Get a free quote now.
How Will Bayesian-Guided AUC Monitoring Work in Practice?
- Bayesian-guided AUC monitoring is the preferred method for optimal management of vancomycin dosing, as it adapts to varying regimens and changes in physiology during courses of therapy
- Two vancomycin serum concentration levels (peak and trough) are preferred for the most accurate monitoring, but single level monitoring may be implemented as well
- Using this method, vancomycin serum concentration levels may be taken prior to steady-state
- In addition to clinical decision support, Bayesian software can provide technology for productive and efficient workflow
Bayesian-guided AUC monitoring refers to using model-informed precision-dosing (MIPD) software to estimate and predict more accurate drug exposures over a period of time. It is based on Bayes’ Theorem, which can be used to update certain measures of interest based on prior knowledge and newly collected information. For therapeutic monitoring of vancomycin, this means interpreting individual patient levels while also applying validated pharmacokinetic (PK) models to account for physiological and dosing regimen changes during therapy.
The PK models are derived from measuring vancomycin serum concentrations during intravenous administration of vancomycin to a particular cohort of patients.This model and the associated distributions of PK parameters make up the “Bayesian prior,” which is used to represent that particular cohort of patients in future PK simulations.
In contrast to AUC monitoring with first-order PK equations, Bayesian-guided dosing can provide a more complete clinical picture of patients beyond the moment when a vancomycin level or serum creatinine was drawn and does not require waiting for steady-state achievement. Additionally, a Bayesian approach also provides measures of model misfit and model uncertainty, which can provide further useful information to the clinician in designing the optimal future dosing regimen.
Bayesian software-guided vancomycin dosing can be approached using two serum vancomycin concentration levels per analysis or using one level as shown in a study by Neely et al. (2014). Being able to obtain one level is advantageous in that only slight changes in pharmacy, nursing, and provider practice would be needed to adopt the Bayesian approach. However, not all populations were well represented in the single-level analysis (e.g. obese, critically ill, pediatrics, and patients with renal instability). Accordingly, obtaining two levels is still the preferred approach as stated in the upcoming vancomycin guidelines. Furthermore, although Bayesian software can predict AUC using levels taken at most time points, in the absence of data surrounding other timepoints, peak and trough levels are still considered the standard.
Implementing Bayesian-guided AUC monitoring requires an institution to decide which software best suits its needs. There are many available options with varying user interfaces, features, and companion applications. The selection of the tool will depend on the needs , wants, and goals of the institution. Once a software tool is chosen, the institution must budget to pay the third-party vendor. Additionally, clinical staff will need to dedicate time for training in proper use of the software to ensure quality of patient care.
If teams decide to calculate AUC via two levels (peak and trough), education and training for all relevant clinical staff will be necessary as discussed in our recent blog post. The benefit of Bayesian-model dosing, however, is that exact timing of levels will need not be as strict. If doses are given or levels are taken outside of scheduled times, the information will still be useful for Bayesian interpretations.
If teams opt for single-level monitoring, the transition to AUC monitoring may be a little easier. Clinical staff will benefit from understanding that levels can be drawn without waiting for steady-state to be achieved. This practice change may increase both efficiency in workflow and flexibility in laboratory interpretation for clinical staff.
Finally, Bayesian software can be integrated into electronic health record systems, allowing for even faster data entry and interpretation. If this path is chosen, information technology support will likely be necessary at the beginning of implementation. Once integrated, data such as prior doses, serum vancomycin concentration levels, serum creatinine, and patient demographics can be automatically imported into the software and incorporated into PK models.
Overall, Bayesian-guided AUC monitoring provides a solution for adherence to the upcoming vancomycin therapeutic monitoring guidelines, and healthcare institutions should be aware of the many practice implementation considerations.
If you found this post helpful, check out previous and upcoming posts.
● How Will AUC Monitoring With First-Order PK Equations Work in Practice?
● Deciding Between Vancomycin AUC Monitoring Approaches
For more information on the new guidelines and related topics, reach out to us today.