Google DeepMind's AI could give doctors a head start on saving lives, study says

Artificial Intelligence can be used to predict acute kidney injury in hospital patients, according to a study DeepMind conducted with the US Department of Veterans Affairs.

Katie Collins Senior European Correspondent
Katie a UK-based news reporter and features writer. Officially, she is CNET's European correspondent, covering tech policy and Big Tech in the EU and UK. Unofficially, she serves as CNET's Taylor Swift correspondent. You can also find her writing about tech for good, ethics and human rights, the climate crisis, robots, travel and digital culture. She was once described a "living synth" by London's Evening Standard for having a microchip injected into her hand.
Katie Collins
2 min read
Magnifying glass examining male internal organs

Acute kidney injury causes around 500,000 deaths annually in the US.

Roy Scott/Getty Images

Artificial intelligence is frequently heralded as the future solution to many of the world's biggest problems, and for many experts developing the technology, health comes top of the list. London-based DeepMind, owned by Google parent company Alphabet, is working on a number of health care-based projects, and on Wednesday it published its latest research showing how doctors may be able to predict a quick-onset, deadly condition to save more patient lives.

As if it's not bad enough being admitted to hospital for one illness or injury, in-patients in medical facilities are also at risk of developing secondary conditions that can pose serious threats to their health. Among them, acute kidney disease claims the lives of 500,000 US patients every year, according to the Center for Disease Control and Prevention. Acute kidney injuries can be deadly, and they pose a real problem for physicians. Not only are they incredibly difficult to detect, but they can also cause patients to deteriorate rapidly. 

But using AI, DeepMind has a solution that could help doctors spot potential kidney injuries 48 hours before they occur, giving them valuable time to get ahead of the problem and potentially allowing them to prevent the condition in up to 30% of patients. In a study published in the journal Nature, DeepMind outlined work it conducted with the US Department of Veterans Affairs in which it used anonymized data to develop machine learning tools that correctly predict nine out of 10 patients who later went on to require dialysis.

In future, DeepMind hopes to combine this technology with its Streams system -- a medical mobile assistant that flags patient deterioration to doctors, as well as enabling communication between clinical teams and the review of medical information that allows them to make more-efficient treatment decisions. DeepMind also announced on Wednesday that researchers at University College London had successfully peer reviewed Streams, which has been used in a local hospital since 2017. They found it saved doctors up to two hours per day, as well as allowing them to review the most urgent cases within 15 minutes.

"These results comprise the building blocks for our long-term vision of preventative health care, helping doctors to intervene in a proactive, rather than reactive, manner," said the company in a blog post.

DeepMind's work on detecting kidney injuries follows similar work the company did last year, in which it used AI to detect over 50 sight-threatening eye conditions. In the long run, it aims to combine these AI detection tools with its Streams system to improve detection and reduce the costs of treating a whole variety of diseases and illnesses.