US researchers have developed an artificially intelligent system that can detect early the risk of developing acute kidney injury in an intensive care patient.
The Artificial Intelligence (AI) was developed by researchers at Northwestern University and the University of Texas Health Science Center to detect acute kidney damage 24 hours in advance in ICU patients. This AI is able to collect and extract data from patient health records to predict the likelihood of developing acute kidney injury in ICU patients.
Critical care patients are more vulnerable to acute kidney injury, a condition by which the kidneys fail to filter out blood waste. They can destroy the kidney system of critically ill patients. The mortality rate can approach 89% if it exceeds stage 2 (out of 3). And if it develops after major abdominal surgery, the risk of death is twelve fold, according to Venturebeat.
To train the machine, the researchers trained the system to automatically learn health data from more than 40,000 patients, taking into account different parameters such as their age and gender. Their creatinine level (renal function indicator) was recorded 72 hours after admission to Beth Israel Deaconess Medical Center in Boston.
As a result, the intelligent system was able to diagnose 50% of cases at risk of having a kidney lesion within 24 hours of the test.
“From a practical point of view, our predictive model could be used to alert clinicians of critically ill patients at high risk of developing acute kidney injury shortly after admission to intensive care unit,” say the researchers. at the Venturebeat site.
Artificial Intelligences are of great interest to researchers. Englishmen are working on an AI capable of diagnosing the loss of motor skills, a symptom of Parkinson’s disease, in just three minutes. In the United States a smart system is able to diagnose diabetic retinopathy with 90% effectiveness.