Certara, Inc. · NASDAQ:CERT

Certara — Technical Presentation (2015)

18 slides from Certara's technical presentation, 2015

Slide 1: Population Analysis of Complex Multidimensional Responses using Phoenix NLME
Population Analysis of Complex Multidimensional Responses using Phoenix NLME Dr. Serge Guzy, President of Pop_Pharm
Slide 2: Responses seen in clinical trials
Responses seen in clinical trials • Trial data include – continuous responses • Pharmacokinetic, pharmacodynamics, disease response models – Non continuous responses • Multi level data (e.g., disease state,0,1,2,3) • Count data (like number of cardiac events accumulated from observation to observati
Slide 3: 3 modeling tools available in Phoenix
3 modeling tools available in Phoenix • Built in libraries – NLME has built in many PK/PD models with a fully integrated interface making very easy to model most of the common PK/PD models seen in Pharmacometrics • Graphical mode – If these models do not describe totally the system, you can shift to
Slide 4: Handling combined PK and multi level data
Handling combined PK and multi level data • Most of the statisticians can easily handle multi level data using SAS – However PK complex dynamics cannot be combined with multi-level data using the SAS system – Other software require you to write the full code that define both the dynamic and the prob
Slide 6: The strategy – PK
The strategy – PK • Plot • NCA • Model 1 compartment • Model 2 compartment – Select best model » Add additional extravascular compartment as prior information suggesting 2 oral routes © Copyright 2012 Certara, L.P. All rights reserved. 6
Slide 11: Categorical response
Categorical response We can see that the probability to be in category 3 increase with concentration: This is a hint to suggest the following equation for the Prob to be 3 Prob(3)=exp(a*C+b)/(1+exp(a*C+b))=ilogit(a*C+b) with a being positive Prob(3or2)=ilogit(a*C+b+d), d>0 Prob(3or2or1)=ilogit(a*C+b
Slide 12: Final PK/multi level responses model
Final PK/multi level responses model Prob(3)=exp(a*C+b)/(1+exp(a*C+b))=ilogit(a*C+b) with a being positive Prob(3or2)=ilogit(a*C+b+d), d>0 Prob(3or2or1)=ilogit(a*C+b+d+e);e>0 © Copyright 2012 Certara, L.P. All rights reserved. 12
Slide 17: Training Courses and Next Webinar in the Series
Training Courses and Next Webinar in the Series • Certara offers training courses for Phoenix NLME • New Courses added in Europe – Sample listing: • Introduction to Population Modeling using Phoenix NLME – 24-25 March, 2015 in Munich, Germany • Intermediate Population Modeling using Phoenix NLME – 2
Slide 18: Contacts & Follow-Up: Thank you!
Contacts & Follow-Up: Thank you! Americas Telephone (US): 888-708-7444 Telephone (Intl): +1-919-852-4685 Email: [email protected] Europe Telephone: +33 1 56 95 17 10 Email: [email protected] India Telephone: +91 984 500 6625 Email: [email protected] Japan Telephone: +81 (0)50-3786-9898 Email: sales@

Certara — Technical Presentation (2015) on 6ix

Certara — Technical Presentation (2015)

Slide 1: Population Analysis of Complex Multidimensional Responses using Phoenix NLME

Slide 2: Responses seen in clinical trials

Slide 3: 3 modeling tools available in Phoenix

Slide 4: Handling combined PK and multi level data

Slide 5: The data

Slide 6: The strategy – PK

Slide 7: NCA results

Slide 8: 1c vs 2c results

Slide 9: Results Final PK Graphical Model

Slide 10: Results Final PK model

Slide 11: Categorical response

Slide 12: Final PK/multi level responses model

Slide 13: Results

Slide 14: Cat descriptive statistics by dosing: Observed vs Predicted

Slide 15: Cat descriptive statistics by dosing: Observed vs Predicted

Slide 16: Q&A

Slide 17: Training Courses and Next Webinar in the Series

Slide 18: Contacts & Follow-Up: Thank you!