Exploratory Analysis of Electronic Intensive Care Unit (eICU) Database

The monitoring of severely ill patients is a crucial procedure for every intensive care unit (ICU). By applying different data exploration methods on monitoring data, some perspective can be gained. In the present research, such monitoring data were explored in the electronic ICU (eICU) Collaborative Research Database--an ICU database collected from more than 200 hospitals and over 139,000 ICU patients across the United States. The eICU database, with its enormous quantity of remote monitoring data, could be a great resource for extracting insightful information that can help to identify potential areas of improvement in the quality of patient treatment. Important information such as patients vital signs, care plan documentation, stage of illness, diagnosis, and treatment is available in the database. In the present study, we explore the distribution of the data, including demographics, conditions, and diseases, and identify important patterns and relationships between features of the data. Through an exploratory analysis of the data, including the relationships between gender, ethnicity, diseases, and quality of care and mortality rates, remarkable insights were obtained. To the best of our knowledge, this is the first comprehensive exploratory analysis of the eICU database. A deep understanding of the ICU data provides the foundation for further predictive and prescriptive analyses of the data with the ultimate goal of improving ICU treatment procedures for future patients.