The Leiden University Medical Center (LUMC) is using artificial intelligence to reduce the chance of re-admission of patients to the Intensive Care Unit. The chance of a readmission of an IC patient is predicted with special AI software. This is done on the basis of available data on IC admissions. If the software proves to work properly, doctors will from now on receive support from the software in making decisions about whether or not to discharge IC patients.
The AI software makes a calculation based on thousands of characteristics of admissions and patients, such as heart rate, blood pressure and diagnosis. This data is then displayed with a risk assessment on a clear dashboard. With the information, risk assessment and the accompanying substantiation, a specialist from the Intensive Care Unit can come to a better and well-considered decision about whether or not to discharge an IC patient. In this way, it is hoped to prevent patients from being discharged too early or too late from the ward. The aim is to optimize the quality of care, the patient experience, the length of stay and the pressure on the IC beds.
A premature discharge comes with great risks, because a readmission often has serious consequences. On the other hand, late discharge results in an unnecessarily long hospital stay, a delayed rehabilitation process, unnecessary psychological complaints and increased bed pressure. The new support could become an indispensable technique for applying more patient-oriented care in the ICU and reducing risks.