Initiating and managing therapy - TDM - Modelling

19 important questions on Initiating and managing therapy - TDM - Modelling

What is the pharmacokinetic modeling?

The model is a mechanistic or mathematical description of pharmacokinetics processes. So, two compartments with different clearances and a dose that is given.

What are the formulas that are used in the 1-compartment model?

1) C = A/V
2) Cl = K*V
dA/dt = -Cl*C (dA/dt = the rate of change of amount in the body.
or
dA/dt = -k * A

What is the difference between an analytical solution and a numerical solution?

An analytical solution is an exact solution in contrast to a numerical solution which is more approximate.
  • Higher grades + faster learning
  • Never study anything twice
  • 100% sure, 100% understanding
Discover Study Smart

Which formulas belong to the analytical solution?

C = C0 * e ^-kt
C0 = Dose/ V

What are the formulas that belong to a 2-compartment model?

1) C = A1/V1
2) k10 = CL/V1
3) k12 = Cl12/V1
4) k21 = Cl21/V2

So, what is the numerical solution?

It is a differential equation approximated by assuming that dA/dt is constant over a time interval delta t. So the step size is delta t counting from t = 0. Then for every single step (dt) the dA1/dt  and the dA2/dt is calculated. Then this can be done all over again with a two times as big delta t to see the differences.

What are some (dis)advantages of the numerical method?

advantages 1) applicable to almost any model (e.g. multi-compartments, non-linear kinetics 2) easy to write model code in programmes

disadvantages: 1) approximate solution (improved by advanced integration methods), 2) slower execution

What are integration methods for the numerical solution?

- Euler
- (4th order) Runge-Kutta
- predictor-corrector methods
- Bulirsch-Stoer

What is the critical factor for the numerical solution?

the step size.
An automated step size control can be the solution or an estimate of error (determined by order).

So, what is data modelling?

You have a certain amount of measurements. (n)
You need to determine the best fitting model.

What is nonlinear regression analysis?

non linear indicates that the parameters can change.
Given n observations (with measurement errors)
Given a (PK) model with m (unknown) parameters
To be estimated: best fitting set of parameters. 

You generate also a set of data with e.g. Monte Carlo simulation.

What will be the solution with the nonlinear regression analysis?

1) exact solution by linear regression analysis: a) possible only for models that are linear in the parameters and for data with homoscedastic error b) Therefore, never applicable in PK and PD model.

2) General solution by nonlinear regression analysis: Approximate solution by an iterative algorithm

What are ordinary least squares (OLS)?

You have an objective function. Within this function the deviations (afwijkingen) of  the observed from the predicted are squared. Because there are several parameters (e.g. Cl, V) this has be summed. The best fitting line represents the O with the lowest value.

What are sources of residual error?

- measurement error
- model misspecification
- errors in timing, dosing, sampling etc.

What is the maximum likelihood (ML)?

The best parameter estimates defined as parameter estimates for which likelihood is maximal. This is the maximum likelihood principle. The likelihood (probability) that we would get this answer. The maximum likelihood is then defined as the parameter estimates for which the probability is maximal.

What is the objective function value (OFV)?

the function to be minimized by nonlinear regression analysis is usually called "objective function (value)". So this is the function you want to minimize or maximize.

What is the weighted least squares (WLS)?

This can be interpreted as weighted least squares with weight factor for each observation wi = 1/ variance

What are the criteria for minimization algorithms?

- robustness, that is avoiding convergence to local minima and insensitive for initial estimates
- speed of convergence

What is the difference between global minimum and local minima?

The global minimum is the real minimum. local minima are also lower points (valleys) in the vicinity of the global minimum but not as low as the global minima.

The question on the page originate from the summary of the following study material:

  • A unique study and practice tool
  • Never study anything twice again
  • Get the grades you hope for
  • 100% sure, 100% understanding
Remember faster, study better. Scientifically proven.
Trustpilot Logo