A machine learning model developed at the University of California Davis was able to quickly and accurately predict acute kidney injury in burn patients, a common complication occurring in around 30 percent of severe burn cases.
In a study published in the journal Burns, the model was used to analyze routine biomarkers typically associated with acute kidney injury, such as urine output and serum/plasma creatinine, as well as neutrophil gelatinase associated lipocalin (NGAL), which has previously not been used to diagnose post-burn kidney injury in the U.S. and thus required a more complex algorithm to interpret.
The model achieved 90-100 percent accuracy in predicting kidney injury in the 50 adult burn patients whose data was analyzed in the study, and did so, on average, within 18.8 hours of admission. In comparison, traditional models that do not analyze NGAL achieve only 80-90 percent accuracy and take nearly a full day longer, 42.7 hours, to predict kidney injury.
More articles about AI:
Memorial Sloan Kettering AI model spots cancer in pathology slides with almost 100 percent accuracy
6 updates in AI ethics: new guidelines, institutes & more
Beth Israel Deaconess' open-source patient database enabled a decade of AI research