Databases of potential digital doubles of the software-modeling complex for predicting leading indicators of patients' condition

In modern personalized medicine, there is a transition from the phenomenological description of clinical cases to the predictive modeling of individual patient trajectories. In this context, the concept of medical digital twins seems to be not so much a technological innovation as a qualitatively new epistemological tool that requires rethinking traditional approaches to the organization and structuring of medical data. In this regard, the fundamental challenge is not the very possibility of computational modeling of physiological processes, which is technically feasible in one way or another, but the formation of an adequate information infrastructure capable of ensuring the representativeness and validity of the source data for predictive algorithms. The methodology of forming databases of potential digital counterparts in medicine is currently under development. Our research has allowed us to develop an approach to the formation of a database of medical digital twins and formulate several fundamental provisions that determine the directions for further development of the information infrastructure for predicting the condition of patients based on them.