Bayesian Precision in Children: Getting Vancomycin Dosing Correct in This Population

We are continuing our new series, Media Day, where we spotlight individual medical institutions and infectious disease (ID) programs. Today, we spotlight UC San Diego’s Skaggs School of Pharmacy and Pharmaceutical Sciences and the university’s extensive research programs and community-based health programs to serve the local population.

Jennifer Le, PharmD, MAS, FIDSA, FCCP, FCSHP professor of Clinical Pharmacy, Skaggs School of Pharmacy and Pharmaceutical Sciences, UC San Diego, points out when the first vancomycin guideline was published in 2009 it left out the pediatric population. All the data they had was in adults with trough concentrations. This guideline led many pediatric hospitals throughout the US to start adopting this because it was the only guideline that was published and there was no data available in the pediatric population. Consequently, some hospitalized children taking vancomycin saw an increase in acute kidney injury.

Le notes that most cases were transient, but she wanted to shift from trough monitoring to AUC Bayesian monitoring to reduce these serious adverse effects. When she joined UC San Diego, she received an NIH grant to support her hypothesis, and she went to work studying the ongoing issue.

“One of the first studies that I published— it was the largest study of its kind—over 700 children, over 1,000 concentration drug concentrations that we had on vancomycin. This was published in 2013 and showed that look, we need to move away from trough monitoring to AUC Bayesian monitoring, and this will be more precise in kids,” Le said. “And in fact, if we use this, we don't need to increase the dose in most of our patients and actually prevent acute kidney injury in approximately 1 out of every 4 children.”

Her work with pediatric patients has taken her overseas to different countries including Vietnam. She talks about the challenges of limited resources associated with low- and middle-income countries.

“In countries where they're resource limited…the eagerness is there; the passion is there; the knowledge is there…but as a resource-limited country, the biggest challenge is affordability,” Le said.

As an example, she points to using Bayesian estimation, which requires a tool.

“It's kind of integrating AI machine learning in a sense. It goes through this algorithm to figure out what the best dose is for kids. And without that tool, you can't really conduct Bayesian estimation or drug monitoring,” Le said. “So the biggest challenge is that most of these products are usually $10,000 per year, per institution, which is not affordable in these countries.”

She collaborated with professor Hai Le Ba at the Hanoi University of Pharmacy in Vietnam to create a Bayesian tool for vancomycin that is now available for free and accessible in Vietnam. Her hope is to repeat this in other countries.

In the next episode, Le continues the conversation around vancomycin prescribing in resource-limited environments as well as dosing for meropenem in critically ill children.