Sparse estimators for the AUC and AUMC to the last measured concentration

These are the FUN_sparse of auclast, aumclast, aucall, and aumcall: with sparse PK, pk.nca() estimates those parameters from the pooled individual samples with the Bailer point estimate and the Nedelman-Jia/Holder standard error rather than integrating the arithmetic-mean profile. They wrap pk.calc.sparse_auc() and pk.calc.sparse_aumc() (with auc.type = "AUClast" or "AUCall"), reporting the results under the unified parameter names. AUCall adds the triangle from tlast to the next time, whose mean is zero.

Usage

pk.calc.auclast_sparse(conc, time, subject, ..., options = list())

pk.calc.aumclast_sparse(conc, time, subject, ..., options = list())

pk.calc.aucall_sparse(conc, time, subject, ..., options = list())

pk.calc.aumcall_sparse(conc, time, subject, ..., options = list())

Arguments

conc: Measured concentrations
time: Time of the measurement of the concentrations
subject: Subject identifiers (may be any class; may not be null). A missing subject marks an imputed concentration (such as the zero that the start_conc0 imputation adds at the start of an interval): a time where every subject is missing has a known concentration, which enters sparse estimates but not their variance.
...: For functions other than pk.calc.auxc, these values are passed to pk.calc.auxc
options: List of changes to the default PKNCA options (see PKNCA.options())

Value

A data.frame with the point estimate, its standard error, and the degrees of freedom, named for the parameter (for example, auclast, auclast_se, and auclast_df)

Details

The sparse variance theory is defined for the linear trapezoidal rule only, so these ignore the auc.method option; pk.nca() says so when the option is set to anything else.

Functions

pk.calc.aumclast_sparse(): Sparse AUMClast
pk.calc.aucall_sparse(): Sparse AUCall
pk.calc.aumcall_sparse(): Sparse AUMCall