Thesis

Building a resource-sparing direct compression formulation optimiser to accelerate medicine development

Downloadable Content

Download PDF
Creator
Rights statement
Awarding institution
  • University of Strathclyde
Date of award
  • 2026
Thesis identifier
  • T18173
Person Identifier (Local)
  • 202294822
Qualification Level
Qualification Name
Department, School or Faculty
Abstract
  • The development of directly compressed oral solid dose (OSD) tablets traditionally relies on resource-intensive, trial-and-error approaches, which are costly, wasteful, and hinder the potential to formulate rarer, more expensive, medications. A major challenge lies in systematically understanding the complex interactions between material properties, process parameters, and critical quality attributes - namely porosity and tensile strength. This thesis introduces a digital approach designed to predict how material contributions affect these quality attributes and support formulation scientists in navigating formulation development through data-informed experimental strategies. The research begins by evaluating common empirical models for compressibility and compactability to determine their ability to predict drug-loaded formulation behaviour. A literature review revealed significant variability in reported fitting parameters for these models, into which further investigation revealed that the existence of multiple parameters was not a detriment to model performance but instead indicated the models’ inherent flexibility. Building on this, previous work on mixture rules for predicting material contributions to blend behaviour was advanced. A newly developed algorithm leverages this flexibility through a global optimization approach, identifying optimal component parameters that, when applied via mixture rules, can accurately predict compression and compaction behaviours based on a blends composition. Finally, a Model-Based Design of Experiments (MBDoE) is proposed to guide experimental campaigns, reducing reliance on trial-and-error methods. Results demonstrate that integrating MBDoE with optimised mixture rules can reduce active pharmaceutical ingredient (API) consumption by over 30% across all formulations studied, with median savings consistently exceeding 75–95% under the Acceptable and Good performance thresholds, as determined through performance threshold crossing analysis of both the Kawakita and Ryshkewitch–Duckworth compaction models. Savings were found to decrease with increasing threshold stringency, with the widest variability observed in the ibuprofen formulations under the Excellent performance criterion. The Kawakita model consistently supported API reductions across all threshold levels, whilst the Ryshkewitch–Duckworth model was more limited in its capacity to achieve crossings beyond the Acceptable threshold for several formulations. These reductions were achieved without compromising tablet quality, although the MBDoE approach did not consistently outperform random selection across all formulation spaces, suggesting that the threshold-based optimisation framework itself is the primary driver of API reduction rather than the experimental design strategy employed. Overall, this work establishes a systematic methodology for tablet formulation development by combining predictive modelling with structured experimental design. By bridging material science and computational optimisation, the proposed framework contributes toward more sustainable and efficient pharmaceutical development.
Advisor / supervisor
  • Markl, Daniel
Resource Type
DOI

Relations

Items

Thumbnail Title Date Uploaded Visibility Actions
file details PDF of Thesis T18173 2026-10-12 Public Download