Prediction of vancomycin dose on high-dimensional data using machine learning techniques

Abstract

Objectives: Despite therapeutic vancomycin is regularly monitored, its dose requirements vary considerably between individuals. Various innovative vancomycin dosing strategies have been developed for dose optimization; however, the utilization of individual factors and extensibility is insufficient. We aimed to develop an optimal dosing algorithm for vancomycin based on the high-dimensional data using the proposed variable engineering and machine-learning methods. Methods: This study proposed a variable engineering process that automatically generates secondorder variable interactions. We performed an initial examination of independent variables and interactive variables using eXtreme Gradient Boosting. The vancomycin dose prediction model was established based on the derived variables. Results: Based on the evaluation of the model performance in the validation cohort, our algorithm accounted for 67.5% of variations in the vancomycin doses. Subgroup analysis showed better performance in patients with medium and high body weight (with the ideal predictive percentage of 72.7% and 73.7%), and low and medium levels of serum creatinine (with the ideal predictive percentage of 77.8% and 73.1%) than in other groups. Conclusion:The new vancomycin dose prediction model is potentially useful for patients whose population profiles are similar to those of our patients and yielded desired reference of clinical indicators with specific breakpoints.

AI Models to Assist Vancomycin Dosage Titration

“…This is a common problem in developing dose recommender of vancomycin injection. Previous studies either used only the data of patients whose corresponding vancomycin trough levels fall into the desired therapeutic range ( Huang et al, 2021 ), or set the “correct” injection dose proportionally to patients’ weight ( Imai et al, 2020 ). These studies either neglected the potential to learn from failures, that is, injections with undesired therapeutic effect, or oversimplified inter-individual variability of injection dose.…”

“…In clinical practice, various approaches have been used to guide clinicians in vancomycin dosing such as dosing nomograms and Bayesian estimation methods. In a study by Huang et al, a vancomycin dose prediction model was established using eXtreme Gradient Boosting (XGBoost) for feature selection and model construction ( Huang et al, 2021 ). Their model did not differentiate between initial dose and subsequent dose predictions.…”

“…Owing to the large heterogeneity of dosage requirements between individual patients, machine learning methods are able to process large amounts of patient data and translate into useful clinical recommendations ( Goh et al, 2020 ). Although a couple of studies published promising results very recently ( Imai et al, 2020 ; Huang et al, 2021 ), AI-assisted vancomycin dosage titration is not widely used as a regular practice in hospitals. Our models contribute to this area in three aspects: First, the models give direct suggestions of appropriate daily dose not only to initial injection but also to subsequent injections.…”

Artificial Intelligence in Infection Management in the ICU

“…Introduction of machine learning into pharmacometrics is still in its infancy although the potential of a partnership is increasingly being recognized [ 27 ]. At the same time, machine learning research is ongoing to improve antimicrobial dosing, as is illustrated by the vancomycin dose prediction model using XGBoosting developed by Huang et al [ 28 ]. However, more research is needed here as the error is rather high to have potential in clinical practice.…”

The future of antimicrobial dosing in the ICU: an opportunity for data science

“…As a vast amount of data is available from TDM programmes for antimicrobials with a narrow therapeutic index, developing dosing guidance models for these antimicrobials would be a convenient first step. Recently, an attempt was made by Huang and colleagues to develop a purely ML-based vancomycin dose prediction model using eXtreme Gradient Boosting [10].…”