BUSINESS INSIGHTS : The latest news, analysis, and trends about protection and care

Nov 29, 2018,  by Allianz Partners Business Insights

Kidney injury: AI capable of accurately estimating risks

Among its many potential areas of use, artificial intelligence has much to offer the medical world. Developed by US researchers, new AI can predict the appearance of acute kidney injury in patients in intensive care. Results have been found to be 50% accurate, a higher rate than that achieved by existing methods and one that should improve as AI integrates and analyses new patient data.


Thanks to artificial intelligence, clinicians can now anticipate the risk of acute kidney injury (AKI) within the first 24 hours following intensive care. A team of researchers from the USA’s Northwestern University and the University of Texas Health Science Center has developed AI capable of calculating the probability of critically ill patients developing AKI, mainly on the basis of clinical notes, as reported by the website Presse-Citron.


Anticipating AKI


AKI is a condition in which the kidneys suddenly lose their ability to filter waste from the blood. The mortality rate can be as high as 89% if the illness progresses beyond stage 2 (out of a total of three stages). According to the researchers behind the study, which was published on 9 November, the risk of death increases 12-fold if AKI develops after major abdominal surgery.

The AI they have developed can extract and analyse risk factors from the electronic health records of patients in intensive care and then predict the likelihood of AKI appearing. This innovative system can alert doctors to such risks as soon as patients are admitted, thus saving precious time.


40,000 patients


The researchers trained their AI using the anonymous health records of 40,000 patients who stayed in the intensive care unit (ICU) at the Beth Israel Deaconess Medical Center. All these patients had kidney dysfunction. The team took into account the age, gender and ethnicity of the patients as well as clinical notes taken during their ICU admission and 72-hour serum creatinine levels.

According to the test results, 50% of cases where there was a risk of AKI appearing were detected within 24 hours, a level of accuracy far greater than that achieved by existing methods, as reported by Venture Beat. The researchers believe the system’s  performances will only improve with the inclusion of new medical files. With AI also capable of diagnosing early-stage Alzheimer’s and detecting skin cancers, it is fast becoming a valuable aid in specialist medical care.


Allianz Partners

Linked Topics
About Allianz Partners Business Insights