mapbayr

mapbayr performs maximum a posteriori Bayesian estimation (MAP-BE) of pharmacokinetic (PK) parameters to support model‑informed precision dosing (MIPD).

Key Features:

• MAP-BE estimation: Implements maximum a posteriori Bayesian estimation of individual PK parameters.
• mrgsolve integration: Operates with population PK models coded in mrgsolve for simulations and likelihood evaluation.
• Supported model elements: Handles first-order and zero-order absorption, lag time, time‑varying covariates, Michaelis‑Menten elimination, parent drug and metabolite dynamics, and models with small or large inter‑individual variability (IIV).
• Residual error structures: Supports combined and exponential residual error models.
• Simulation-based validation: Validated on simulated "test" models with 4,000 PK profiles and direct comparisons to NONMEM reporting a 98% concordance rate.
• Complex model assessment: Reports discrepancies for dose‑related parameters and large IIV, with objective function value analyses indicating cases where mapbayr may outperform NONMEM.
• Published-model validation: Evaluated on seven previously published PK models with near‑100% concordance on PK outcomes relevant to MIPD.
• Data handling functions: Includes functions for data formatting and reporting.

Scientific Applications:

• Model‑Informed Precision Dosing (MIPD): Provides individual PK parameter estimates to inform dosing decisions in clinical pharmacology and personalized medicine.
• PK model evaluation and comparison: Enables performance assessment and benchmarking of population PK models through simulation and comparison with NONMEM.

Methodology:

Performs MAP-BE on mrgsolve-coded population PK models, uses simulation of PK profiles (4,000 in test cases), compares estimates to NONMEM including objective function value analyses, and validates results on seven published PK models; includes functions for data formatting and reporting.