Offer Description

Severity and mortality in critically injured ICU patients depend on complex, dynamic physiological processes that current scoring systems capture only partially. Most existing models rely on static, admission-time variables and offer limited insight into the underlying biological mechanisms driving recovery or deterioration, which restricts both their predictive accuracy and their clinical interpretability.

This thesis offer aims to characterize the physiological trajectories of trauma patients admitted to the ICU by identifying biomarker profiles associated with distinct clinical courses and outcomes. Using multicenter, prospective data from the national RETRAUCI registry, dynamic clinical and laboratory variables collected throughout ICU stay will be structured, analyzed, and integrated into a longitudinal dataset (RETRASeries). Artificial intelligence techniques - including unsupervised clustering and pattern-recognition methods - will be applied to detect recurring physiological signatures and biomarker combinations linked to survival, complications, and recovery patterns.

By moving beyond single-timepoint predictors toward a dynamic, data-driven understanding of trauma physiology, this work seeks to identify clinically meaningful biomarker profiles that can inform risk stratification and support the explainability of predictive models developed within the broader project (RETRALogic). The findings are expected to contribute to more personalized, mechanism-informed severity assessment, improving communication with clinical teams and families and supporting future research on modifiable risk factors in trauma outcomes.

Where to apply Website

Requirements





Research Field Computer science Education Level Master Degree or equivalent

Skills/Qualifications

- Programming in R (data analysis, statistical modeling)
- Experience working with clinical/health data (structured datasets, ideally ICU or hospital registries)
- Strong background in statistics or biostatistics
- Fluent written and spoken English
- Master's degree in a relevant field (e.g., Biomedical Engineering, Data Science, Statistics, Medicine, Bioinformatics, Computer Science)
- Ability to work with large, multicenter datasets

Specific Requirements

- Experience with machine learning / artificial intelligence techniques (e.g., clustering, classification, explainable AI)
- Knowledge of Python
- Familiarity with fuzzy logic or explainable AI methods
- Experience with electronic health records (EHR) or medical registries
- Knowledge of critical care / trauma medicine terminology and clinical workflows
- Experience with data visualization or dashboard development (e.g., Shiny, web platforms)
- Spanish or Catalan language skills (useful for working with clinical teams and registry data in Spain, though not always mandatory)
- Prior publications or research experience in biomedical data science
- Experience with reproducible research tools (Git, R Markdown, etc.)

Languages ENGLISH Level Good

Additional Information

Website for additional job details

Work Location(s)

Number of offers available 1 Company/Institute Universitat de Lleida Country Spain State/Province Catalunya City Lleida Postal Code 25198 Street Av. Alcalde Rovira Roure, 80 Geofield

Contact State/Province

Lleida City

Lleida Website

Street

Av. Alcalde Rovira Roure, 80 Postal Code

25198 E-Mail

[email protected]

[email protected] Phone

+34 973 702441

STATUS: EXPIRED

Share this page

📌 AI-Driven Biomarker Profiling for Understanding Trauma Physiology in ICU Patients (Lérida)
🏢 Reconocida empresa
📍 Lérida