Predicting Acute Kidney Injury Using DeepMind’s New AI

DeepMind researchers have created a machine learning algorithm that is capable of detecting Acute Kidney Injury (AKI) in hospital patients up to 48 hours before its occurrence. In 90% of the worst cases identified when patients resulted in requiring dialysis, the algorithm’s predictions were accurate. The AI system was trained on over 620,000 distinct data points, and it identified 3,600 of them that were good predictors of AKI. DeepMind’s clinical lead, Dr Dominic King, hopes that using AI to predict patient deterioration would allow clinicians to intervene much sooner to prevent AKI, a condition linked to 100,000 deaths a year in the UK and over 1.4 million deaths worldwide. In the case of AKI, rehydration, antibiotics or altering medications can help restore patients’ kidney function fairly easily. DeepMind Health is set to become part of Google, under the oversight of Google Health, suggesting that there will be a greater emphasis to commercialize its creations – not a huge surprise considering recent reports of £470M losses and £1.03bn debt in 2018 at the company.