0
Y = F + SD*ERR(1))
ELSE ; BQL values
F_FLAG=1
Y = PHI((LLOQ-F)/SD)
ENDIF
Best wishes,
Nick
On 29/10/2010 9:50 p.m., Ann Rigby-Jones wrote:
Hello NMUsers
I am trying to reanalyse some data using the M3 method (Ahn et al,
Likelihood based approaches to handling data below the quantification
limit using NONMEM VI, J Pharmacokinet Pharmacodyn. 2008). I had
previously used M6 and wanted to try something potentially more robust.
I'm having problems with a "Hessian of posterior density is non
positive definite during search" error occurring as soon as NONMEM is
started. It is suggested that this may be caused by including time
zero records in the analysis (when MDV or EVID = 0) but I've
double-checked and I can't find any incidence of this.
My control stream is pasted below and I can sendan extract of data if
that helps?
Any ideas very gratefully received!
Many thanks J
Ann
$PROBLEM PK 3comp LAG BLQ M3
$INPUT ID DOSE AMT RATE DUR TIME DV EVID ART FLG AGE WGT MDV
$DATA step1&2PKBLQ.csv IGNORE=#
$SUBROUTINES ADVAN9 TOL=9
$MODEL
COMP(CENTRAL, DEFOBS) ;1
COMP(PERIPH1) ;2
COMP(PERIPH2) ;3
$PK
CL=THETA(1)*EXP(ETA(1))
Q2=THETA(2)*EXP(ETA(2))
Q3=THETA(3)*EXP(ETA(3))
V1=THETA(4)*EXP(ETA(4))
V2=THETA(5)*EXP(ETA(5))
V3=THETA(6)*EXP(ETA(6))
K10=CL/V1
K12=Q2/V1
K13=Q3/V1
K21=Q2/V2
K31=Q3/V3
S1=V1
IF (DUR.EQ.10) THEN
ALAG1=THETA(7)*EXP(ETA(7))
ELSE
ALAG1=THETA(8)*EXP(ETA(7))
ENDIF
$DES
DADT(1)=A(2)*K21 + A(3)*K31 - A(1)*(K10+K12+K13)
DADT(2)=A(1)*K12 - A(2)*K21
DADT(3)=A(1)*K13 - A(3)*K31
$ERROR
SIG1=THETA(9)
LOQ = 0.4 ; nm/L
IPRED = F
DUM = (LOQ-IPRED)/(SIG1*IPRED)
CUMD=PHI(DUM)
IF(FLG.EQ.0)THEN ; FLG=0, MDV=0 non BQL values
F_FLAG=0
Y = F*(1+SIG1*ERR(1))
ENDIF
IF(FLG.EQ.1)THEN ; FLG=1, MDV=0 BQL values
F_FLAG=1
Y=CUMD
ENDIF
$THETA (0, 696) ;CL
$THETA (0, 213) ;Q2
$THETA (0, 2600) ;Q3
$THETA (0, 5600) ;V1
$THETA (0, 50500) ;V2
$THETA (0, 38000) ;V3
$THETA (0.2167, 0.548,1) ;ALAG 10MIN
$THETA (0.00833, 0.101, 1) ;ALAG 1MIN
$THETA (0, 0.2) ;SIG1
$OMEGA (0.0121) ; ETA CL
$OMEGA (0.455) ; ETA Q2
$OMEGA (0.118) ; ETA Q3
$OMEGA (0.0919) ; ETA V1
$OMEGA (0.324) ; ETA V2
$OMEGA (0.0664) ; ETA V3
$OMEGA (1.02) ; ETA LAG SHARED
$SIGMA (0.0622)
$ESTIMATION METHOD=1 LAPLACIAN NUMERICAL SLOW PRINT=1 MAX=9999 NOABORT
SIG=3
$COVA
$TABLE ID EVID AMT TIME IPRED
NOPRINT FILE=AllRecords.txt
$TABLE ID CL Q2 Q3 V1 V2 V3
ETA1 ETA2 ETA3 ETA4 ETA5 ETA6 ETA7
FIRSTONLY NOPRINT NOAPPEND FILE=FirstRecords.txt
_______________________________________________________________________
*Ann Rigby-Jones PhD MRSC*
Research Fellow in Pharmacokinetics & Pharmacodynamics
Peninsula College of Medicine & Dentistry, UK
_______________________________________________________________________
--
Nick Holford, Professor Clinical Pharmacology
Dept Pharmacology& Clinical Pharmacology
University of Auckland,85 Park Rd,Private Bag 92019,Auckland,New Zealand
tel:+64(9)923-6730 fax:+64(9)373-7090 mobile:+64(21)46 23 53
email: n.holf...@auckland.ac.nz
http://www.fmhs.auckland.ac.nz/sms/pharmacology/holford
