Big data and machine learning in critical care: Opportunities for collaborative research.

Area of Science:

• Critical Care Medicine
• Health Informatics
• Data Science

Background:

• Clinical Information Systems (CIS) generate vast amounts of data in Intensive Care Units (ICUs), offering potential for improved patient outcomes and resource allocation.
• Clinicians often lack the expertise for large-scale data analysis, facing challenges with patient privacy, consent, and data quality.
• Multidisciplinary collaboration is essential to overcome these hurdles and leverage ICU data effectively.

Framework:

• A Critical Care Datathon was organized to foster collaboration between clinicians, data scientists, and legal experts.
• The event utilized the anonymized MIMIC III database to address predefined clinical research questions.
• Collaboration aimed to bridge the gap between clinical needs and data analysis capabilities.

Implementation:

• Eight multidisciplinary groups worked on distinct clinical research questions during the Datathon.
• Participants included clinicians, data engineers, machine-learning experts, statisticians, epidemiologists, information scientists, and lawyers.
• The event facilitated hands-on experience with big data analysis tools in a critical care context.

Implications:

• The Datathon successfully generated analyses for posed clinical research questions, identifying future research opportunities.
• It established a foundation for the future utilization of ICU databases in Spain.
• The event highlighted the significant potential of big data analysis tools in critical care and set a precedent for future collaborative meetings.