Semantic Analysis of Medical Records

Abstract:

A program (LogStory) is described that was developed for the automatic semantic analysis of clinical narratives, stored in a computerized problem-oriented medical record (PROMED). The diagnoses were written in a free-text format during consultation, and later collected into diagnostic classes, e.g., diseases. A lexical parser automatically created dictionaries from the clinical narrative associated with each disease. Automatic (fuzzy) set operations were performed on the words associated with each class. The manifestations of 16 diseases were automatically extracted by pairwise operations on the word sets. The correlation between diseases and corresponding signs, symptoms and treatment was highly significant (p <0.001). Applying the difference operation on diseases with disjunct sets of clinical findings allowed the recovery of disease-specific knowledge. The evolution of a disease was accounted for, and the system was able to generalize its findings. The PROMED-LogStory concept enables the processing of natural language and may be a powerful tool for knowledge acquisition and clinical research.

Key-Words

Medical Record - Expert System - Fuzzy Set Theory - Knowledge Acquisition - General Practice - Natural Language Processing

• REFERENCES

• 1 van der Lei J, Musen MA, van der Does E, Man In’t Veld AJ, van Bemmel JH. Comparison of computer-aided and human review of general practitioners’ management of hypertension. Lancet 1991; 338: 1504-8.
• 2 Payne TH, Goroll AH, Morgan M, Barnett GO. Conducting a matched-pairs historical cohort study with a computer-based ambulatory medical record system. Comput Biomed Res 1990; 23: 455-72.
• 3 Nilsen JF, Bassøe C-F. Contraception in general practice analyzed by a computerized medical record. Scand J Prim Health Care 1987; 05: 217-20.
• 4 Sager N, Friedman C, Lyman MS. Medical Language Processing. Reading Mass: Addi-son-Wesley Publ Comp; 1987
• 5 Weed LL. Medical records that guide and teach. N Engl J Med 1968; 278: 593-600.
• 6 Bassøe C-F. A problem-oriented medical record worked, whereas a chronological did not. In: Hansen R, Solheim BG, O’Moore RR, Roger FH. eds. MIE88. Lecture Notes in Medical Informatics. Berlin: Springer Verlag; 1988: 1904.
• 7 Saltan G. Developments in automatic text retrieval. Science 1991; 253: 974-80.
• 8 Bonnet A. Strategies for understanding structured English. In: Buchanan BG, Shortliffe EH. eds. MYCIN. Reading Mass: Addison-Wesley Publ Comp; 1985: 613-34.
• 9 Straley SJ. Programming in Clipper. Reading Mass: Addison-Wesley Publ Comp; 1989
• 10 Skagestein GM, Thorvaldsen A. Fra Vir-kelighet til Datamodell. Oslo: Universitets-forlaget; 1988
• 11 Aho AV, Sehti R, Ullmann JD. Compilers Principles, Techniques and Tools. Reading Mass: Addison Wesley Publ Comp; 1986
• 12 Elmasri R, Navathe SB. Fundamentals of Database Systems. Redwood City Calif: Benjamin/Cummings Publ Comp; 1989
• 13 Dougherty ER, Giardina CR. Mathematical Methods for Artificial Intelligence and Autonomous Systems. Englewood Cliffs NJ: Prentice Hall; 1988: 59-68.
• 14 Negoita CV. Expert Systems and Fuzzy Systems. Menlo Park Calif: Benjamin/Cummings Publ Comp; 1995: 6-9.
• 15 Rosner B. Fundamentals of Biostatistics. 2nd ed.. Boston Mass: Duxbury Press; 1986: 467-72.
• 16 Schank RC. Dynamic Memory. Cambridge Mass: Cambridge Univ Press; 1982
• 17 Joshi AK. Natural language processing. Science. 1991 253. 1242-9.
• 18 Charniak E, McDermott D. Introduction to Artificial Intelligence. Reading Mass: Addison-Wesley Publ Comp; 1985 169-245, 393-5, 591-602, 609-59.
• 19 Luger GF, Stubblefield WA. Artificial Intelligence and the Design of Expert Systems. Redwood City Calif: Benjamin/Cummings Publ Comp; 1989 333-6, 363-7, 377-405.
• 20 Bassøe C-F. A combinatorial diagnostic system for general practice: Evaluation of the social impact of disease. In: Hansen R, Solheim BG, O’Moore RR, Roger FA. eds. MIE 88. Lecture Notes in Medical Informatics. Berlin: Springer Verlag; 1988: 212-4.
• 21 Bassøe C-F. Skinache: A new, common disorder identified using a computerized, problem-oriented medical record. In: Hansen R, Solheim BG, O’Moore RR, Roger FA. eds. MIE 88. Lecture Notes in Medical Informatics. Berlin: Springer Verlag; 1988: 212-4.