The method of primary processing of poorly structured medical data

The article deals with the approach to the primary processing of poorly structured medical protocol textual data stored and disseminated as pdf files. The relevance of this work is due to the lack of a universal structure for the presentation of medical protocols and methods of their processing. In the course of the work, the problem of primary processing of clinical protocol data was solved by the example of a unified clinical protocol of primary, secondary (specialized) and tertiary (highly specialized) medical care. The method of primary data processing was developed to create a clear structure of the symptoms of the disease. The first step in structuring clinical protocol data is to divide the protocol information into four basic parts, which allows it to be quickly converted to other formats. This process is implemented using an algorithm developed in C # programming language. The proposed algorithm parses the information from a pdf file and converts it to a txt file. After that, the received information is processed, which consists in the syntactic analysis of the text of the protocol and selection of the structural parts of the protocol corresponding to the headings of the sections: title page; introduction; a list of abbreviations used in the protocol; the main part of the protocol; list of literary sources. The identification of the disease name in the medical protocol is performed by comparing the protocol data and the list of disease names, presented in the world classification MKH-10. The headings “Introduction”, “List of abbreviations used in the protocol” and the main part of the protocol were analyzed and the algorithm for removing uninformed sections from the beginning of the protocol, for example, literature sources, was proposed. An algorithm for finding information in the main part of the medical protocol by processing input data by: tables, diagrams, headings, words, phrases and special symbols are also proposed. As a result of the clinical protocol processing algorithms, a new clinical protocol file is generated, which is three times smaller than the original file. It contains only meaningful information from clinical protocols that will speed up further work on this file, namely its use in medical decision support. The disease card based on a medical protocol in JSON format is presented.