icuARM-An ICU Clinical Decision Support System Using Association Rule Mining

摘要: The rapid development of biomedical monitoring technologies has enabled modern intensive care units (ICUs) to gather vast amounts multimodal measurement data about their patients. However, processing large volumes complex in real-time become a big challenge. Together with ICU physicians, we have designed and developed an clinical decision support system icuARM based on associate rule mining (ARM), publicly available research database MIMIC-II (Multi-parameter Intelligent Monitoring Intensive Care II) that contains more than 40,000 records for 30,000+patients. is constructed multiple association rules easy-to-use graphical user interface (GUI) providers perform information the setting. To validate icuARM, investigated associations between patients' conditions such as comorbidities, demographics, medications outcomes length stay. Coagulopathy surfaced most dangerous co-morbidity leads highest possibility (54.1%) prolonged In addition, women who are older 50 years (38.8%) For treatable drugs, suggests medication choice can be optimized patient-specific characteristics. Overall, provide valuable insights physicians tailor patient's treatment his or her status real time.