Medical data compatibility problems in the tasks of information systems integration. Conceptualization

Abstract

Integration of medical data is a critical component of the functioning of modern healthcare systems and a primary task of personalized medicine. Aggregating data from disparate sources, such as electronic medical records and medical devices, allows service providers to obtain a complete picture of patients' health status and optimize workflows. It is noted that the strategy of integration is closely linked with the logic of medical data compatibility. The problems of integration also reflect in the tasks of creating Portfolios for physicians and pharmacists. Research objective. Summary of biomedical data compatibility issues. It is emphasized that data compatibility depends on the consistency of standards applied in programs. The quality of data also requires special attention. This directly affects the quality of the decisions made. Although data interoperability is one of the primary requirements of information and communication systems (ICS), it is often overlooked. As a result, data exchange is not performed, significantly limiting the flow of information. The problem of large dimensionality is also serious. It is evident that such multidimensional and labor-intensive computational processes are a primary task for modern algorithms and models of artificial intelligence (AI) and machine learning (ML). It is highlighted that one of the ways to solve the problems of standardization and integration of large-scale medical data are metadata, which are also useful for improving statistical analysis, probabilistic models, and ML models. Conclusions. 1. Precision Medicine (4P Medicine) was introduced as a new paradigm approach to healthcare with a more predictable, preventive, personalized, and participatory manner. Precision medicine is closely related to data-intensive approaches, as well as to ML and AI. 2. Integration of data for placement in structures useful for precision medicine is only possible after several previous stages of their processing, namely: data (metadata) collection, processing, obtaining 'clean' data, data compression. 3. To realize the prospects of precision medicine, approaches to computational learning must evolve with the help of well-chosen and wellintegrated digital data ecosystems.