Department of Biostatistics Seminar/Workshop Series

A Bayesian Hierarchical Nonlinear Mixture Model in the Presence of Artifactual Outliers in a Population Pharmacokinetic Study

Leena Choi, PhD

Assistant Professor, Department of Biostatistics
Vanderbilt University School of Medicine

Wednesday, November 12, 1:45-2:55pm, MRBIII Conference Room 1220

Intended Audience: Persons interested in applied statistics, statistical theory, epidemiology, health services research, clinical trials methodology, statistical computing, statistical graphics, R users or potential users

A Bayesian Hierarchical Nonlinear Mixture Model in the Presence of Artifactual Outliers in a Population Pharmacokinetic Study
by Leena Choi, Utkarsh Kohli, E. Wesley Ely, and C. Michael Stein

---

The purpose of this study is to develop statistical methodology to estimate pharmacokinetic (PK) parameters in the presence of large proportion of artifactual outliers. The motivating PK data were obtained from a population PK study to examine associations between PK parameters such as clearance of dexmedetomidine and Cytochrome P450 2A6 variants. The blood samples were sparsely sampled from patients in Intensive Care Units (ICUs) while dexmedetomidine was continuously infused to achieve sedation goal. The conventional population PK analysis of this PK data revealed several challenges and intricacies. Especially, there was strong evidence that the plasma concentrations were contaminated with high concentration of the infused drug due to unavoidable sampling process in ICU setting. If not addressed, these could lead to biased estimates of PK parameters and miss important relationships between PK parameters and covariates due to increased variability. We propose a novel population PK model, a Bayesian hierarchical nonlinear mixture model, to accommodate the artifactual outliers using a finite mixture as the residual error model. Our results showed that the proposed model handles the outliers well. We also conducted simulation studies with varying proportion of the outliers. The simulation results also supported that the proposed model can accommodate the outliers well so that the estimated PK parameters are unbiased. There has been increasing evidence of the dangers of sedatives and analgesics which are nearly universally provided to critically ill patients in the ICU to relieve suffering. A large observational cohort study has been ongoing to examine the risk of these medications. Due to sparse sampling (for patient safety) of the PK data from this study, population pharmacokinetic and pharmacodynamic modeling approach needs to be employed to examine the associations between drug exposure and drug response such as cognitive impairment. The proposed method will provide critical methodology to address the challenges observed in observational PK studies which will provide important information to optimize doses of potent medications in order to reduce acute and long-term brain dysfunction.