Bayesian parameter scan: a new method for better estimation of pharmacokinetic parameters in badly designed pharmacokinetic studies

Postgraduate Thesis uoadl:1319126 399 Read counter

Unit:
Κατεύθυνση Κλινική Φαρμακευτική
Library of the School of Science
Deposit date:
2013-07-18
Year:
2013
Author:
Νικοπούλου Δήμητρα
Supervisors info:
Αριστείδης Δοκουμετζίδης Λέκτορας ΕΚΠΑ (επιβλέπων), Παναγιώτης Μαχαίρας Καθηγητής ΕΚΠΑ, Γεωργία Βαλσαμή Αναπλ. Καθηγήτρια ΕΚΠΑ
Original Title:
Μπαεσιανή σάρωση παραμέτρων: Μια νέα μέθοδος για τη βελτιωμένη εκτίμηση φαρμακοκινητικών παραμέτρων σε λάθος σχεδιασμένες φαρμακοκινητικές μελέτες
Languages:
Greek
Translated title:
Bayesian parameter scan: a new method for better estimation of pharmacokinetic parameters in badly designed pharmacokinetic studies
Summary:
In a study, where collected data are restricted or/and collected without proper
design, a
NONMEM parameter estimation may lead to large standard errors, bias and to
overall
unstable performance. In the present study, various sampling scenarios where
simulated
and analyzed, using methods of population analysis. It was also assumed that
the theoretical
drug of interest follows one-compartment pharmacokinetic model with first order
absorption, and for this drug 40 simulated patients were studied. A bootstrap
distribution
was created and nine different percentiles (10% to 90% increasing by 10%) of
this
distribution of clearance were determined. The next step was NONMEM parameter
estimation using Bayesian priors for clearance. The criterion for determining
bias and for
selecting the percentile most likely to be closest to the correct parameter
value was
considered to be the value of the objective function, which should be the
smallest possible.
Thus, for various scenarios a plot was created illustrating the values of the
objective function
versus the respective percentiles. At this plots the minimum of the objective
function was
the linked to the percentile with the most accurate values of the parameters.
However, this
was not the case for extreme scenarios where sampling was planned to be
discontinued very
early, and our criterion was unable to locate the correct parameter values
after NONMEM
execution. Briefly, a method was developed for a more accurate estimation of
population
pharmacokinetic parameters, without need of additional interventions.
Keywords:
Pharmacokinetics, Bias, Bayesian, Parameters, Estimation
Index:
No
Number of index pages:
0
Contains images:
Yes
Number of references:
26
Number of pages:
94
File:
File access is restricted.

document.pdf
4 MB
File access is restricted.