In clinical practice there is often a need for a universal system that could analyze heterogeneous data, both in a structured form and in an unstructured one. In addition, it should allow to automate a wide range of activities within the framework of the treatment process. This will improve the quality of extraction and use of clinical data. One of the solutions may be the use of a system of intelligent analysis of medical information. It allows the aggregation of various methods of analyzing medical data and texts; therefore, it will help optimize the work with clinical data.
According to Hernandez, Roman, and White (2020), to improve the quality of the use of clinical data, systems for the intelligent analysis of medical information can be used. They process clinical data according to the sets of rules defined in the system and are able to solve diagnostic problems. The system can also extract and organize clinical analyses based on rules and regular expressions. First, the results of various clinical tests, which are string or numeric values, would be extracted and optimized. Then the system would recognize medical terms in the text according to the specified codifiers (for example, diseases, drugs, or medical procedures). The data sources would be structured string fields and numeric data, as well as sizeable unstructured text blocks: anamnesis and examination results.
To ensure the quality of the data of the machine method of working with clinical analyses, the procedure of multiple cross-checking would be used. In addition, for this purpose, regular verification of the correct configuration of the system would be carried out. The methods used to extract information from clinical texts and analyze medical data and their aggregation allow to speed up the process of differential diagnostic search in severe, chronic, disabling diseases. Therefore, the value for use of the obtained data can be the absence of delay in the appointment of pathogenetic therapy.
The use of the system of intelligent analysis of structured and textual data within a single procedure can significantly improve the quality of diagnosis of chronic diseases. The use of preliminary categorization of numerical indicators of patients’ health would also make it possible to achieve a steady increase in the quality of diagnostics. Automation of working with clinical data would allow optimizing the process of obtaining and processing them, eliminating the factor of human error.
Hernandez, A. V., Roman, Y. M., & White, C. M. (2020). Developing criteria and associated instructions for consistent and useful quality improvement study data extraction for health systems. Journal of General Internal Medicine, 35(6), 802-807.