A real-time integrated framework to support clinical decision making for covid-19 patients.

Area of Science:

• Health Informatics
• Artificial Intelligence in Medicine
• Epidemiology

Background:

• The COVID-19 pandemic strained global healthcare systems.
• Artificial Intelligence (AI) predictive models require timely, standardized patient data for improved management.
• A robust data architecture is crucial for effective pandemic response.

Purpose of the Study:

• To detail the architecture and technologies of a COVID-19 Data Mart.
• To demonstrate a hospital's approach to managing the COVID-19 pandemic using integrated data.
• To establish a foundation for AI-driven predictive studies and real-time clinical decision support.

Main Methods:

• Developed an information technology framework using SAS and open-source R/Python.
• Extracted and integrated clinical variables from real-world patient data.
• Created a data mart for retrospective analysis and a real-time dashboard.

Main Results:

• The Data Mart includes 5528 SARS-CoV-2 infected patients.
• Demographic and clinical factors associated with mortality were identified (age, comorbidities, dyspnea, d-dimer, CRP, urea nitrogen).
• A multi-level dashboard was implemented for hospital, ward, and individual patient management.

Conclusions:

• An AI-based COVID-19 Data Mart was successfully developed and integrated into clinical practice.
• The Data Mart facilitates machine learning for phenotype discovery and predictive modeling.
• A real-time dashboard enhances understanding of COVID-19 epidemiology and clinical features, aiding future pandemic preparedness.