We propose a three year program of investigation into the use of expert systems for the diagnosis of liver disease, using deep reasoning methods based on structure/function models. This research would lead to the refinement of our prototype system Pathex/Liver-1 for liver disease diagnosis. As part of this refinement process, techniques and methodologies of artificial intelligence will be used to investigate the problem-solving of experts in the diagnosis of liver disease. Specifically, research will be directed toward: 1. the expansion of the current Pathex/Liver-1 expert system in order to recommend diagnoses in a larger class of liver diseases and associated disorders. 2. the expansion of the current Pathex/Liver-1 expert system in order to increase the types of data it can gather. 3. The development of a deeper understanding of the representations and problem-solving methods needed to integrate structure/function models with traditional techniques for constructing medical knowledge-based systems, and 4. the continuing development of high level knowledge-engineering tools to support the implementation of medical diagnostic systems. The resulting prototype system Pathex/Liver-2 will represent a significant advance in the state of the art in medical knowledge- engineering demonstrating the way structure/function models may be integrated with compiled diagnostic knowledge to improve diagnostic reasoning performance. The high level knowledge-engineering tools developed will also represent an advance in the state of the art, representing a flexible, extensible, and modularized tools environment for representing and integrating diagnostic medical knowledge in other domains.

National Institute of Health (NIH)
National Library of Medicine (NLM)
Research Project (R01)
Project #
Application #
Study Section
Biomedical Library and Informatics Review Committee (BLR)
Project Start
Project End
Budget Start
Budget End
Support Year
Fiscal Year
Total Cost
Indirect Cost
Ohio State University
Schools of Medicine
United States
Zip Code
Smith Jr, J W; Bayazitoglu, A; Johnson, T R et al. (1995) One framework, two systems: flexible abductive methods in the problem-space paradigm applied to antibody identification and biopsy interpretation. Artif Intell Med 7:201-25
Smith Jr, J W; Bayazitoglu, A (1993) Exploring the relationship between rationality and bounded rationality in medical knowledge-based systems. Artif Intell Med 5:125-42
Amra, N K; Smith Jr, J W; Johnson, K A et al. (1992) An approach to evaluating heuristics in abduction: a case study using RedSoar--an abductive system for red blood cell antibody identification. Proc Annu Symp Comput Appl Med Care :690-4
Bayazitoglu, A; Smith Jr, J W; Johnson, T R (1992) A diagnostic system that learns from experience. Proc Annu Symp Comput Appl Med Care :685-9
Fischer, O; Smith, J W (1991) Contextual pathognomony: a computationally useful extension of pathognomony. Comput Methods Programs Biomed 35:311-9
Johnson, K A; Johnson, T R; Smith Jr, J W et al. (1991) RedSoar--a system for red blood cell antibody identification. Proc Annu Symp Comput Appl Med Care :664-8
Bylander, T; Smith Jr, J W; Svirbely, J R (1988) Qualitative representation of behavior in the medical domain. Comput Biomed Res 21:367-80