Identifying the clinically relevant covariates that drive inter-individual variability is fundamental to precision dosing. However, standard approaches such as stepwise covariate modeling are often limited by the laborious process of evaluating numerous covariate-parameter relationships and primarily focus on statistical significance rather than the clinical relevance of covariate effects on dosing requirements. To address these limitations, we developed the Systematic Covariate Observational Uncovering Technique (SCOUT), an approach that streamlines covariate exploration by shifting the focus from influences on multiple pharmacokinetic (PK) or pharmacodynamic (PD) parameters to those on individual optimal dose, which is a clinically essential metric. This shift simplifies the analysis from a complex many-to-many relationship to a one-to-many evaluation, facilitating rapid identification of factors that substantially influence therapeutic effects. We validated SCOUT across three distinct scenarios: simulation-based verification, real-world amikacin PK data analysis, and complex eribulin PK/PD modeling. Estimated individual optimal doses matched theoretical values with minimal bias. Real-world applications successfully revealed established covariates such as weight and renal function for amikacin and baseline neutrophil count for eribulin. Furthermore, SCOUT provided objective evidence for dosing-interval optimization. By serving as an efficient hypothesis-generating approach, SCOUT enables pharmacometricians to prioritize clinically impactful factors and rationally narrow the search space in formal model building. This approach functions as a "nautical chart" in navigation toward optimal dosing, ultimately supporting more rational and efficient clinical decision-making in drug development and precision medicine.
Ideno et al. (Tue,) studied this question.