BARITONE
BARITONE
Boosting digital transfer in healthcare: integration of clinical data and AI technologies for high accuracy phenotyping of complex diseases (BARITONE)
Driving the digital transition in health: integration of clinical data and Artificial Intelligence technologies for high-precision phenotyping of complex diseases (BARITONE)
Welcome to BARITONE
The BARITONE project aims to promote the digitization of the health care value chain, with special emphasis on the exploitation of data by clinical staff, researchers and Artificial Intelligence scientists, to improve patient care .
BARITONE will focus on minority autoimmune diseases in adults, where the digitization of the data will serve as a valuable proof of concept. The technological infrastructure used will be used for various other research and clinical applications, and will benefit the different actors of the health value chain.
The project has been published on the web Orphanet of minority diseases, a unique resource that brings together and improves knowledge about minority diseases to improve the diagnosis, care and treatment of patients who suffer from them.
Financing
Ecological Transition and Digital Transition Projects 2021. File No. TED2021-129974B-C21
Duration of the project: 2 years
Project IP in the I3PT: Dr. Jordi Gratacós, Rheumatology Service, Parc Taulí University Hospital
Economic endowment: €287.500
Participating centers
Parc Taulí Research and Innovation Institute (I3PT)
SUBPROJECT 1 IP: Jordi Gratacós (project coordinator)
TITLE: Optimization of a data lake clinical through the incorporation of validated digital tools to improve high-precision phenotyping of complex diseases.
Barcelona Supercomputing Center (BSC)
SUBPROJECT 2 IP: Martin Krallinger
TITLE: Development of an interoperable infrastructure of Spanish health data digitization technologies based on language processing and artificial intelligence.
hypothesis
The use of Artificial Intelligence algorithms to integrate all health data into one on-premise datalake will facilitate a precise clinical profile of the patients.
Therefore, it is hypothesized that the use of machine learning algorithms in a data lake clinic for adult minority autoimmune diseases (ARADs) will optimize the classification of these diseases.
This technological implementation is essential to increase the likelihood of finding new predictive and prognostic biomarkers to facilitate the evolution towards personalized medicine and improve the well-being of society.
Objectives and methodology
The main purpose of BARITONE is adapt a digital infrastructure (i.e. a data lake) using Artificial Intelligence (AI) tools for improve the clinical profile of complex diseases such as ARADs.
The transformation towards an efficient digitized system requires collaboration between different disciplines. Therefore, both partners will contribute with their experience to successfully achieve the following goals:
Subproject 1. The I3PT will contribute clinical and care information systems expertise.
- Determine the maximum clinical data related to ARADs from hospital sources to allow their integration, interoperability and exploitation in a data lake clinical
- Identify clinical longitudinal phenotypic patterns of ARADs for a more precise classification within each specific disease.
- Design a strategy of scale-up to export technological developments to other hospitals accredited by ARADs of the Health System.
- Define future omics studies derived from clinical phenotypes to advance towards personalized and precision medicine with a higher chance of diagnostic success.
- Incorporate the results communicated by the patient to obtain their state of health without any interpretation of the answer (collaboration with Foundation29).
Subproject 2. BSC will be responsible for developing the AI tools needed to facilitate data mining models (mining) for its exploitation and analysis.
- Developing Advanced Machine Learning Semantic Annotation Tools (Machine Learning) for the identification of entities with relevance for the detection of ARAD in EHR.
- Generation of AI entity normalization models capable of mapping clinical entities to medical terminologies that provide semantic interoperability to EHR textual data.
- Creation of a health knowledge graph to extract value from the entities detected in the patients' electronic clinical history.
- Generating synthetic data to provide the community with representative and anonymized data on ARADs that can be used to promote research in this field.
- Integration of the systems developed in data lake hospital taking into account the scalability and usability of the generated Artificial Intelligence models.