A Simple and Replicable Framework for the Implementation of Clinical Data Science

Abstract

138A large amount of clinical data generated in the hospital environment allows us to imagine the possibilities that their analysis would offer. Different technologies such as Big Data, Data Science (DS), machine learning (ML) or artificial intelligence (AI) can be applied in order to improve prevention and early detection, diagnosis, treatment and monitoring, research and management in multiple pathologies of high incidence and high human and economic cost. Although these technologies promise to revolutionize medicine as we know it, the reality that we can currently observe in the hospitals of the Spanish National Health System (SNS) shows us that we are not yet able to implement these technologies in an effective and broad way. There are different reasons why their real implementation in the hospital environment is still scarce, among which are organizational reasons and availability of resources, the highly fragmented map of information systems at hospitals and therefore of clinical data, problems of governance and related to the very nature of clinical data and the need to guarantee the protection of personal data. However, despite the importance of all these issues, we consider that the key problem is the lack of a framework, understood as a set of concepts, procedures and tools, which facilitates the availability of adequate data and the multidisciplinary analysis of clinically relevant problems. Without having this framework implemented in the hospital, the different techniques and algorithms (from linear regressions to deep learning) lack practical utility.

Ähnliche Arbeiten