Google’s DeepMind will deploy Streams at Taunton and Somerset

The artificial intelligence branch of Google has entered into an agreement with another global digital exemplar (GDE) trust to deploy its technology.

DeepMind Health has signed plans with Taunton and Somerset NHS Foundation Trust to implement its alerting app, Streams, over the next five years.

Announced 21 June, the trust said the app will be used by healthcare professionals to alert them to conditions such as acute kidney injury.

The deployment comes as the previous agreement between DeepMind and the Royal Free London NHS Foundation Trust remains under investigation by the Information Commissioner’s Office (ICO). The ICO confirmed on Wednesday that the investigation is “close to conclusion”.

Luke Gompels, joint chief clinical information officer at the trust, said: “Obviously, we have looked at the experience of the Royal Free and we have written to the National Data Guardian with details of how we will use Streams. We have not begun any work with patient data, so we are in a position to take advice.”

“With this app, we are following standard practice, which is to test using only high quality synthetic data. Only during the live pilot, in direct patient care, will clinicians use patient identifiable data.”

Taunton and Somerset use IMS Maxims for its open source electronic patient record, and are the only GDE trust to do so.

“We have chosen to use the Streams app developed by DeepMind because it offers our clinicians improvements in the safety of their patients by making data accessible at patients’ bedside, where they need it most. We have the opportunity to develop the app, working with DeepMind, and our long-term digital partner, IMS Maxims, keeping patients informed and involved”, said Gompels.

Tom Edwards, a consultant surgeon at Musgrove Park Hospital, said that “fast access to information about patients is absolutely crucial for our doctors, nurses, and other clinical staff”.

“Safety alerts will be immensely useful, but it is important to remember that – whatever technology we use – it will still be our highly trained and expert staff who are making decisions about diagnosis, treatment and patient care”, Edwards said.

Last month, it was also revealed that the national data guardian, Dame Fiona Caldicott, had said that DeepMind had received NHS patient data on an inappropriate legal basis.

Dame Fiona says in the letter that she informed Royal Free and DeepMind in December that she “did not believe that when the patient data was shared with Google DeepMind, implied consent for direct care was an appropriate legal basis”.

Royal Free and DeepMind were subject to media scrutiny following an investigation by New Scientist in May last year. Concerns included the transferred five years worth of data covering 1.6 million patients, without their knowledge.

It is only in the last few months that DeepMind has changed its stance on the Royal Free controversy. In May, DeepMind said on its website that the company should have publicised its intentions before starting work with the Royal Free.

“We should also have done more to engage with patients and the public at that time, including by proactively providing information about how patient data would be processed, and the safeguards around it.”

On the latest agreement with the West Country trust, Dominic King, clinical lead at DeepMind Health said, that doctors and nurses are telling the company that Streams is “helping them deliver faster and better care for their patients”.

“[Taunton and Somerset] is well known for its pioneering approach to healthcare technology, so it’s incredibly exciting to be working with the outstanding clinical team there, on the shared goal of improving outcomes for patients.”

The AI company is also due to hold events at the trust, prior to any data processing happening, to inform patients about its work.

The Royal Free deal was also heavily criticised for its lack of transparency by an academic paper in March, published in Health and Technology, where it said “inexcusable” mistakes were made.

The trust and DeepMind quickly hit back saying the research contains major errors.

In a joint statement from the trust and DeepMind it said: “This paper completely misrepresents the reality of how the NHS uses technology to process data. It makes a series of significant factual and analytical errors, assuming that this kind of data agreement is unprecedented.”

The authors of the paper, Julia Powles and Hal Hodson, said in a joint response that “a subsidiary of Google/Alphabet making a strong move into the heart of managing the world’s most significant and intimate health data repository is, indeed, unprecedented”.

Other trusts have not been deterred from working with DeepMind however, and the company has signed with Imperial College Healthcare NHS Trust to deploy streams, alongside research partnerships at University College London Hospitals NHS Foundation Trust, Moorfields Eye Hospital NHS Foundation Trust.

DeepMind have attempted to allay health data transparency fears by creating a new data audit infrastructure and having a panel of independent reviewers look into its work. The panel has yet to publish any findings.

DeepMind is a London-based AI company that Google brought for £360 million in 2014.

UPDATED 21/06/17 to include Gompels’ comments. 

UPDATED 27/07/17 to include Powles and Hodson comments.