1st Edition

Exposure-Response Modeling
Methods and Practical Implementation




ISBN 9781466573208
Published July 17, 2015 by Chapman and Hall/CRC
351 Pages 27 B/W Illustrations

USD $105.00

Prices & shipping based on shipping country


Preview

Book Description

Discover the Latest Statistical Approaches for Modeling Exposure-Response Relationships

Written by an applied statistician with extensive practical experience in drug development, Exposure-Response Modeling: Methods and Practical Implementation explores a wide range of topics in exposure-response modeling, from traditional pharmacokinetic-pharmacodynamic (PKPD) modeling to other areas in drug development and beyond. It incorporates numerous examples and software programs for implementing novel methods.

The book describes using measurement error models to treat sequential modeling, fitting models with exposure and response driven by complex dynamics, and survival analysis with dynamic exposure history. It also covers Bayesian analysis and model-based Bayesian decision analysis, causal inference to eliminate confounding biases, and exposure-response modeling with response-dependent dose/treatment adjustments (dynamic treatment regimes) for personalized medicine and treatment adaptation.

Many examples illustrate the use of exposure-response modeling in experimental toxicology, clinical pharmacology, epidemiology, and drug safety. Some examples demonstrate how to solve practical problems while others help with understanding concepts and evaluating the performance of new methods. The provided SAS and R codes enable readers to test the approaches in their own scenarios.

Although application oriented, this book also gives a systematic treatment of concepts and methodology. Applied statisticians and modelers can find details on how to implement new approaches. Researchers can find topics for or applications of their work. In addition, students can see how complicated methodology and models are applied to practical situations.

Table of Contents

Introduction
Multifaceted exposure-response relationships
Practical scenarios in ER modeling
Models and modeling in exposure-response analysis
Model-based decision-making and drug development
Drug regulatory guidance for analysis of exposure-response relationship
Examples and modeling software

Basic exposure and exposure-response models
Models based on pharmacological mechanisms
Statistical models
Transformations
Semiparametric and nonparametric models
Comments and bibliographic notes

Dose-exposure and exposure-response models for longitudinal data
Linear mixed models for exposure-response relationships
Modeling exposures with linear mixed models
Nonlinear mixed ER models
Modeling exposure with a population PK model
Mixed effect models specified by differential equations
Generalized linear mixed model and generalized estimating equation
Generalized nonlinear mixed models
Testing variance components in mixed models
Nonparametric and semiparametric models with random effects
On distributions of random effects
Bibliographic notes

Sequential and simultaneous exposure-response modeling
Joint models for exposure and response
Simultaneous modeling of exposure and response models
Sequential exposure-response modeling
Measurement error models and regression calibration
Instrumental variable methods
Modeling multiple exposure and response
Internal validation data and partially observed and surrogate exposure measures
Comments and bibliographic notes

Exposure-risk modeling for time-to-event data
An example
Basic concepts and models for time-to-event data
Dynamic exposure model as a time varying covariate
Multiple TTE and competing risks
Models for recurrent events
Frailty: Random effects in TTE models
Joint modeling of exposure and time to event
Interval censored data
Model identification and misspecification
Random sample simulation from exposure-risk models
Comments and bibliographic notes

Modeling dynamic exposure-response relationships
Effect compartment models
Indirect response models
Disease process models
Fitting dynamic models for longitudinal data
Semiparametric and nonparametric approaches
Dynamic linear and generalized linear models
Testing hysteresis
Comments and bibliographic notes

Bayesian modeling and model-based decision analysis
Bayesian modeling
Bayesian decision analysis
Decisions under uncertainty and with multiple objectives
Evidence synthesis and mixed treatment comparison
Comments and bibliographic notes

Confounding bias and causal inference in exposure-response modeling
Introduction
Confounding factors and confounding biases
Causal effect and counterfactuals
Classical adjustment methods
Directional acyclic graphs
Bias assessment
Instrumental variable
Joint modeling of exposure and response
Study designs robust to confounding bias or allowing the use of instrument variables
Doubly robust estimates
Comments and bibliographic notes

Dose-response relationship, dose determination, and adjustment
Marginal dose-response relationships
Dose-response relationship as a combination of dose-exposure and exposure-response relationships
Dose determination: Dose-response or dose-exposure-response modeling approaches?
Dose adjustment
Dose adjustment and causal effect estimation
Sequential decision analysis
Dose determination: Design issues
Comments and bibliographic notes

Implementation using software
Two key elements: Model and data
Linear mixed and generalized linear mixed models
Nonlinear mixed models
A very quick guide to NONMEM

Appendix

Bibliography

Index

...
View More

Author(s)

Biography

Jixian Wang is a principal statistician at Celgene International, Switzerland. He worked on drug development for 14 years at GSK and Novartis Pharma and was an academic researcher at Edinburgh University and Dundee University, where he is still an honorary research fellow. His research interests include statistical methodology and its applications to real problems in pharmaceuticals, including exposure-safety, PKPD modeling, treatment/dose selection, health economics, benefit-risk and health technology assessments, and optimal trial design.

Reviews

"...the book is worth reading as it takes the reader all the way from basic to state-of-the-art exposure-response modeling approaches and challenges. It focuses on detailed mathematical derivations, with many insights based on practical experience. Moreover, many data examples are accompanied by software code ..."
~Biometrical Journal

" . . . the greatest strength of this book is that the models and methodologies are always motivated and explained by applications and examples, which effectively communicates to readers the basic ideas behind complex methodologies. Also, practical implementation and computer code are discussed alongside the methods, which will help readers to apply the methods to their own data."
~University of Texas Health Science Center at Houston

"In summary, this book provides a good overview of various scenarios of ER relationship assessment and modelling, and the appropriate statistical approaches. With a lot of hints and tips and numeric examples that illustrate various aspects it is easy to read for both statisticians and nonstatisticians; the numerous programming code examples also make the book notably expedient."
~Dirk Lindner, ISCB