Health team

Building technologies that can make a difference

Overview

Launched in 2016, DeepMind Health is dedicated to helping clinicians make quicker and more accurate decisions through digital tools that improve health outcomes and give them more time to care for patients.

We have two separate strands of work – the immediate development and deployment of our mobile assistant for clinicians, Streams, and a set of long-term AI research projects with world-leading partners.

Google Health

In November 2018, we announced that the team behind Streams, as well as those working on translational research, will join Google Health under the leadership of  Dr David Feinberg.

The enhanced scale, resource, and expertise offered by our colleagues at Google will accelerate our work, helping us continue building technologies that can make a significant difference in the lives of millions of patients and clinicians around the world.

Streams

Each year, thousands of people die in hospitals from preventable conditions, like sepsis and acute kidney injury, because the warning signs aren't picked up and acted on in time. We built Streams in close collaboration with nurses and doctors to help address this problem.

At the moment, the app doesn't use AI. It brings together important medical information from a range of existing hospital IT systems in one place, like blood test results or vital signs. This allows clinicians to identify patients who are deteriorating and to take early action. However, our intention is to develop Streams into an AI-powered assistant for nurses and doctors everywhere – combining the best predictive algorithms with intuitive design to predict a range of conditions. To get there, our team of expert engineers, clinicians, and researchers will be working collaboratively with brilliant colleagues from Google to make this vision a reality. 

The first version of Streams has been in use at the Royal Free London NHS Foundation Trust since 2016, where doctors are using the app to identify acute kidney injury. We’re delighted that the early anecdotal feedback from nurses, doctors, and patients has been really positive, something that was backed up in our recent peer-reviewed service evaluation.

Streams is also in use at Imperial College Healthcare NHS Trust. At the Trust, healthcare professionals can use Streams to access important information about their patients’ health, such as test results, while on the move. We have also signed partnerships with Taunton and Somerset NHS Foundation Trust and Yeovil District Hospital NHS Foundation Trust.

Streams FAQ

Does Streams use artificial intelligence (AI)?

Streams doesn’t currently use AI. Right now, we’re focusing on getting the right test results to the right nurse or doctor via a secure mobile app. This is an essential first step before any more advanced technology like AI can be introduced. This first step can also bring enormous patient benefits. We’re already hearing stories of people whose care has been helped by Streams, and nurses say that the technology is already saving them two hours each day.

In the future, Streams will develop into an AI-powered mobile assistant. Not only will it provide clinicians with the information they need at their fingertips, but it will also predict those patients at risk of acute deterioration using machine-learning technology. We aspire to help clinicians prevent serious illness, not just react to it.

Research

Detecting eye disease with Moorfields Eye Hospital NHS Foundation Trust

Our work with Moorfields began in mid-2016 and the first results were published two years later in the journal Nature Medicine. They demonstrated that our technology can automatically detect eye conditions in seconds and prioritise patients in urgent need of care, matching the accuracy of doctors with over 20 years experience. It also addressed the “black box” problem by explaining its decisions to clinicians through visual representation and a percentage recommendation. Our next phase of work is focused on predicting eye disease before symptoms set in.

Planning radiotherapy for head and neck cancer with UCLH NHS Foundation Trust

Our partnership with University College London Hospitals  began in 2016 and is focussed on exploring the potential benefits that AI technology could have in planning treatment for patients suffering from head and neck cancers. In September 2018, we published early results that suggest we can develop an AI system that can analyse and segment medical scans of head and neck cancer to a similar standard as expert clinicians. It can also do this in a fraction of the time of traditional methods.

Predicting patient deterioration with the US Department of Veterans Affairs

Our work with the US Department of Veterans Affairs, which began in 2018, explores whether we can use AI to predict patient deterioration. Our project is initially focussing on predicting Acute Kidney Injury (AKI), a disease that contributes to 100,000 deaths in hospitals per year in the UK, and over 250,000 in the US. The next phase of work will be an evaluation of these AI algorithms to test how they might perform in a clinical environment.

Spotting breast cancer with Cancer Research UK Imperial Centre and Jikei University Hospital:

We began working with the Cancer Research UK Centre at Imperial College London in 2017. The focus of our work is to see if our machine learning technology could help improve the detection of breast cancer - a disease that affects two million people every year and claims the lives of 500,000 people around the world. In 2018, we announced that The Jikei University Hospital in Japan will join the research consortium.

Security

The security of patient data is, and will always be, our top priority throughout all of our work.

We only use patient data to help improve care under the explicit instructions of our partners. Each of our partnerships have strict rules about who can access data and how it can be used, in accordance with the UK and European laws and GDPR regulations. We will never use patient data outside of these rules.

All data is held to world-leading standards of security and encryption, and our systems have passed multiple NHS audits. They have been built by some of the world’s leading security experts. Identifiable data is fully encrypted, stored in a high-security facility, and separated at all times from any other systems. All access to the data is logged and accessible to our partners.

Data and governance FAQs

What regulations cover your work?

We will only use patient data to help improve care, under the instructions of our partners, who will continue to act as data controllers at all times. Each of our partnerships have strict rules about who can access data and how data can be used, and our work is governed by UK and EU laws and information governance regulations. We will never use patient data outside of these rules.