Klartextanalyse pathologischer Biopsiebefunde mit Bildschirmabfrage

Abstract

Beschreibung eines Klartextsystems zur Dokumentation von Befunden in der Pathologie, das sich seit 1. 1. 1971 im Testbetrieb befindet und bisher 8 000 Biopsiebefundberichte analysiert und gespeichert hat. Der zugrundeliegende Thesaurus verfügt über 39 000 Eingangsworte; weitere 10 000 Worte werden demnächst eingefügt.Das entwickelte System ist das erste Dokumentationssystem mit Bildschirmabfrage, das in der Pathologie eingesetzt wird.In der Diskussion werden Hinweise auf den notwendigen Aufwand an Maschinenkonfiguration und Manpower gegeben und die verschiedene Struktur der in der Pathologie verwendeten Thesauri erörtert.Die Analyse der Möglichkeiten eines Klartextsystems ergibt drei Stufen der Entwicklung von der wortweise erfolgenden Erfassung bis zur Erfassung des Begriffsinhaltes; praktisch alle Klartextsysteme in der Pathologie befinden sich auf der ersten Stufe dieser Entwicklung, ‘

Computer-Linked Memorandum of Pathology System (CLIMPS 7909) for Interactive Storage and Analysis of Surgical Pathology Data

“…To make it a fruitful effort for providing material for reference in histopathological diagnosis, the data must be pooled in a readily retrievable form and shared amongst pathologists. A basic requirement for successful cooperation is that the system can accept every pathologist as a data producer I user so that the labor of data supply is immediately rewarded by a return of necessary information [1,5]. In tIns respect, on-line systems are decidedly advantageous compared to batch processing methods.…”

“…The essential requirement for the system to be used in routine work is the ease with which the data are stored and retrieved by pathologists untrained in computer technology [5]. A sequence of the processes that is flexible enough to meet users' demands also is an important fea• ture.…”

“…A sequence of the processes that is flexible enough to meet users' demands also is an important fea• ture. Other factors that lower the threshold of man• machine interface and thus encourage pathologists would include (I) immediate accessibility of the data base [2,4,5,19,20,29], (2) interactive dialogue that permits machine•assisted operation [6,10,14,15,19,20], (3) storage of the essential minimum information, resorting to hand•written documents for details, (4) ease of cor• rection and updating of data, (5) automatic coding capability which eliminates need for coding by hand of diagnostic terms [S,10,16,IS,20,21,2S,30], and (6) workability frOIn the very begiruling without prelinlinary construction of a thesaurus for coding. The speed of data acquisition and the number of records that can be stored are of secondary importance.…”

Research at the Department of Medical Informatics, Statistics and Documentation of the University of Graz

“…In 1989 the section became an independent Department of Medical Informatics where duties comprise research and education and, in cooperation with the informatics services of the hospital association, the planning of infrastructure and the planning and implementation of clinical IT projects. Ourresearchprojects and clinicaldevelopments include, amongst others: quantitative modeling in nuclear medicine [3], computer-assisted radiotherapy [3][4][5][6][7][8], clinical information systems [9][10][11][12][13][14], PACS [14][15][16][17], image processing [ 18], and computer-assisted diagnosis [19,20].…”

“…We therefore started a project to implementautomaticindexingfromfree text to SNOP (Systematic Nomenclature of Pathology) and to a local radiology code. In pathology for example, a comprehensive thesaurus with over 60,000 entries was constructed for the indexing ofhistology findings [9]. Despite some success, the approach towards automatic indexing was eventually abandoned in favor of a simple search for words in the reports, where the handling of synonyms, homonyms, negations, etc.…”