Errors in time in pharmacokinetic studies

Abstract

Pharmacokinetic models with errors in the time variable were developed and explored by simulation in NONMEM. The precision of estimated parameters decreased with increasing error magnitude. The usual asymptotic standard error estimates were biased low by 10-50% Accuracy and precision were not greatly diminished by ignoring errors in time when fitting models.

Keywords:

• Errors in variables
• Measurement-error models
• Population pharmacokinetics
• Maximum likelihood
• NONMEM