Publications and Presentations

This study examines how age and body mass index (BMI) influence the selection of pharmacokinetic models for dosing vancomycin, a commonly used antibiotic, using data from a large, multi-site retrospective study. It found that the accuracy of vancomycin dosing predictions varies significantly across different age and BMI groups, with certain models performing better for specific populations. The research highlights the need for personalized model selection in medication management to improve treatment outcomes, especially noting the necessity for better models for younger adults.

This study compares parametric and nonparametric statistical methods for developing precision dosing models using vancomycin in patients with class 3 obesity. It found minimal differences in model structure and predictive error between the two methods. However, the parametric model showed superior performance in external validation, suggesting that custom models could enhance drug management in specific patient populations.

This study focuses on optimizing posaconazole dosing for fungal infections through model-informed precision dosing (MIPD), aiming to personalize treatment and improve outcomes. By evaluating seven pharmacokinetic models using data from 143 patients, the researchers identified a model that accurately predicts posaconazole levels with minimal bias and imprecision, even with limited patient samples. This advancement marks a significant step towards personalized antifungal therapy, potentially enhancing the effectiveness of posaconazole treatments in clinical settings.

This study compares two methods, non-compartmental analysis (NCA) and maximum a posteriori Bayesian (MAP) approaches, for optimizing busulfan dosing, a medication requiring precise dosing due to its narrow therapeutic index and high variability among individuals. The research, using retrospective patient data and simulated treatment courses, found that while both methods provided similar estimates in real-world data, MAP approaches resulted in higher accuracy in achieving target drug concentrations in simulations. The findings suggest that MAP methods may offer superior precision in dosing adjustments for drugs like busulfan, potentially improving patient outcomes.

This study evaluated the effectiveness of eight different population pharmacokinetic models for predicting the appropriate dose of infliximab, a medication used in treating children with inflammatory bowel disease, in a Dutch pediatric cohort. The Fasanmade model was identified as the most suitable for clinical use, demonstrating good predictive performance in maintaining target infliximab serum concentrations. The findings suggest that model-informed precision dosing could be beneficial in managing infliximab therapy in children with inflammatory bowel disease, potentially enhancing treatment efficacy.

This study introduces a novel approach combining pharmacokinetic/pharmacodynamic (PKPD) models with machine learning (ML) to improve predictions of chemotherapy-induced neutropenia, a common and severe side effect of cancer treatment. By integrating individual patient data from electronic health records with advanced modeling techniques, the researchers were able to enhance the accuracy of neutropenia risk predictions in patients undergoing chemotherapy for lymphoma, breast, or thoracic cancer. The hybrid model showed significant improvements in predicting severe neutropenia, suggesting a promising method for personalizing chemotherapy dosing and potentially reducing treatment-related complications.

This pilot study explores the potential of early at-home measurement of adalimumab levels to improve dosing accuracy for patients with rheumatic or inflammatory bowel diseases. By using population pharmacokinetic model-based Bayesian forecasting, researchers were able to predict steady-state adalimumab concentrations after the initial dose with a high degree of accuracy. The findings suggest that early measurement can guide precision dosing, potentially enhancing treatment efficacy and reducing the risk of underdosing.

This article discusses the role of Clinical Decision Support Software (CDSS) in facilitating precision dosing, which aims to customize medication doses for individual patients to enhance treatment outcomes and minimize toxicity. It highlights the challenge of underrepresentation of diverse patient populations in clinical trials and data sources that inform CDDS, potentially limiting the effectiveness of precision dosing for these groups. The authors propose methods to improve the inclusivity of CDSS tools, including better study designs and the use of biological indicators over demographic ones, to ensure more equitable healthcare outcomes.

This study evaluates the effectiveness of various pharmacokinetic models for dosing vancomycin in neonates and infants, considering the challenges posed by their rapid growth and changing organ function. It was found that two models were particularly effective in predicting drug concentrations, highlighting the importance of accurate patient data, such as serum creatinine levels and gestational age, for individualized dosing. The research underscores the potential of model-informed precision dosing (MIPD) to improve medication dosing in this vulnerable population by accounting for individual variability.

This study focuses on refining a population pharmacokinetic model for fludarabine, a drug used in pediatric hematopoietic cell transplantation, to achieve more precise dosing. By analyzing fludarabine plasma concentrations in a multicenter study, the researchers were able to improve the model by incorporating factors like fat-free mass and a maturation function for drug clearance. The refined model promises better dosing recommendations, especially for children under 2 years and those with a higher body mass index, potentially making therapeutic drug monitoring unnecessary for achieving targeted drug exposure.

This study explores the effectiveness of model-based follow-up dosing for tacrolimus, a critical medication for kidney transplant recipients, compared to standard therapeutic drug monitoring (TDM) methods. By incorporating patient-specific characteristics and pharmacological data, the research aimed to achieve more accurate drug exposure levels. The findings suggest that combining an initial algorithm-based dose with model-based follow-up dosing could potentially improve the precision of tacrolimus dosing in the early post-transplant phase, although the improvement over initial dosing methods appears to be modest. This research highlights the potential benefits of personalized medicine approaches in managing tacrolimus therapy for kidney transplant patients, aiming for better drug exposure control to minimize risks of under- or over-dosing.

This study evaluates the effectiveness of six neonatal gentamicin pharmacokinetic models using data from routine clinical care across nine U.S. sites, involving 475 patients. It found that four models accurately predicted gentamicin concentrations, which could help in making more precise dosing decisions for neonates. The research highlights the potential of model-informed precision dosing (MIPD) to improve treatment outcomes by refining these models through a continuous learning process, although using informative Bayesian priors did not consistently enhance model performance.

This study aimed to develop a population pharmacokinetic (PopPK) model for administering tacrolimus, a drug with a narrow therapeutic window, to pediatric and young adult patients undergoing hematopoietic cell transplantation. By analyzing data from 111 patients, the researchers created a model that accounts for individual differences in drug clearance and distribution, improving the precision of dosing. The model, which shows moderate bias in predictions, has been implemented in a Bayesian dosing tool to facilitate personalized medication management at the point of care, aiming to enhance treatment efficacy and reduce toxicity.

A novel approach combining machine learning (ML) with pharmacokinetic (PK) modeling was shown to improve the accuracy of personalized dosing predictions for patients treated with vancomycin, compared to traditional methods. This hybrid method was particularly effective when the standard PK models did not adequately describe the patient, by adjusting the influence of prior population data on the prediction. The research highlights the potential of integrating ML with PK models to enhance the precision of dosing regimens, potentially leading to better patient outcomes.

This study introduces a new model-based method for estimating renal function through iohexol plasma clearance, aimed at improving the accuracy of renal function assessments in the context of renal transplantation. By developing a population pharmacokinetic model and limited sampling schedules, the researchers offer a practical tool for measuring glomerular filtration rate (GFR) that could replace less accurate or feasible conventional methods. The model, validated both internally and externally, shows high predictive performance and is made accessible online for clinical use, although further validation is suggested for patients with GFRs below 30 mL/min.

This study aimed to develop a pharmacokinetic model for Envarsus, a formulation of tacrolimus used in liver transplant patients, to improve drug monitoring and dosage accuracy. Despite exploring various genetic factors, no significant covariates were found that could help in optimizing the initial dose for patients. The research concluded that a three-point limited sampling strategy could accurately predict drug exposure, suggesting that genetic factors like CYP3A status do not significantly impact dose optimization for Envarsus.

This study investigates the optimal exposure of busulfan, a chemotherapy drug, for patients with non-malignant disorders undergoing bone marrow transplantation to achieve stable myeloid chimerism, a condition where both donor and recipient blood cells coexist, leading to successful disease correction. The research, conducted at the University of California San Francisco, found that a higher busulfan exposure (cumulative area under the curve ≥70 mg·h/L) was associated with better outcomes and minimal toxicity, suggesting a target busulfan exposure range of 70 to 80 mg·h/L for these patients. The findings highlight the importance of precise busulfan dosing to improve transplantation success rates while minimizing side effects.

This study evaluates the impact of using age-adjusted serum creatinine levels in vancomycin pharmacokinetic models on the accuracy of drug dosing in elderly patients. The findings suggest that incorporating age-adjusted serum creatinine decreases the predictive performance of these models in elderly populations. This indicates a potential need for alternative methods to improve vancomycin dosing accuracy in older patients.

This study explores the concept of model-informed precision dosing (MIPD) for pediatric patients receiving vancomycin, emphasizing the importance of continuous learning to improve dosing accuracy. By analyzing data from 273 pediatric patients and employing continuous learning models, the research demonstrates a potential reduction in prediction error by 2-13% compared to existing models. The findings suggest that using a continuous learning approach in MIPD can enhance drug efficacy and safety by better tailoring dosing to individual patient needs, particularly in pediatric intensive care settings.

Researchers developed a mathematical model to explore shorter, more efficient drug therapy options for treating active pulmonary tuberculosis (TB), aiming to reduce treatment duration and the risk of bacterial resistance. This model, which simulates the human immune response, drug interactions, and bacterial resistance, was validated against clinical trial data and is available as open-source software. It offers a promising tool for analyzing new TB treatments and improving patient adherence by potentially shortening therapy durations.

This review article discusses the pharmacokinetic and pharmacodynamic properties of hydroxychloroquine (HCQ) in the context of treating COVID-19, highlighting the lack of randomized trial data and the need for further research to understand its efficacy and safety against the virus. Despite HCQ's potential in vitro activity against SARS-CoV-2, the causative agent of COVID-19, there is a significant gap in knowledge regarding the optimal dosing and its overall impact on patient outcomes. The authors call for more studies to fill these gaps and help guide effective and safe use of HCQ in COVID-19 treatment.

This study evaluates the effectiveness of a population pharmacokinetic model for precision dosing of everolimus in adult renal transplant recipients, aiming to improve therapeutic drug monitoring by considering haematocrit-normalised whole-blood concentrations. The model demonstrated accurate and precise predictions of everolimus exposure, suggesting that this approach could enhance dose adaptation and potentially improve patient outcomes. The findings support the clinical implementation of this model-informed precision dosing strategy for more effective management of everolimus therapy in renal transplantation.

This study evaluates the effectiveness of a cloud-based clinical decision support tool for dosing busulfan in pediatric patients undergoing hematopoietic cell transplantation. The research found that using this model-informed precision dosing platform significantly improved the accuracy of achieving target drug exposure levels compared to traditional dosing methods. The platform's use resulted in a higher percentage of patients reaching desired busulfan concentrations early in treatment, demonstrating the benefits of personalized medicine in improving treatment outcomes.

This study aimed to improve the dosing of vancomycin, a critical antibiotic for patients with severe infections, by using a model-informed precision dosing tool. The researchers selected the most effective model from a retrospective analysis and then validated it in a prospective multicentre study with critically ill patients. The results showed that the model could accurately predict the necessary dose of vancomycin, suggesting that this approach could enhance treatment efficacy and safety in clinical practice.

The study describes the successful implementation and use of a model-informed precision dosing (MIPD) tool for vancomycin in hospitalized children at an academic children's hospital. By integrating this tool with the electronic health record system and utilizing clinical pharmacists as key operators, the hospital was able to individualize vancomycin dosing for pediatric patients, including neonates and children, enhancing therapeutic drug monitoring efficiency. The adoption of this technology was well-received by users, with a significant majority finding it efficient and expressing satisfaction with its performance in clinical settings.

This study evaluates the effectiveness of a Bayesian Clinical Decision Support System (CDSS) compared to the standard of care (SOC) in determining personalized gentamicin dosing for neonates. The research found that the CDSS was more effective in achieving desired gentamicin peak concentrations in neonates than the SOC, particularly at the initiation of treatment and after adjusting doses based on measured drug concentrations. The findings suggest that using a CDSS could enhance the safety and efficacy of gentamicin therapy in newborns.

This article discusses the importance of incorporating real-world data (RWD) into patient care to enhance treatment outcomes. The authors argue that using RWD can lead to more personalized and effective healthcare strategies. This approach aims to bridge the gap between clinical research and everyday clinical practice, ultimately improving patient care.

This study compared the effectiveness of a Bayesian clinical decision support (CDS) software against traditional clinician judgement in determining the optimal dosing of vancomycin for pediatric patients to meet specific pharmacokinetic/pharmacodynamic targets. The research found that the CDS-guided dosing was significantly more successful in achieving desired vancomycin levels and pharmacodynamic outcomes than clinician-guided dosing. This suggests that using Bayesian CDS tools can improve the precision of vancomycin dosing in children, potentially leading to better therapeutic outcomes.

This study introduces a model-informed approach for optimizing the dosing of rifampicin in tuberculosis treatment, utilizing therapeutic drug monitoring (TDM). By analyzing pharmacokinetic data from patients, the research found that a simplified blood sampling strategy at 2 and 4 hours, when combined with model-based predictions, could replace the traditional method without losing accuracy. This new approach not only reduces the sampling time and complexity but also allows for more flexible sampling times, potentially improving personalized tuberculosis treatment.

The article discusses the advancement and significance of model-informed precision dosing (MIPD) tools, which are designed to optimize medication dosing for patients directly at their bedside through the use of applications. These tools promise substantial clinical benefits by utilizing available clinical data and improving medication dosing accuracy. However, the implementation of MIPD faces unique scientific challenges, including the need for adjustments in pharmacometrics diagnostics to enhance their predictive capabilities.

This study introduces a model-based approach, named Neo-Vanco, for individualizing empiric vancomycin dosing in neonates, aiming to address the challenges posed by the wide variation in their pharmacokinetics. The approach, based on a population pharmacokinetic model, was shown to significantly increase the likelihood of achieving optimal drug exposure compared to traditional dosing recommendations. The findings suggest that this individualized dosing strategy could potentially improve vancomycin therapy outcomes in neonates, highlighting the need for further clinical validation.

This study investigates the relationship between busulfan exposure and clinical outcomes in children and young adults undergoing allogeneic haemopoietic cell transplantation (HCT). It was found that targeting busulfan exposure to a specific range (78-101 mg × h/L) using a validated pharmacokinetic model significantly improves event-free survival and reduces the risk of graft failure or relapse, while higher exposures increase the risk of acute toxicity and transplantation-related mortality. The research suggests that precise dosing of busulfan, guided by pharmacokinetic modeling, can enhance the success of HCT in pediatric and young adult patients.