AnestAssist PK/PD Testing


Testing and Results

The results of AnestAssist PK/PD model testing are given here so that users and potential users can see and judge how the PK/PD models perform, and how they were verified.

As part of normal quality control Palma Healthcare runs an automated series of tests of AnestAssist's PK/PD Module. These tests involve presenting the following inputs to the models

  • Bolus doses of varying amounts
  • Infusion doses of varying rates
  • Inspired doses of varying concentrations
  • Combinations of propofol and opioids (for interaction testing)
  • Combinations of sevoflurane and remifentanil (for interaction testing)
  • Patient ages and weights varied for models which use these as parameters (covariates)

The models are run and the following output is collected at 1 second intervals

  • Plasma concentration (IV)
  • Inspired and Expired Concentrations (Inhaled)
  • Effect site concentration
  • Probability of effect - drug w/o interactions
  • Probability of effect - interaction with other drugs

The output is plotted and visually checked against expected values. The results of our AnestAssist version 1.0 and 1.1.1 (note release 1.1 had an error in the sevoflurane PK model) release test runs are given below.

Quick Links:


Propofol Model (Schnider) Test Results

Below are the results of testing AnestAssist's implementation of the propofol model of Schnider et al (see reference [1] in AnestAssist Scientific Basis).

Propofol PropofolPropofol
Propofol
Propofol
Propofol
Propofol
Propofol
Propofol

Back to top


Propofol Model (Marsh) Test Results

Below are the results of testing AnestAssist's implementation of the propofol model of Marsh et al (see reference [2] in AnestAssist Scientific Basis).

Propofol PropofolPropofolPropofolPropofolPropofol

Back to top

Remifentanil Model Tests

Below are the results of testing AnestAssist's implementation of the Remifentanil model of Minto et al (see reference [3] in AnestAssist Scientific Basis).

Below are the results of testing AnestAssist's implementation of the Alfentanil model of Scott and Stanski (see reference [4] in AnestAssist Scientific Basis).
Remifentanil
Remifentanil
Remifentanil
Remifentanil
Remifentanil
Remifentanil

Back to top.

Fentanyl Model Tests

Below are the results of testing AnestAssist's implementation of the Fentanyl model of Scott and Stanski (see reference [4] in AnestAssist Scientific Basis).

Fentanyl
Fentanyl
Fentanyl
Fentanyl

Back to top.

Alfentanil Model Test Results

Below are the results of testing AnestAssist's implementation of the Alfentanil model of Scott and Stanski (see reference [4] in AnestAssist Scientific Basis).

Alfentanil
Alfentanil
Alfentanil
Alfentanil

Back to top

Sufentanil Model Tests

Below are the results of testing AnestAssist's implementation of the Sufentanil model of Gepts et al (see reference [5] in AnestAssist Scientific Basis).

Sufentanil SufentanilSufentanilSufentanil

Back to top.

Inhaled Agent Model Tests

Below are the results of testing AnestAssist's implementation of the inhaled agent model of Lerou et al (see reference [10] in AnestAssist Scientific Basis). Probability of effect estimates, including interactions, are from the model of Manyam et al, reference [11]. Note that these results are from a test run using 5 sample administrations of sevoflurane, starting at time = 1 minute and constant for 60 minutes as shown in the first figure below. The results in subsequent figures are from this input sequence.

Sevo Model InputSevo expired concsevo brain concsevo single prob effect

Back to top.