nca_control

Control options for non-compartmental analysis (NCA)

Provides automated methods for generating initial parameter estimates
in population pharmacokinetic modeling. The pipeline integrates adaptive single-point
methods, naive pooled graphic approaches, noncompartmental analysis methods, and parameter
sweeping across pharmacokinetic models. It estimates residual unexplained variability
using either data-driven or fixed-fraction approaches and assigns pragmatic initial values
for inter-individual variability. These strategies are designed to improve model
robustness and convergence in 'nlmixr2' workflows. For more details see
Huang Z, Fidler M, Lan M, Cheng IL, Kloprogge F, Standing JF (2025)
<doi:10.1007/s10928-025-10000-z>.

Zhonghui Huang

nlmixr2autoinit

Automatic Generation of Initial Estimates for Population
Pharmacokinetic Modeling

Joseph Standing

Matthew Fidler

Frank Kloprogge

nca_control function

<dl><dt>trapezoidal.rule</dt>
<dd>Character. Method for trapezoidal AUC integration:<ul>
<li><code>"linear"</code> - Linear trapezoidal rule (default)</li>
<li><code>"linear_up_log_down"</code> - Linear-up / log-down rule</li>
</ul></dd>
<dt>duration</dt>
<dd>Numeric. Optional. Duration of the observation window (same
units as time). Used to restrict the integration or define the evaluation
range.</dd>
<dt>nlastpoints</dt>
<dd>Integer. Number of terminal points for half-life
regression (default = 3).</dd>
<dt>slope.method</dt>
<dd>Character. Method for estimating the terminal slope
(lambdaz). Options are:<ul>
<li>"bestfit": Performs automated terminal phase selection based on adjusted
R-squared. If no acceptable segment is found, lambdaz is returned as NA.</li>
<li>"bestfitforce": First attempts the "bestfit" method. If no valid lambdaz
is obtained, the function applies a fallback log-linear regression using
progressively fewer terminal points to force estimation. This is the default.</li>
</ul></dd></dl>

Arguments

Author

Control options for non-compartmental analysis — nca_control

<dl>

<dt>trapezoidal.rule</dt>
<dd>Character. Method for trapezoidal AUC integration:<ul>
<li><code>"linear"</code> - Linear trapezoidal rule (default)</li>
<li><code>"linear_up_log_down"</code> - Linear-up / log-down rule</li>
</ul></dd>


<dt>duration</dt>
<dd>Numeric. Optional. Duration of the observation window (same
units as time). Used to restrict the integration or define the evaluation
range.</dd>


<dt>nlastpoints</dt>
<dd>Integer. Number of terminal points for half-life
regression (default = 3).</dd>


<dt>slope.method</dt>
<dd>Character. Method for estimating the terminal slope
(lambdaz). Options are:<ul>
<li>"bestfit": Performs automated terminal phase selection based on adjusted
R-squared. If no acceptable segment is found, lambdaz is returned as NA.</li>
<li>"bestfitforce": First attempts the "bestfit" method. If no valid lambdaz
is obtained, the function applies a fallback log-linear regression using
progressively fewer terminal points to force estimation. This is the default.</li>
</ul></dd>

</dl>

nca_control: Control options for non-compartmental analysis

Description

Usage

Value

Arguments

Author

Examples