poso_* functions. Once the model has been parsed by rxode2() with this package the model$posologyr gives the list needed for poso_* functionsposo_dose_conc(), poso_dose_auc() and poso_inter_cmin() where the returned estimate of the target value to be optimized against was always equal to zero.poso_time_cmin(), poso_dose_conc(), and poso_dose_auc() now explicitly states the consequences of setting tdm to TRUE: which parameters are required, which parameters are ignored, and which parameters behave differently.poso_time_cmin(), poso_dose_conc(), and poso_dose_auc() now return a warning if any of the input parameters are ignored.poso_dose_auc()posologyr() (as requested by CRAN)parent.frame() (as requested by CRAN)poso_estim_map(), poso_estim_sir() and poso_estim_mcmc() can now estimate individual PK profiles for multiple endpoints models (eg. PK-PD, parent-metabolite, blood-CSF…), using a different residual error model for each endpoint.poso_time_cmin(), poso_dose_conc(), poso_dose_auc() and poso_inter_cmin() now allow you to select the end point of interest for which you want to optimise, provided it is defined in the model.vignette("a_priori_dosing") illustrates a priori dose selectionvignette("a_posteriori_dosing") illustrates a posteriori dose selection, using TDM datavignette("auc_based_dosing") shows how to select an optimal dose for a given target AUC using data from TDMvignette("multiple_endpoints") introduces the new multiple endpoints featureposo_time_cmin() can now estimate time needed to reach a selected trough concentration (Cmin) using the data from TDM directlyposo_dose_conc() can now estimate an optimal dose to reach a target concentration following the events from TDMposo_dose_auc() can now estimate an optimal dose to reach a target auc following the events from TDMposologyr() is now an internal function, all exported functions take patient data and a prior model as input parametersposo_estim_map() provides an rxode2 model using MAP-EBE and the input dataset, with interpolation of covariates, to make plotting easierposologyr() functionposo_time_cmin(), poso_dose_auc(), poso_dose_conc(), and poso_inter_cmin() no longer fail for models with IOVposo_estim_sir() estimates the posterior distribution of individual parameters by Sequential Importance Resampling (SIR). It is roughly 25 times faster than poso_estim_mcmc() for 1000 samples.poso_estim_map() allows the estimation of the individual parameters by adaptive MAP forecasting (cf. doi: 10.1007/s11095-020-02908-7) with adapt=TRUE.poso_simu_pop(), poso_estim_map(), and poso_estim_sir() now support models with both inter-individual (IIV) and inter-occasion variability (IOV).MASS:mvrnorm is replaced by mvtnorm::rmvnorm for multivariate normal distributions.poso_estim_map() now uses method=“L-BFGS-B” in optim for better convergence of the algorithm.poso_inter_cmin() now uses method=“L-BFGS-B” in optim for better convergence of the algorithm.poso_dose_conc() is the new name of poso_dose_ctime().poso_time_cmin(), poso_dose_auc(), poso_dose_conc(), and poso_inter_cmin() now work with prior and posterior distributions of ETA, and not only with point estimates (such as the MAP).nocb parameter is added to posologyr(). The interpolation method for time-varying covariates can be either last observation carried forward (locf, the RxODE default), or next observation carried backward (nocb, the NONMEM default).vignette("uncertainty_estimates") is removed.poso_time_cmin(), poso_dose_ctime(), and poso_dose_auc() now work for multiple dose regimen.poso_inter_cmin() allows the optimization of the inter-dose interval for multiple dose regimen.vignette("case_study_vancomycin") illustrates AUC-based optimal dosing, multiple dose regimen, and continuous intravenous infusion.First public release.