Background
UNN has acquired a license for a commercial CE-marked software that utilizes artificial intelligence (Bayesian statistics and machine learning) for personalized dosing of several medications, including vancomycin.
Vancomycin is an important antibiotic used in infections caused by microbes resistant to other antibiotics. The dosage should be individualized, and newer international guidelines recommend area under the curve (AUC)-based dose adjustment using Bayesian software. The main advantage of this approach is a reduction in kidney damage compared to traditional dose adjustment based on trough concentration.
Many hospitals in the Northern Norway Health Region do not have access to vancomycin analysis in their laboratories. For patients in these hospitals, a service offering personalized dosing using Bayesian software would allow for much earlier optimization of the dosage compared to traditional dose adjustment, mitigating the delay associated with sending blood samples to another laboratory.
Clinical pharmacologists, in collaboration with the infectious diseases department at UNN, have conducted a small quality improvement project involving the implementation of a workflow for personalized, AUC-based vancomycin dosing (see figure).
However, before the method can be implemented, the software needs to be clinically validated using patient data, as it has not been developed for the Norwegian population. There is also a need for workflow optimization, extensive training, and knowledge updates.
Objectives
Main Objective
To offer AUC-based vancomycin dosing as decision support throughout the Northern Norway Health Region.
Sub-Objectives
- To complete the improvement project "Personalized Vancomycin Dosing."
- To validate the estimation of vancomycin AUC from the software.
- To offer AUC-based vancomycin dosing to multiple units at UNN.
- To document the benefits for patients in Northern Norway.
- To offer AUC-based vancomycin dosing to multiple institutions in the Northern Norway Health Region.
Additionally, it may be relevant to expand the project to include the integration of the software into electronic medical records and the development of prediction models based on Norwegian data.
Project Partners
This is a collaborative project involving physicians and pharmacists from UNN, Finnmark Hospital, Helgeland Hospital, Western Norway Health Authority/University of Bergen, and Northern Norway Hospital Pharmacy. Lena Aronsen, a clinical pharmacologist at UNN, leads the project. SPKI provides AI expertise and is intended to have a role in the development of next-generation prediction algorithms based on our own Norwegian data.
