A common longitudinal intensive care unit data format (CLIF) for critical illness research

Area of Science:

• Critical care medicine
• Health informatics
• Data science

Background:

• Millions of lives are impacted annually by critical illness.
• Electronic health record (EHR) data offer valuable insights into critical illness.
• Existing data barriers hinder large-scale EHR studies in critical care.

Purpose of the Study:

• To overcome data management, security, and standardization challenges in critical illness EHR research.
• To develop a harmonized data format for critical care studies.
• To demonstrate the utility of a federated research architecture for critical illness data.

Main Methods:

• Developed the Common Longitudinal Intensive Care Unit (ICU) data Format (CLIF), an open-source database format.
• Utilized a federated research architecture for data analysis.
• Conducted external validation of an in-hospital mortality prediction model.
• Assessed 72-hour temperature trajectories and their association with outcomes using group-based trajectory models.

Main Results:

• Harmonized longitudinal EHR data from 111,440 critically ill patients across 9 health systems into CLIF databases.
• The mortality prediction model showed variable performance across sites (AUCs: 0.73-0.81).
• Similar temperature trajectories were observed across health systems, with hypothermia and hyperthermia linked to higher mortality.

Conclusions:

• The CLIF format facilitates transparent, efficient, and reproducible critical care research.
• Federated case studies highlight CLIF's potential for disease sub-phenotyping and decision-support evaluation.
• Future applications include EHR-based trials, AI model development, and real-time quality dashboards.