Dynamic Disease, Fragmented Processes, and System Risk Why Patient Pathways Are Not Processes — and Why This Matters for Clinical Evaluation and Digital Systems

This paper examines a structural mismatch in healthcare systems: diseases behave as dynamic biological processes, while care delivery is organised as fragmented patient pathways composed of multiple, loosely coupled processes. When patient pathways are implicitly treated as single processes, standardisation, quality improvement, and digital system design risk optimising local steps while increasing system-level risk. Using secondary failure to biologic therapy in psoriatic arthritis as an illustrative example, the paper shows how loss of treatment effectiveness over time reflects disease dynamics interacting with fragmented care processes, rather than isolated treatment failure. The analysis highlights how assumptions of stability at fixed response checkpoints obscure temporal drift, partial control, and evolving disease trajectories. The paper argues that safe clinical evaluation, quality improvement, and post-adoption governance of digital health systems and clinical AI require an explicit distinction between disease-as-process and care-as-pathway. Recognising this distinction is essential to avoid hidden drift, local optimisation, and false assumptions of long-term safety and effectiveness. This work builds on a dynamic disease perspective and is conceptually related to: Domargård A. The Universal Resonance Model (URM): A Mathematical Kernel for Dynamic Disease Systems. Zenodo. DOI: 10.5281/zenodo.18419250