Ninety patients in three arms of thirty: intravenous only, subcutaneous only, and an intravenous loading dose followed by subcutaneous maintenance. The third arm is the reason the dataset exists — a patient who receives both routes is what a mixed model is for, and no public dataset in the package survey carries an administration column at all.
Format
A data frame with 1800 rows and 11 columns:
- ID
Patient identifier, 1 to 90.
- TIME
Recorded time in days, drifting off the plan the way visits do.
- NTIME
Nominal (protocol) time in days.
- DV
Concentration in mg/L;
NAon dose records.- AMT
Dose in mg, 100 at each of days 0, 7 and 14.
- EVID
1 on a dose record, 0 on an observation.
- CMT
1 for a dose entering the depot, 2 for central and for every observation.
- ADM
Administration identifier: 1 intravenous, 2 subcutaneous. Declare it with
pmx_roles(adm = "ADM", routes = c("1" = "iv", "2" = "extravascular")).- CENS
1 where the value is at the quantification limit.
- ARM
"IV only","SC only"or"IV then SC".- WT
Baseline weight in kg.
Details
Simulated from a one-compartment model with first-order absorption, so the truth is known: clearance 2 L/day, volume 10 L, absorption 0.5 /day, and bioavailability 0.7, with clearance and volume scaled allometrically on weight. Bioavailability is identifiable here precisely because the same study doses both ways.
Three weekly doses, sampled richly after the first and last and at troughs in
between. Concentrations below 0.05 mg/L are reported at the limit with CENS
set, which is the convention nlmixr2 and Monolix share.