School of Pharmacy
Design Space, Figures of Merit, Monte Carlo Simulation, Process Analytical Technology, Quality by Design, Risk Assessment
The design of drug delivery systems and their corresponding dosing guidelines are critical product development functions supported by clinical pharmacokinetic (PK) and pharmacodynamic (PD) data. Largely, the importance of variance and covariance in product and patient attributes is poorly understood. The existence of PK/PD diversity among myriad patient sub-populations further complicates efforts to gauge the importance of product quality variation. Nevertheless, a platform capable of evaluating the effects of product and patient variability on clinical performance was constructed. This dissertation was predicated on requests to re-define pharmaceutical quality in terms of risk by relating clinical attributes to production characteristics.
To avoid in vivo studies, simulated experimental trials were conducted using the model drug, theophylline, for which data and models could be acquired from the literature. Where comprehensive data were unavailable (e.g., production variability statistics), initial estimates were acquired via laboratory-scale experiments. Model asthmatic patients were generated using Monte Carlo simulation and published population distributions of various anothropometric measurements, disease rates, and lifestyle factors.
Mathematical constructs for in vitro-in vivo correlations provide a linkage between Quality by Design (QbD) product and process models, PK/PD models, and patient population statistics. The combined models formed the foundation for Monte Carlo risk assessments, which characterized the risk of inefficacy and toxicity for dosing of extended-release theophylline tablets. Sensitivity analyses revealed that patient compliance and content uniformity significantly influenced the probability of observing an adverse event.
The Monte Carlo risk assessment platform defined the link between the critical quality attributes (CQAs) and clinical performance (i.e., performance-based quality specifications (PBQS)). The PBQS were subsequently utilized to generate process independent design spaces conditioned on inefficacy and toxicity risk. These design spaces, which directly account for the conditional relationships between product quality and patient variability, can be transferred to a specific process via models that relate process critical control parameters to the CQAs. Process Analytical Technology, therefore, can be integrated into the QbD production environment to control the safety and efficacy of the final product. This work demonstrated that process and product knowledge can be used to estimate the risk that final product quality imparts to clinical performance.
Short, S. (2009). Performance-Based Quality Specifications: The Link between Product Development and Clinical Outcomes (Doctoral dissertation, Duquesne University). Retrieved from https://dsc.duq.edu/etd/1190