Abstract
Patients readmitted to an intensive care unit (ICU) during the same hospitalization have an increased risk of death, length of stay, and are associated with higher costs. Previous studies have demonstrated overall readmission rates of 4-14%, of which nearly a third can be attributed to premature discharge from the ICU. In this project, a prediction algorithm based on intelligent modeling techniques is proposed, considering numerical and textual data. Both sources of information can be complimentary and will be used in a decision support system framework. The specific aims of the proposed research are: identify important features that lead to unfavorable or favorable clinical conditions for readmissions of patients in ICU; design specific decision models that will support clinicians? decisions in terms of suitability of releasing a patient from the ICU, in order to achieve more favorable clinical outcomes.
This project will use data from the Multi-parameter Intelligent Monitoring for Intensive Care (MIMIC II) database. This is a large database of ICU patients admitted to the Beth Israel Deaconess Medical Center, collected from 2001 to 2006. The MIMIC II database is currently formed by 26,655 patients, of which 19,075 are adults. It includes high frequency sampled data of bedside monitors, clinical data (laboratory tests, physicians’ and nurses’ notes, imaging reports, medications and other input/output events related to the patient) and demographic data.