Probabilistic Pharmacokinetic Models for Diagnosis, Prognosis, and Personalized Treatment

Clinicians have recognized that the nature of diseases can be highly individual resulting in different patterns of onset and progression. In turn, the response of an individual to drugs is also unique and governed by a variety of factors. Pharmacokinetic models represent the movement of a drug through the body, and personalized pharmacokinetic models aim to capture the unique responses of specific individuals. The overall goal of the Phase I research effort is to develop a methodology for using probabilistic modeling in silico to construct personalized pharmacokinetic models. The team of Barron Associates and Virginia Commonwealth University will leverage prior research to accomplish this goal. A general model that represents a large cohort of individuals will first be generated based on existing data and known model structures. Any available information regarding a specific patient can then be used to personalize the model for that patient by computing conditional distributions over expected outcomes. Phase I proof-of-concept demonstrations will apply the proposed methodology to two different application areas: diabetes and traumatic brain injury. The team will also develop initial clinical strategies for how personalized pharmacokinetic models might guide diagnosis, prognosis, and treatment.