Development and Evaluation of Multimodel Informed Precision Dosing Tool for Optimizing Vancomycin Therapy in...

The narrow therapeutic window and high pharmacokinetic (PK) variability of vancomycin may lead to trough concentrations outside the usual therapeutic range, requiring dose adjustments. In this study, we aimed to identify suitable pediatric vancomycin models, evaluate their predictive performance, and develop an RShiny-based multimodel informed precision-dosing (multi-MIPD) tool.

A systematic literature search was undertaken to identify pediatric vancomycin PK models, which were graded according to published quality-assessment criteria. Retrospective vancomycin therapeutic drug monitoring data were used to evaluate the performance of high-quality models. Consensus models (mean, median, and weighted) were constructed. In addition, a MIPD tool was developed using the free R package Shiny and validated for both initial dosing and dose adjustment. This tool was evaluated using a prospective dataset.

Nine models demonstrated excellent predictive performance in the retrospective data set (311 concentrations from 192 patients), with root-mean-square error values ranging from 1.00 to 1.97 mg/L and median individual prediction errors from -0.46 to 0.42 mg/L. The multi-MIPD tool incorporating 9 models is presented in the Supplemental Digital Content 1 (see Appendix , http://links.lww.com/TDM/A894 ). The optimal model achieved a median individual prediction errors of 0.02 mg/L, and an root-mean-square error of 0.12 mg/L in the prospective data set (42 concentrations from 35 patients). The mean consensus model significantly improved target area under the curve attainment compared with empirical dosing, with 68.73% versus 36.53% for initial dosing and 55.56% versus 22.22% after dose adjustments.

The multi-MIPD tool provided accurate concentration predictions and, compared with empirical dosing, significantly improved vancomycin target attainment, offering a more effective and individualized dosing strategy for pediatric patients.