Health data spaces
Federated architectures, integration with European nodes and technical data governance, from the clinical data dictionary and the ontologies (SNOMED CT, LOINC) to the infrastructure where the data lives.
Data science
We interpret clinical and molecular data with machine learning to anticipate and decide better.
Machine learning applied to clinical and molecular data makes it possible to anticipate how a disease, care demand or a waiting list will evolve, identify which variable explains it and measure the effect of acting on it from day one.
The clinical hypothesis, validated from the outset
Every project starts from the clinical hypothesis: what data, from what sources, how they relate to one another and how much each weighs on the decision.
We validate it with the professional before building, so that the result meets the clinical expectation from the first version.
2031 is already on the calendar
The European Health Data Space Regulation has been in force since March 2025 and applies progressively from March 2027 through to 2031. It establishes health data access bodies, secure processing environments and access rights for patients and researchers. Carrying it into the day-to-day of healthcare organisations calls for decisions on architecture, governance and operating model.
We have taken part in Spain's first health data space projects and we operate one of the nodes of a European omics data space.
What we do
Health data office
Strategic definition, operating model, data sources and the economic model that sustains the space over time, for administrations structuring clinical data governance.
Predictive models and epidemiology
Surveillance, population projections and predictive models validated on real clinical data, to anticipate how a disease or care demand will evolve.
Clinical and operational intelligence
Dashboards that show what has happened, explain why, check whether the decision worked and recommend what to do next.
What we bring on the data
Molecular data, as well as clinical
Adding molecular data to demographic and administrative variables multiplies the level of detail, allows new questions to be asked and reveals relationships that until now went unnoticed.
The variable that really matters
Clinical judgement decides which variables enter the hypothesis and the model measures how much each weighs on the real outcome. The scenario is recalculated as soon as a condition changes.
Acting on the cause
We tell apart the variables that signal a problem from those that explain it, simulate the effect of acting on each one and measure the result from day one.
New applications
The same capabilities open up new fields, such as clinical image analysis with computer vision or the genomic stratification of patients in clinical trials.
How we do it
Medical and technical training in the same analyst
Our analyst discusses the what-for and the why of the analysis with the client as competently as they build it.
We model with the clinician
We compare dozens of statistical strategies in shared working sessions, free to use the most effective tool in each case, and generate the code once the model is validated.
More than five models in six months
That is our commitment, against the single model a conventional project usually delivers in a year.
Data science projects
Multi-omic clinical system for controlled longevity
Personalised health Data science
Biomechanical digital twin and genetic characterisation in rehabilitation
Personalised health Data science Clinical processes and governed AI
Psychoeducational mediators and auditable machine learning
Personalised health Data science Clinical processes and governed AI
Epigenetic modulation of PRRSV susceptibility in pigs
Bioinformatics services Data science
Training in generative AI and machine learning
Bioinformatics services Data science Clinical processes and governed AI
Machine learning models in production
Data science Clinical processes and governed AI