s'authentifier
version française rss feed
Metrics for external model evaluation with an application to the population pharmacokinetics of gliclazide.
Brendel K., Comets E., Laffont C., Laveille C., Mentré F.
Pharmaceutical Research / Pharmaceutical Research (Dordrecht) 23, 9 (2006) 2036-49 - http://www.hal.inserm.fr/inserm-00189557/en/
 (16906454) 
Metrics for external model evaluation with an application to the population pharmacokinetics of gliclazide.
Karl Brendel () 1, 2, Emmanuelle Comets1, 2, Céline Laffont3, Christian Laveille3, 4, France Mentré1, 2
1 :  Modèles et méthodes de l'évaluation thérapeutique des maladies chroniques
INSERM : U738 – Université Paris VII - Paris Diderot
Faculté de médecine Paris 7 16, Rue Henri Huchard 75018 Paris
France
2 :  Département d'épidémiologie, biostatistique et recherche clinique
Assistance publique - Hôpitaux de Paris (AP-HP) – Hôpital Bichat - Claude Bernard
46 rue Henri Huchard 75018 Paris
France
3 :  Institut de Recherches Internationales Servier
SERVIER
Courbevoie
France
4 :  EXPRIMO NV
Exprimo NV
Lumnen
Belgique
PURPOSE: The aim of this study is to define and illustrate metrics for the external evaluation of a population model. MATERIALS AND METHODS: In this paper, several types of metrics are defined: based on observations (standardized prediction error with or without simulation and normalized prediction distribution error); based on hyperparameters (with or without simulation); based on the likelihood of the model. All the metrics described above are applied to evaluate a model built from two phase II studies of gliclazide. A real phase I dataset and two datasets simulated with the real dataset design are used as external validation datasets to show and compare how metrics are able to detect and explain potential adequacies or inadequacies of the model. RESULTS: Normalized prediction errors calculated without any approximation, and metrics based on hyperparameters or on objective function have good theoretical properties to be used for external model evaluation and showed satisfactory behaviour in the simulation study. CONCLUSIONS: For external model evaluation, prediction distribution errors are recommended when the aim is to use the model to simulate data. Metrics through hyperparameters should be preferred when the aim is to compare two populations and metrics based on the objective function are useful during the model building process.
Sciences du Vivant/Bio-Informatique, Biostatistique
Informatique/Bio-informatique
Anglais
0724-8741

Articles dans des revues avec comité de lecture
10.1007/s11095-006-9067-5
Pharmaceutical Research / Pharmaceutical Research (Dordrecht)
internationale
09/2006
12/08/2006
23
9
2036-49

model evaluation – metrics – external validation – population pharmacokinetics – posterior predictive check
Algorithms – Artificial Intelligence – Biological Availability – Clinical Trials – Phase II – Computer Simulation – Data Interpretation – Statistical – Gliclazide – Humans – Hypoglycemic Agents – Models – Population – Reproducibility of Results
Liste des fichiers attachés à ce document : 
DOC
Pharmres1612_reviewed110506.doc(2.9 MB)
PDF
Pharmres1612_reviewed110506.pdf(502.4 KB)
XHTML
index.xhtml(102.8 KB)

tous les articles de la base du CCSd...