TerminAI - Optimization of healthcare data management through Artificial Intelligence

Context:

Nowadays, the management, use and sharing of genetic data face multiple challenges, among others, due to the lack of unified clinical standards. This issue limits their potential for diagnostics, biomedical research and integration into advanced healthcare infrastructures. Initiatives such as the European Health Data Space (EHDS) and the AI Act seek to transform this scenario, promoting the safe, interoperable and ethical use of data.

In this context, the need arises to improve the management of genetic data, through initiatives such as mapping genetic data to standard clinical terminologies using artificial intelligence, enhancing its applicability in accurate diagnoses and personalized treatments.

Aim of the project:

And this is where TerminAI is born, a project that seeks to improve the accuracy and efficiency in the coding of clinical information through Artificial Intelligence. The project uses natural language processing and machine learning to analyze clinical data, identifying patterns and concepts that map to SNOMED CT terms or other standard terminologies.

It also applies bioinformatics tools and incremental learning to process genetic data, improving disease diagnosis and prevention. In addition, it digitizes informed consents to facilitate access and reduce manual typing errors.

TerminAI is also exploring the integration capabilities of these solutions with European infrastructures such as health data spaces, ensuring compatibility, secure exchange and compliance with European requirements.

Role of Vicomtech:

Vicomtech is responsible for the analysis and evaluation of technologies for the development of an AI-assisted clinical coding system. This includes the analysis of clinical terminology models such as SNOMED CT, ICD-10 and LOINC, as well as the investigation of software tools to integrate multiple terminologies and manage large volumes of data accurately and efficiently. In addition, it explores the use of advanced algorithms such as natural language processing (NLP), machine learning (ML) and neural networks to ensure an adaptive and efficient system that optimizes clinical coding.

Work is also underway to explore synthetic generation (SDG) of genetic data, an innovative technology that allows the creation of artificial-genetic data representative of the population. This approach allows the privacy of individuals to be preserved while maintaining the quality and integrity of the original data, facilitating the secure sharing of information in data spaces and overcoming the privacy limitations associated with real genetic data. 

In addition, the center's involvement extends to exploring the integration capabilities of these solutions with European infrastructures such as BBMRI-ERIC, GAIA-X and EOSC, ensuring their interoperability in healthcare data spaces. This includes analysis of advanced cybersecurity measures and compliance with key regulations such as GDPR, the AI Act and EHDS. By connecting Vicomtech's technologies with these initiatives, we facilitate secondary use of data, foster international collaboration and contribute to the development of innovative medical solutions that drive the advancement of biomedical research.

Partners:

Vicomtech participates in the consortium formed by NorayBio, BIOLAN, EProcessMed, Devol, AseBio and Basque Health Cluster, which come together to investigate the potential of AI and its ability to optimize the management, codification and harmonization of health data.