A Reference Model Based Patient Management System: Opportunities and Challenges

摘要

Object oriented design (OOD) and hierarchical layered architecture are the dominant paradigms for building computer-based applications. The biological system deviates significantly from these paradigms, while adding self-adaptability as a paradigm that is worthy of emulation in the computer domain. Biological system components may be considered to be nearly decomposable, but not modular; further, the top-down command and control, the bottom-up information flow, and the fully submissive & reactive lower levels, as envisioned with a layered architecture, is replaced in biology with overlapping and self-aware/adaptive/intelligent layers. These differences pose challenges in integrating seamlessly computer and biological systems. A seamless system could help build personalized computer-based patient management systems for optimal management of chronic conditions such as Diabetes and Hypertension. The lower levels of sensing, data acquisition, data processing, and state detection would be based on the biological system while the higher levels of health assessment, prognostics, and advisory generation, generally the domain of medical personnel, could be substantially supplemented by a computer based expert system. With a better appreciation of the two domains, decisions made at the higher levels (in the computer domain) would be tempered with adaptive strategies that comprehend the adaptive and partially-reactive responses of the lower levels (of the biological system) for maximum benefit to the patient. This requires a seamless model that integrates the computer and biological domains. In this paper, we present our attempts at building such a seamless model for integration of computer and biological systems. We used OPD (object process diagrams) to build a systems biology OOD hierarchy at lower levels. The scope of this paper is limited to mapping glucose metabolism to an object oriented model, which is a prerequisite to building a viable multi-layer model.