AI vs. population pharmacokinetic models to predict the concentrations of antiepileptic drug using TDM records

Population pharmacokinetic (PK) models are commonly used to predict drug concentrations, but artificial intelligence (AI) models offer advantages in handling complex patterns without requiring mathematical assumptions. This study compares the predictive performance of AI and population PK models using therapeutic drug monitoring (TDM) records of four antiepileptic drugs (AEDs): carbamazepine, phenobarbital, phenytoin, and valproic acid. Ten AI models were evaluated against published population PK models using metrics such as Root Mean Squared Error (RMSE). The predictive performance of AI models generally exceeded that of population PK models, with ensemble models such as Adaboost, eXtreme Gradient Boosting, Random Forest showing the lowest RMSE. The most influential covariate was time after last drug administration. Our AI models may serve as clinical decision-support tools to optimize AED dosing, improving therapeutic outcomes while minimizing adverse effects.