Linear Discriminatory Analysis: A Patient Classifying for Research and Production Control

Administrative and medical information about the patient forms, in each case, a pattern, the complexity of which increases as the number of data grows. Even when the data are 4—5 in number, the human ability to recognize and distinguish between different patterns begins to fail, A mathematical method (linear discriminatory analysis) has been worked out. This system of analysis appears to provide opportunities of placing patients with the same or similar patterns in classes which are diagnostically, prognostically or therapeutically homogeneous.

Administrative und medizinische Informationen über den Patienten stellen in jedem Einzelfalle ein Erscheinungsbild dar, dessen Komplexität mit der Anzahl der einzelnen Daten zunimmt. Schon bei nur 4—5 verschiedenartigen Daten beginnt die menschliche Fähigkeit, unterschiedliche Konstellationsmuster zu erkennen und sie sicher voneinander zu unterscheiden, zu versagen. Verfasser haben unter Benutzung einer mathematischen Differenzierungsmethode (lineare Diskriminanzanalyse) ein Analysen. verfahren ausgearbeitet, das geeignet erscheint, Patienten mit gleichen oder ähnlichen Symptomenbildern in diagnostisch, prognostisch und therapeutisch homogene Klassen einzuordnen.

Key-words

Patient Classification - Discriminatory Analysis - Medical Data Bank - Medical Record System J5 - Anaesthesia Protocols - Anaesthssia Risk Groups

Schlüssel-Wörter

Patienten-Klassifizierung - Diskriminanzanalyse - Medizinische Datenbank - Krankenblatt-System J5 - Anaesthesie-Protokolle - Risiko-Gruppen (Narkose)

• Reference

• 1 Brain A. E. A Large Self-contained Learning Machine. (WESCON paper 6. 1. 1963).
• 2 Danielsson T, Eklund J, Hall P, Hallen B, Selandeh H. Anestesi och Informationssystem J5. Läkartidningen 64 1967; 3214-3221.
• 3 Enslein K. An Approach to Pattern Recognition. In: Enslein K. (Edit,), Data Acquisition and Processing in Biology and Medicine. Vol. 1, 91-116. Pergamon Press; Oxford, London, New York, Paris: 1962
• 4 Frank O. Databehandlingsmetoder for medicinsk diagnosticering. Mimeograph. (Statistiska institutionen, Stockholms Universitet, 1967).
• 5 Frank O. Metoder for klassificering. Forskningsrapport. (Statistiska Institutionen. Stockholms Universitet, 1968).
• 6 Gohdh T. An anesthesiological study of 222 gastric resections. Acta chir. scand 98 1949; 248-249.
• 7 Hall P, Mellner C, Danielsson T. JS - a data processing system for medical information. Meth. Inform, Med 06 1967; 1-6.
• 8 Hall P. et al. J5 Medical Information System, User’s Manual. Report from Department of Information Processing, Karolinska Hospital, Stockholm 1967).
• 9 Halldin M. Narkos och bedömning. 3. Edit. Svenska Bokförlaget Bonniers; Stockholm: 1968
• 10 Nilsson N. J. Learning Machines: Foundations of Trainable Pattern Classifying Systems. McGraw-Hill Book Company; New York: 1963
• 11 Petterson W. W. Error-correcting Codes. M.I.T. Press; Cambridge/Mass: 1961
• 12 Saklad M. Grading of patients for surgical procedures. Anesthesiology 02 1941; 281-284.
• 13 Stark L, Okajima M, Whipple G. Computer pattern recognition tediniques: electrocardiographic diagnoses. Comm. Ass. comput. Mach 05 1962; 529-532.
• 14 Sterling T. D, Saenger E. L, Gleser M. Robot data screening. In: Enslein K. (Edit.), Data Acquisition and Processing in Biology and Medicine. Vol. 2, 233-244 Pergamon Press; Oxford, London, New York, Paris: 1964