Get a quote

Get a free quote now.

Thank you! Your message has been received!
Oops! Something went wrong while submitting the form.

How Bayesian Modeling for Vanco Dosing Works

Megan N. Freeland, PharmD
August 27, 2020

The new vanco dosing guidelines have brought Bayesian dosing to the forefront of conversations for many institutional pharmacy departments. But perhaps you’re still not 100% clear about exactly how Bayesian dosing works. You might be wondering:

  • How does Bayesian dosing help estimate area under the curve (AUC)? 
  • What makes Bayesian dosing so much more accurate than first-order pharmacokinetic (PK) equations? 
  • How do I actually implement Bayesian dosing in practice?
  • Is there a Bayesian vancomycin calculator I can use?

Let’s discuss the answers to those questions in a straightforward, easy-to-understand way.

Bayesian Dosing, Simplified

You might hear different terms used to describe this concept: Bayesian dosing, Bayesian forecasting, Bayesian modeling, and model-informed precision dosing (MIPD), just to name a few. Each of these terms refers to the same set of principles that helps clinical pharmacists take better care of their patients through dose optimization. 

In the most basic sense, here’s how Bayesian modeling works:

  1. You feed the model with patient-specific information. 
  2. Bayesian forecasting processes that information via a PK model.
  3. You receive an output of more patient-specific information that guides vanco dosing. 

Now let’s flesh that out a bit to understand how this supports vancomycin dosing.

  1. You input patient-specific information. Specifically, you enter individual patient data like clinical characteristics, previous vanco dosing history, drug concentrations, and lab results, as well as a potential vanco dosing regimen.
  2. Bayesian forecasting processes that information via a PK model. Bayesian models, which are based upon existing populational PK data, take your data input and use it to estimate a patient’s individual PK profile—e.g., absorption, metabolism, and clearance—for vancomycin.
  3. You receive an output of more patient-specific information that guides vanco dosing. The information you get out includes a prediction of the patient-specific AUC exposure corresponding to any dosing regimen. These predictions enable identification and selection of a patient-specific regimen predicted to meet AUC targets.

As you continue to input patient-specific data over a patient’s treatment course, the AUC prediction becomes more and more accurate.

The Science Behind Bayesian Dosing

The quality of Bayesian dosing software is highly dependent upon the models used to generate the dosing predictions. As real-world data evolves, these models and algorithms can be updated to make the Bayesian predictions more accurate and remain current. 

Our clinical pharmacists—the experts behind InsightRX Nova, a Bayesian dosing software solution—employ models that have undergone rigorous scientific verification and external validation. We also continuously review, update, and validate our models to ensure they represent the best that science has to offer.

Free, online Bayesian vancomycin calculators are unlikely to provide the same level of high-quality modeling, may not offer the ability to adjust the model’s predictions based on a patient’s historical data, and may not have the ability to account for changing renal function and other time-varying covariates.

What about First-Order Pharmacokinetics for Vancomycin?

Throughout this discussion of Bayesian dosing, you might be wondering where that leaves the first-order PK equations you may have traditionally used for vanco dosing. The new 2020 vancomycin consensus guideline team actually states a preference for Bayesian dosing over first-order PK equations, considering that Bayesian dosing is more accurate and more efficient than using first-order PK equations.

Learn more about the benefits of Bayesian dosing over first-order PK equations.

How to Use Bayesian Dosing for Vancomycin in Practice

Implementing Bayesian dosing in practice requires access to Bayesian dosing software. Here are a few suggestions to guide your search:

  • Consider what types of patients you serve in your institution—e.g., adults, pediatrics, neonates, obese, critically ill—and whether the software accommodates those special populations.
  • Balance your budgetary needs with the value provided by the Bayesian dosing software solution. 
  • Guarantee that your software will grow with you through continuous, automatic updates.

InsightRX Nova offers dosing models for a variety of different patient populations and future-proofs our software through constantly-evolving real-world data and human-assisted AI.

Recent news