An app called Streams, built with Google’s DeepMind, can look for signs of acute kidney injury and identify the condition within 14 minutes.

The Problem

Kidney injuries kill around 100,000 in the UK alone every year. Fatal kidney diseases, such as AKI, involve sudden change or decreased blood flow to the kidneys that is normally the result of another illness. At least one in three AKI deaths may be preventable if doctors can intervene earlier. However, present detection methods are slow, taking several hours with reliance on NHS’ limited technology offerings. A dire problem, up to one in five may fall to AKI. The condition costs the NHS £1.2 billion annually.

What is DeepMind’s Streams?

Streams was developed by DeepMind (now owned by Google).

An AI-powered, secure mobile phone app that aims to address what clinicians call “failure to rescue,” which is when the right nurse or doctor doesn’t get to the right patient in time. In the UK alone, many thousands die annually due to this problem—when conditions such as sepsis and acute kidney injury (AKI) strike as the warning sides are not located.

How  did they Design Streams?

DeepMind/Google report they designed in close collaboration with nurses and doctors to help address key problems. The app brings together important medical information, like patients’ blood test results, chart data and much more, all in one place, supporting clinicians in real time.

First deployed at Royal Free London NHS

Royal Free has adopted Streams to send secure alerts and support clinicians’ review of their patients’ test results. Nurses report that Streams is saving them up to two hours each day. 


Royal Free is using Streams to detect AKI within 14 minutes—this represents a real path-breaking technology. Presently, with existing methods it can take several hours to detect AKI. Based on a recent study published in the journal Nature Digital Medicine, Streams cut the costs by around £2000 per hospital patient with AKI—from £11, 772 to £9,761. In various studies, Streams can detect 96.7% of cases considered an emergency—compared to 87.6% with current methods.

DeepMind (Google)

Google acquired DeepMind for $520 million in 2014. The technology has previously raised privacy concerns. Based in London, it shares operations with U.S-based Google’s Health unit. Google states that the health data of DeepMind remains not in effect. There are some concerns out there—just in 2017, a deal between DeepMind and the Royal Free NHS Foundation Trust was found to misuse patient data. Apparently Royal Free didn’t comply with the Data Projection Act.

Source: BrinkWire

Pin It on Pinterest