DeepMind AI Predicts Kidney Disease 48 Hours in Advance: A Breakthrough in Early Diagnosis

Acute Kidney Injury (AKI) and AI Innovation

Acute Kidney Injury (AKI) claims the lives of 500,000 individuals in the US and 100,000 in the UK each year, largely due to late detection. In an effort to change this alarming trend, researchers are leveraging artificial intelligence. DeepMind, an AI research company owned by Alphabet, has collaborated with the US Department of Veterans Affairs (VA) to create an AKI prediction algorithm.

In a study published today in Nature, the team revealed their algorithm can forecast the onset of AKI up to 48 hours in advance. Remarkably, the model accurately identified 90% of patients whose conditions were likely to deteriorate to the point of requiring dialysis.

Additionally, DeepMind's mobile app, Streams, was assessed by researchers at University College London. The findings indicated that specialists could review urgent cases within 15 minutes through the app, a significant improvement compared to the several hours required previously. Furthermore, Streams reduced missed AKI cases from 12.4% to just 3.3%. This tool also contributed to healthcare cost savings, reinforcing its value in clinical settings.

With the combination of the new AKI detection algorithm and the Streams app, DeepMind aims to shift healthcare from a reactive model to a proactive prevention strategy. The initiative doesn't end with AKI; the DeepMind team intends to tackle other diseases responsible for millions of preventable deaths annually. This work will continue as part of Google Health, a partnership dedicated to making a global impact.