The Standard Two-Stage and ELS

4 messages 4 people Latest: Mar 03, 2009

The Standard Two-Stage and ELS

From: Kim Kristensen Date: March 03, 2009 technical
Dear user Searching the user site I found a pre-published method by Dr's Beal and Shafer (link given below) to do The Standard Two-Stage Estimation using ELS with NONMEM. Out of curiosity I tried it, but I could not get it to work with NM6. Could anyone tell what I do wrong?? Control stream and test data are given below. http://www.cognigencorp.com/nonmem/nmo/topic035.html $PROBLEM Standard two stage $INPUT SUB AMT TIME DV EVID CMT MDV $DATA ASA_NM3.txt IGNORE=# $SUBROUTINE ADVAN1 TRANS2 $PK CL=THETA(1)*EXP(ETA(1)) V =THETA(2)*EXP(ETA(2)) $ERROR Y=F+F*EXP(ETA(3))*ERR(1) $THETA 5 50 ;these are rough typical population values $OMEGA 100 100 100 $SIGMA 1 FIXED $ESTIMATION MAXEVAL=0 POSTHOC METH=COND INTERACTION $TABLE SUB TIME CL V # SUB AMT TIME DV EVID CMT MDV 1 650 0 0 4 1 1 1 . 1 98 . 1 0 1 . 2 82 . 1 0 1 . 4 58 . 1 0 1 . 8 33 . 1 0 1 . 12 22 . 1 0 1 . 15 17 . 1 0 1 . 20 12 . 1 0 1 . 30 7.5 . 1 0 1 . 45 3.7 . 1 0 1 . 60 1.8 . 1 0 1 . 70 1.2 . 1 0 2 650 0 0 4 1 1 2 . 1 101 . 1 0 2 . 2 85 . 1 0 2 . 4 57 . 1 0 2 . 8 33 . 1 0 2 . 12 22 . 1 0 2 . 15 17 . 1 0 2 . 20 12 . 1 0 2 . 30 7.5 . 1 0 2 . 45 3.7 . 1 0 2 . 60 1.8 . 1 0 2 . 70 1.1 . 1 0 ________________________________ Kim Kristensen, M.Sc(pharm),Ph.D. Pharmacokineticist, Clinical Pharmacology & DMPK Medical Science Sweden AstraZeneca R&D Lund SE-221 87 Lund, Sweden Tel: +46 46 33 86 63 Fax: +46 46 336110 > Mobil: +46 (0) 705 33 77 18 > e-mail: [email protected] ________________________________

RE: The Standard Two-Stage and ELS

From: Brian Sadler Date: March 03, 2009 technical
Dear Kim, There is no ID variable in the $INPUT statement. Try: $INPUT SUB=ID AMT TIME DV EVID CMT MDV Cheers... Brian Brian M. Sadler, Ph.D. Strategic PK Consulting, LLC 1209 Seabrook Avenue Cary, North Carolina 27511 USA [email protected] +1 (919) 523-5485 (mobile) +1 (508) 408-0250 (fax)
Quoted reply history
-----Original Message----- From: [email protected] [mailto:[email protected]] On Behalf Of Kristensen, Kim Sent: Tuesday, March 03, 2009 4:26 AM To: [email protected] Subject: [NMusers] The Standard Two-Stage and ELS Dear user Searching the user site I found a pre-published method by Dr's Beal and Shafer (link given below) to do The Standard Two-Stage Estimation using ELS with NONMEM. Out of curiosity I tried it, but I could not get it to work with NM6. Could anyone tell what I do wrong?? Control stream and test data are given below. http://www.cognigencorp.com/nonmem/nmo/topic035.html $PROBLEM Standard two stage $INPUT SUB AMT TIME DV EVID CMT MDV $DATA ASA_NM3.txt IGNORE=# $SUBROUTINE ADVAN1 TRANS2 $PK CL=THETA(1)*EXP(ETA(1)) V =THETA(2)*EXP(ETA(2)) $ERROR Y=F+F*EXP(ETA(3))*ERR(1) $THETA 5 50 ;these are rough typical population values $OMEGA 100 100 100 $SIGMA 1 FIXED $ESTIMATION MAXEVAL=0 POSTHOC METH=COND INTERACTION $TABLE SUB TIME CL V # SUB AMT TIME DV EVID CMT MDV 1 650 0 0 4 1 1 1 . 1 98 . 1 0 1 . 2 82 . 1 0 1 . 4 58 . 1 0 1 . 8 33 . 1 0 1 . 12 22 . 1 0 1 . 15 17 . 1 0 1 . 20 12 . 1 0 1 . 30 7.5 . 1 0 1 . 45 3.7 . 1 0 1 . 60 1.8 . 1 0 1 . 70 1.2 . 1 0 2 650 0 0 4 1 1 2 . 1 101 . 1 0 2 . 2 85 . 1 0 2 . 4 57 . 1 0 2 . 8 33 . 1 0 2 . 12 22 . 1 0 2 . 15 17 . 1 0 2 . 20 12 . 1 0 2 . 30 7.5 . 1 0 2 . 45 3.7 . 1 0 2 . 60 1.8 . 1 0 2 . 70 1.1 . 1 0 ________________________________ Kim Kristensen, M.Sc(pharm),Ph.D. Pharmacokineticist, Clinical Pharmacology & DMPK Medical Science Sweden AstraZeneca R&D Lund SE-221 87 Lund, Sweden Tel: +46 46 33 86 63 Fax: +46 46 336110 > Mobil: +46 (0) 705 33 77 18 > e-mail: [email protected] ________________________________

RE: The Standard Two-Stage and ELS

From: Mats Karlsson Date: March 03, 2009 technical
Dear Kim, Using your model and data file (with ID in $INPUT results in the desired estimation. See TABLE from lst file: **************************************************************************** ******************************************** ******************** ******************** ******************** TABLES OF DATA AND PREDICTIONS ******************** ******************** ******************** **************************************************************************** ******************************************** 1TABLE NO. 1 LINE NO. SUB TIME CL V DV PRED RES WRES 1 + 1.00E+00 0.00E+00 5.03E+00 4.97E+01 0.00E+00 6.50E+02 0.00E+00 0.00E+00 2 + 1.00E+00 1.00E+00 5.03E+00 4.97E+01 9.80E+01 5.88E+02 -4.90E+02 -7.81E-02 3 + 1.00E+00 2.00E+00 5.03E+00 4.97E+01 8.20E+01 5.32E+02 -4.50E+02 -7.29E-02 4 + 1.00E+00 4.00E+00 5.03E+00 4.97E+01 5.80E+01 4.36E+02 -3.78E+02 -6.00E-02 5 + 1.00E+00 8.00E+00 5.03E+00 4.97E+01 3.30E+01 2.92E+02 -2.59E+02 -3.47E-02 6 + 1.00E+00 1.20E+01 5.03E+00 4.97E+01 2.20E+01 1.96E+02 -1.74E+02 -1.64E-02 7 + 1.00E+00 1.50E+01 5.03E+00 4.97E+01 1.70E+01 1.45E+02 -1.28E+02 -6.71E-03 8 + 1.00E+00 2.00E+01 5.03E+00 4.97E+01 1.20E+01 8.80E+01 -7.60E+01 5.01E-03 9 + 1.00E+00 3.00E+01 5.03E+00 4.97E+01 7.50E+00 3.24E+01 -2.49E+01 1.96E-02 10 + 1.00E+00 4.50E+01 5.03E+00 4.97E+01 3.70E+00 7.22E+00 -3.52E+00 2.99E-02 11 + 1.00E+00 6.00E+01 5.03E+00 4.97E+01 1.80E+00 1.61E+00 1.89E-01 3.57E-02 12 + 1.00E+00 7.00E+01 5.03E+00 4.97E+01 1.20E+00 5.93E-01 6.07E-01 4.21E-02 13 + 2.00E+00 0.00E+00 5.27E+00 4.74E+01 0.00E+00 6.50E+02 0.00E+00 0.00E+00 14 + 2.00E+00 1.00E+00 5.27E+00 4.74E+01 1.01E+02 5.88E+02 -4.87E+02 -7.75E-02 15 + 2.00E+00 2.00E+00 5.27E+00 4.74E+01 8.50E+01 5.32E+02 -4.47E+02 -7.24E-02 16 + 2.00E+00 4.00E+00 5.27E+00 4.74E+01 5.70E+01 4.36E+02 -3.79E+02 -6.02E-02 17 + 2.00E+00 8.00E+00 5.27E+00 4.74E+01 3.30E+01 2.92E+02 -2.59E+02 -3.46E-02 18 + 2.00E+00 1.20E+01 5.27E+00 4.74E+01 2.20E+01 1.96E+02 -1.74E+02 -1.64E-02 19 + 2.00E+00 1.50E+01 5.27E+00 4.74E+01 1.70E+01 1.45E+02 -1.28E+02 -6.64E-03 20 + 2.00E+00 2.00E+01 5.27E+00 4.74E+01 1.20E+01 8.80E+01 -7.60E+01 5.12E-03 21 + 2.00E+00 3.00E+01 5.27E+00 4.74E+01 7.50E+00 3.24E+01 -2.49E+01 1.98E-02 22 + 2.00E+00 4.50E+01 5.27E+00 4.74E+01 3.70E+00 7.22E+00 -3.52E+00 3.06E-02 23 + 2.00E+00 6.00E+01 5.27E+00 4.74E+01 1.80E+00 1.61E+00 1.89E-01 3.74E-02 Best regards, Mats Mats Karlsson, PhD Professor of Pharmacometrics Dept of Pharmaceutical Biosciences Uppsala University Box 591 751 24 Uppsala Sweden phone: +46 18 4714105 fax: +46 18 471 4003
Quoted reply history
-----Original Message----- From: [email protected] [mailto:[email protected]] On Behalf Of Kristensen, Kim Sent: Tuesday, March 03, 2009 5:36 PM To: Brian M. Sadler; [email protected] Subject: RE: [NMusers] The Standard Two-Stage and ELS Dear Brian Thanks for the suggestion. I've tried changing the ID and the first row to EVID=1. However, the posthoc (ELS) modelling of data, as suggested by Dr's Beal and Shafer, dosn't seem to go through as the parameter are all left unchanged i.e. as initial estimates! /Kim -----Original Message----- From: Brian M. Sadler [mailto:[email protected]] Sent: den 3 mars 2009 12:49 To: Kristensen, Kim; [email protected] Subject: RE: [NMusers] The Standard Two-Stage and ELS Dear Kim, There is no ID variable in the $INPUT statement. Try: $INPUT SUB=ID AMT TIME DV EVID CMT MDV Cheers... Brian Brian M. Sadler, Ph.D. Strategic PK Consulting, LLC 1209 Seabrook Avenue Cary, North Carolina 27511 USA [email protected] +1 (919) 523-5485 (mobile) +1 (508) 408-0250 (fax) -----Original Message----- From: [email protected] [mailto:[email protected]] On Behalf Of Kristensen, Kim Sent: Tuesday, March 03, 2009 4:26 AM To: [email protected] Subject: [NMusers] The Standard Two-Stage and ELS Dear user Searching the user site I found a pre-published method by Dr's Beal and Shafer (link given below) to do The Standard Two-Stage Estimation using ELS with NONMEM. Out of curiosity I tried it, but I could not get it to work with NM6. Could anyone tell what I do wrong?? Control stream and test data are given below. http://www.cognigencorp.com/nonmem/nmo/topic035.html $PROBLEM Standard two stage $INPUT SUB AMT TIME DV EVID CMT MDV $DATA ASA_NM3.txt IGNORE=# $SUBROUTINE ADVAN1 TRANS2 $PK CL=THETA(1)*EXP(ETA(1)) V =THETA(2)*EXP(ETA(2)) $ERROR Y=F+F*EXP(ETA(3))*ERR(1) $THETA 5 50 ;these are rough typical population values $OMEGA 100 100 100 $SIGMA 1 FIXED $ESTIMATION MAXEVAL=0 POSTHOC METH=COND INTERACTION $TABLE SUB TIME CL V # SUB AMT TIME DV EVID CMT MDV 1 650 0 0 4 1 1 1 . 1 98 . 1 0 1 . 2 82 . 1 0 1 . 4 58 . 1 0 1 . 8 33 . 1 0 1 . 12 22 . 1 0 1 . 15 17 . 1 0 1 . 20 12 . 1 0 1 . 30 7.5 . 1 0 1 . 45 3.7 . 1 0 1 . 60 1.8 . 1 0 1 . 70 1.2 . 1 0 2 650 0 0 4 1 1 2 . 1 101 . 1 0 2 . 2 85 . 1 0 2 . 4 57 . 1 0 2 . 8 33 . 1 0 2 . 12 22 . 1 0 2 . 15 17 . 1 0 2 . 20 12 . 1 0 2 . 30 7.5 . 1 0 2 . 45 3.7 . 1 0 2 . 60 1.8 . 1 0 2 . 70 1.1 . 1 0 ________________________________ Kim Kristensen, M.Sc(pharm),Ph.D. Pharmacokineticist, Clinical Pharmacology & DMPK Medical Science Sweden AstraZeneca R&D Lund SE-221 87 Lund, Sweden Tel: +46 46 33 86 63 Fax: +46 46 336110 > Mobil: +46 (0) 705 33 77 18 > e-mail: [email protected] ________________________________

RE: The Standard Two-Stage and ELS

From: Mouksassi Mohamad-Samer Date: March 03, 2009 technical
Kim, The first stage estimates are those calculated POSTHOC and output in the table. Then you need to conduct stage two and conduct a sort of summary on the parameters calculated in Stage I. Maxeval =0 and the large OMEGA are just to avoid the POSTHOC being influenced by the population value. As a sensitivity you may want to try different values of Omega e.g: 1000 and see how your individual PK parameters change. Bests, Samer
Quoted reply history
-----Original Message----- From: [email protected] on behalf of Brian M. Sadler Sent: Tue 3/3/2009 13:46 To: 'Kristensen, Kim'; [email protected] Subject: RE: [NMusers] The Standard Two-Stage and ELS Kim, In the $ESTIMATION statement, you need to change MAXEVAL=0 to MAXEVAL=9999 (or some other non-zero integer) before the estimation process will proceed beyond the initial estimates. Cheers... Brian -----Original Message----- From: [email protected] [mailto:[email protected]] On Behalf Of Kristensen, Kim Sent: Tuesday, March 03, 2009 11:36 AM To: Brian M. Sadler; [email protected] Subject: RE: [NMusers] The Standard Two-Stage and ELS Dear Brian Thanks for the suggestion. I've tried changing the ID and the first row to EVID=1. However, the posthoc (ELS) modelling of data, as suggested by Dr's Beal and Shafer, dosn't seem to go through as the parameter are all left unchanged i.e. as initial estimates! /Kim -----Original Message----- From: Brian M. Sadler [mailto:[email protected]] Sent: den 3 mars 2009 12:49 To: Kristensen, Kim; [email protected] Subject: RE: [NMusers] The Standard Two-Stage and ELS Dear Kim, There is no ID variable in the $INPUT statement. Try: $INPUT SUB=ID AMT TIME DV EVID CMT MDV Cheers... Brian Brian M. Sadler, Ph.D. Strategic PK Consulting, LLC 1209 Seabrook Avenue Cary, North Carolina 27511 USA [email protected] +1 (919) 523-5485 (mobile) +1 (508) 408-0250 (fax) -----Original Message----- From: [email protected] [mailto:[email protected]] On Behalf Of Kristensen, Kim Sent: Tuesday, March 03, 2009 4:26 AM To: [email protected] Subject: [NMusers] The Standard Two-Stage and ELS Dear user Searching the user site I found a pre-published method by Dr's Beal and Shafer (link given below) to do The Standard Two-Stage Estimation using ELS with NONMEM. Out of curiosity I tried it, but I could not get it to work with NM6. Could anyone tell what I do wrong?? Control stream and test data are given below. http://www.cognigencorp.com/nonmem/nmo/topic035.html $PROBLEM Standard two stage $INPUT SUB AMT TIME DV EVID CMT MDV $DATA ASA_NM3.txt IGNORE=# $SUBROUTINE ADVAN1 TRANS2 $PK CL=THETA(1)*EXP(ETA(1)) V =THETA(2)*EXP(ETA(2)) $ERROR Y=F+F*EXP(ETA(3))*ERR(1) $THETA 5 50 ;these are rough typical population values $OMEGA 100 100 100 $SIGMA 1 FIXED $ESTIMATION MAXEVAL=0 POSTHOC METH=COND INTERACTION $TABLE SUB TIME CL V # SUB AMT TIME DV EVID CMT MDV 1 650 0 0 4 1 1 1 . 1 98 . 1 0 1 . 2 82 . 1 0 1 . 4 58 . 1 0 1 . 8 33 . 1 0 1 . 12 22 . 1 0 1 . 15 17 . 1 0 1 . 20 12 . 1 0 1 . 30 7.5 . 1 0 1 . 45 3.7 . 1 0 1 . 60 1.8 . 1 0 1 . 70 1.2 . 1 0 2 650 0 0 4 1 1 2 . 1 101 . 1 0 2 . 2 85 . 1 0 2 . 4 57 . 1 0 2 . 8 33 . 1 0 2 . 12 22 . 1 0 2 . 15 17 . 1 0 2 . 20 12 . 1 0 2 . 30 7.5 . 1 0 2 . 45 3.7 . 1 0 2 . 60 1.8 . 1 0 2 . 70 1.1 . 1 0 ________________________________ Kim Kristensen, M.Sc(pharm),Ph.D. Pharmacokineticist, Clinical Pharmacology & DMPK Medical Science Sweden AstraZeneca R&D Lund SE-221 87 Lund, Sweden Tel: +46 46 33 86 63 Fax: +46 46 336110 > Mobil: +46 (0) 705 33 77 18 > e-mail: [email protected] ________________________________