JMIR Publications

ABSTRACT

Vancomycin is an intravenous antibiotic with a narrow therapeutic window where concentrations must be monitored for safety and efficacy. Traditional methods of dosing vancomycin use a weight-based starting dose with interval based on renal function. Dose adjustments are often based on trough concentrations measurements and heuristics or nomograms. It has been suggested that vancomycin dose adjustments using Bayesian methods is optimal. Unfortunately, many of these Bayesian tools are costly. In the hopes of democratizing Bayesian vancomycin dosing, we created VancoCalc, a free website, as a proof of concept vancomycin dosing tool which leverages Bayesian inference. It uses individual concentrations to estimate pharmacokinetic parameters and suggest optimal dosing options based on these concentrations and estimated pharmacokinetic parameters. To gain confidence in the Bayesian process and evaluate the tool, it was used to predict vancomycin concentrations using retrospective clinical data from 52 patients. Our hope is to increase awareness of this tool and further improve on vancomycin dosing with this freely available instrument, which may have beneficial implications for therapeutic monitoring and improving patient outcomes.