A web-based data visualization tool for the MIMIC-II database

Area of Science:

• Medical Informatics
• Data Visualization
• Clinical Research

Background:

• The Medical Information Mart for Intensive Care II (MIMIC-II) database is a valuable resource for researchers.
• Proficiency in MIMIC-II requires SQL knowledge and schema understanding, posing challenges for clinicians and health researchers.
• Limited computer proficiency among some users hinders access to MIMIC-II data.

Purpose of the Study:

• To develop an interactive, web-based data visualization tool for MIMIC-II.
• To enable first-time MIMIC-II users to easily explore the database.
• To lower the barrier to entry for accessing and analyzing intensive care unit data.

Main Methods:

• Developed an interactive, web-based visualization tool.
• Implemented an 'Explore' feature for cohort analysis and variable distribution visualization.
• Implemented a 'Compare' feature for side-by-side cohort analysis.

Main Results:

• The tool provides 'Explore' and 'Compare' functionalities for MIMIC-II data.
• Users can visualize distributions of administrative, demographic, and clinical variables within selected cohorts.
• Experienced researchers can accelerate data exploration and reduce time spent on SQL queries.

Conclusions:

• The tool offers free, convenient preliminary investigation of MIMIC-II data.
• Researchers can quickly learn characteristics of MIMIC-II patients.
• Future work includes adding analytics capabilities and utilizing MIMIC-III data.