Michael L. Williams dissertation defense – September 30
***Updated 9.29. The on-site location is now Room 1502.***
PhD candidate Michael L. Williams will be defending his dissertation on Friday, September 30, at 10:00 a.m. Central Time, on campus (2525 West End Avenue, 15th Floor, Room 1502) and online (contact biostatistics[at]vumc[dot]org for access via Zoom). His advisor is Leena Choi.
Using Electronic Health Record Derived Medication Data for Population Pharmacokinetic Studies
The Electronic Health Record (EHR) is a source of vast information about clinical drug dosing in patients. However, the extraction and use of this data in research poses significant challenges due to its observational nature and clinically focused purpose. This dissertation is centered on the implementation of EHR derived medication data to perform population pharmacokinetic (PopPK) studies. In this dissertation, first we will discuss the effect of incorporation of the last-dose time associated with tacrolimus therapeutic drug monitoring (TDM) blood-draws in the EHR. Specifically, we evaluated if the incorporation of that information into the PopPK dataset yields model estimates that vary from those obtained with assumed timing of the previous dose. We will then present an analysis of genetic effects on fentanyl pharmacokinetics. We demonstrated an effect on fentanyl clearance and investigated the significance for clinical dosing decisions. Next, we will present an R Shiny application we developed for assisting TDM decision-making for pediatric drug dosing. We will conclude with a discussion on an extension of our previously reported software system for extracting and processing medication data from the EHR. We developed and tested several modeling solutions to resolve the appearance of inaccurate dosing data in the EHR to obtain the actual dosing schedule as represented in the EHR.