Optimization under uncertainty of melatonin dosing for critically ill patients

Computer-aided modelling and simulation are effective tools to provide guidance in the design of clinical experiments and treatments. Simulations with physiologically-based pharmacokinetic (PBPK) models combine the drug material balances within the body to its real physiological and anatomical features and can be used to optimize drugs dosing and administration timing. We focus on melatonin administration to critically ill patients, a challenging population because of their high inter-individual variability in the pharmacokinetics (due to their heterogeneous and severe conditions). We show how the optimization problem can be suitably formulated to tackle this uncertainty, and compare the results obtained for critically ill patients and healthy individuals. The approach can be easily transferred to any other drug routinely administered in intensive care units whenever a desired pharmacokinetic profile is available.