Covariance step failed with BLQ method M3 using F_FLAG

1 messages 1 people Latest: Aug 29, 2012

Covariance step failed with BLQ method M3 using F_FLAG

From: Yan Xu Date: August 29, 2012 technical
Dear NONMEM users, I am trying to use the method M3 for including BLQ observation in the estimation process in NONMEM. I am using NM 7.1. The code is attached in the end of the e-mail. At the very beginning, the run failed. And then I made log transformation of DV, the run went through, with THETA similar estimation for CL/Ka/V to the model with no BLQ data included. However, the covariance step failed, with error message shown below: 0R MATRIX ALGORITHMICALLY SINGULAR AND ALGORITHMICALLY NON-POSITIVE-SEMIDEFINITE 0COVARIANCE STEP ABORTED My questions are: 1. Is log transformation of DV required for M3 method? 2. Why covariance step failed? What the error message means? Anything to fix? Any comment is highly appreciated. ------------------------------------------------------------------------------------------ $INPUT TIME ID EVID MDV DOSE AMT ADDL II DVO=DROP SEX BLQ FLAG DV $DATA xxxx.csv IGNORE=@ $SUBROUTINES ADVAN2 TRANS2 $PK STRT=DOSE TVCL=THETA(1) CL=TVCL*EXP(ETA(1)) TVV=THETA(2) V=TVV*EXP(ETA(2)) TVKA=THETA(3) KA=TVKA*EXP(ETA(3)) TVED50=THETA(4) F1=DOSE/(DOSE+TVED50) S2=V*46500/1000/1000/1000 $ERROR SIG=THETA(5) LOQ=LOG(184.4) ; define the LLOQ IPRED = LOG(0.025) IF(F.GT.0) IPRED = LOG(F) W=SIG IRES=IPRED-DV IWRES=IRES/W DUM=(LOQ-IPRED)/SIG CUMD = PHI(DUM) IF(BLQ.EQ.0) THEN F_FLAG=0 Y = IPRED+W*ERR(1) END IF IF (BLQ.EQ.1) THEN F_FLAG=1 Y=CUMD END IF $THETA ; (0,0.002) ; THETA(1) CL (0,0.19) ; THETA(2) V (0,0.1) ; THETA(3) KA (0,0.05) ; THETA(4) A (0,0.33); THETA(5) RES_ERR $OMEGA DIAGONAL(1) 0.09 ; [P] iivCL $OMEGA BLOCK(2) 0.09 0.01 0.09 $SIGMA 1 FIXED ;prop $EST METHOD=1 INTERACTION LAPLACE SLOW NUMERICAL SIG=4 MSFO=xxx.msf $COVARIANCE SLOW PRINT=E MATRIX=R -------------------------------------------------------------------- Thanks, Yan 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.