inDemand Stories | ACRA for Intensive Care Units

Sep 15, 2020

ACRA: An effective algorithm that identifies patients susceptible to get complications that cause readmissions in Intensive Care Units

ACRA is one of the healthcare challenges launched by Servicio Murciano de Salud in the Region of Murcia (Spain) as part of the inDemand project. After a successful co-creation process between healthcare professionals from SMS and the company BigML, with the financial support of INFO Murcia and the business support from Ticbiomed, these are the main results:

The need

Unscheduled re-entry of the patient in the ICU is an adverse event that causes an increase in morbidity, mortality and consumption of hospital resources. A recent meta-analysis has reported an average re-entry rate between 4 and 7%, although it can reach up to 14%. The average mortality of patients with unintended readmission in the 72 hours after discharge from the ICU is 33%.

 

The solution

The solution consists of a supervised learning model where each patient with unintended readmission in the 72 hours after discharge from the ICU are predicted as TRUE or FALSE. Additionally, reports with field importance are included (e.g. related complications, oxygen saturation level, text from the discharge summary, among others). The aim is to identify which factors have more influence in readmission helping ICU professionals to better understand what triggers readmission and enabling early detection of risk factors.

 

If you want to know more about the impact of this solution and read first-hand testimonies of the professionals who have participated, download for free the complete ACRA inDemand Story now!

Download the ACRA inDemand Story

This website uses cookies to ensure you get the best experience on our website. More info.

Los ajustes de cookies de esta web están configurados para "permitir cookies" y así ofrecerte la mejor experiencia de navegación posible. Si sigues utilizando esta web sin cambiar tus ajustes de cookies o haces clic en "Aceptar" estarás dando tu consentimiento a esto.

Cerrar