AUC-based dosing with validated Bayesian software is recommended as a good approach to guide bedside vancomycin dosing. To compare treatment and vancomycin-associated acute kidney injury (AKI) costs between Bayesian AUC-based dosing and conventional therapeutic drug monitoring (TDM) using steady-state plasma concentrations of vancomycin administered as continuous infusion in hospitalized non-critically ill patients with severe Gram-positive infection. A cost-benefit analysis presented as a return on investment (ROI) analysis from a hospital perspective was conducted using a decision tree model (TDM versus AUC-based dosing) to simulate treatment cost (personnel, serum sampling and drug cost), vancomycin-associated AKI risk and cost up to 14 days. ROI was calculated against AUC-based software cost. One-way and probabilistic sensitivity analyses (respectively OWSA and PSA) were performed to check for robustness. In base case, an overall cost per patient of €621.0 with TDM and €543.6 with AUC-based dosing resulted in a treatment saving of €77.4 per patient when applying AUC-based dosing. This saving against the software cost (€26.9/patient) generated an ROI per patient of €1.9 per invested € in software [€1.9 (95% CI 1.6-2.2) in PSA]. Enrolling 900 AUC-based dosed patients annually translated to a net saving of €45 469. Software break-even was reached after 313 patients. In OWSA, a higher AKI risk with TDM strongly contributed to a positive ROI. AUC-based dosing appeared a cost-saving strategy compared with conventional TDM when applying base-case settings of vancomycin-associated AKI risk, treatment and AKI costs.
Read full abstract