Comparison study of model evaluation methods: normalized prediction distribution errors vs. visual predictive check / 药学学报
Acta Pharmaceutica Sinica
;
(12): 1123-31, 2011.
Article
in Chinese
| WPRIM
| ID: wpr-414983
ABSTRACT
The objective of this study is to compare the normalized prediction distribution errors (NPDE) and the visual predictive check (VPC) on model evaluation under different study designs. In this study, simulation method was utilized to investigate the capability of NPDE and VPC to evaluate the models. Data from the false models were generated by biased parameter typical value or inaccurate parameter inter-individual variability after single or multiple doses with the same sampling time or multiple doses with varied sampling time, respectively. The results showed that there was no clear statistic test for VPC and it was difficult to make sense of VPC under the multiple doses with varied sampling time. However, there were corresponding statistic tests for NPDE and the factor of study design did not affect NPDE significantly. It suggested that the clinical data and model which VPC was not fit for could be evaluated by NPDE.
Full text:
Available
Index:
WPRIM (Western Pacific)
Type of study:
Prognostic study
Language:
Chinese
Journal:
Acta Pharmaceutica Sinica
Year:
2011
Type:
Article
Similar
MEDLINE
...
LILACS
LIS