M3 method with log transformed data

2 messages 2 people Latest: Nov 22, 2012

M3 method with log transformed data

From: Ayyappa Chaturvedula Date: November 21, 2012 technical
Dear Group, I want to implement M3 method with log transformed data. I want to use the additive error model in log domain and wrote the following code and the runs are fine. I want to make sure that I am doing this right. $ERROR CALLFL=0 IPRED=-3 IPRED=LOG(F) W=THETA(4) ;LOG ADDITIVE ERROR LOQ=LOG(0.03);loq is 0.3 DUM=(LOQ-IPRED)/W CUMD=PHI(DUM) IF (BQL.EQ.0) THEN IWRES = (DV-IPRED)/W F_FLAG=0 SIMDV = LOG(F)+W*ERR(1) Y=F+W*ERR(1) ENDIF IF (BQL.EQ.1) THEN IWRES = 0 F_FLAG=1 Y=CUMD ENDIF $EST METHOD=COND INTER LAPLACIAN $SIGMA 1 FIX Thank you for your help. Regards, Ayyappa

RE: M3 method with log transformed data

From: Mats Karlsson Date: November 22, 2012 technical
Dear Ayappa, This looks fine. However, with NONMEM 7.2 you can also write it as below. Estimating SIGMA rather than fixing it and estimate a multiplicative THETA may have speed advantages (and simpler coding for the new methods). $ERROR CALLFL=0 IPRED=-3 IPRED=LOG(F) W=SIGMA(1,1) ;LOG ADDITIVE ERROR LOQ=LOG(0.03);loq is 0.3 DUM=(LOQ-IPRED)/W CUMD=PHI(DUM) IF (BQL.EQ.0) THEN IWRES = (DV-IPRED)/W F_FLAG=0 SIMDV = LOG(F)+ERR(1) Y=F+W*ERR(1) ENDIF IF (BQL.EQ.1) THEN IWRES = 0 F_FLAG=1 Y=CUMD ENDIF $EST METHOD=COND INTER LAPLACIAN $SIGMA 1 Best regards, Mats Mats Karlsson, PhD Professor of Pharmacometrics Dept of Pharmaceutical Biosciences Faculty of Pharmacy Uppsala University Box 591 75124 Uppsala Phone: +46 18 4714105 Fax + 46 18 4714003
Quoted reply history
From: [email protected] [mailto:[email protected]] On Behalf Of Ayyappa Chaturvedula Sent: 22 November 2012 00:07 To: [email protected] Subject: [NMusers] M3 method with log transformed data Dear Group, I want to implement M3 method with log transformed data. I want to use the additive error model in log domain and wrote the following code and the runs are fine. I want to make sure that I am doing this right. $ERROR CALLFL=0 IPRED=-3 IPRED=LOG(F) W=THETA(4) ;LOG ADDITIVE ERROR LOQ=LOG(0.03);loq is 0.3 DUM=(LOQ-IPRED)/W CUMD=PHI(DUM) IF (BQL.EQ.0) THEN IWRES = (DV-IPRED)/W F_FLAG=0 SIMDV = LOG(F)+W*ERR(1) Y=F+W*ERR(1) ENDIF IF (BQL.EQ.1) THEN IWRES = 0 F_FLAG=1 Y=CUMD ENDIF $EST METHOD=COND INTER LAPLACIAN $SIGMA 1 FIX Thank you for your help. Regards, Ayyappa