This article is from the source 'guardian' and was first published or seen on . It last changed over 40 days ago and won't be checked again for changes.

You can find the current article at its original source at https://www.theguardian.com/technology/2016/may/06/deepmind-best-privacy-infrastructure-handling-nhs-data-says-co-founder

The article has changed 5 times. There is an RSS feed of changes available.

Version 0 Version 1
DeepMind has best privacy infrastructure for handling NHS data, says co-founder DeepMind has best privacy infrastructure for handling NHS data, says co-founder
(about 5 hours later)
Google’s DeepMind has hit back at criticism of its partnership with London’s Royal Free hospital to develop an app that helps doctors and nurses rapidly identify and treat acute kidney injuries.Google’s DeepMind has hit back at criticism of its partnership with London’s Royal Free hospital to develop an app that helps doctors and nurses rapidly identify and treat acute kidney injuries.
DeepMind’s co-founder, Mustafa Suleyman, said the company was better placed than any other to handle sensitive medical data, given its long history of securing highly personal information from other fields. He said: “As Googlers, we have the very best privacy and secure infrastructure for managing the most sensitive data in the world. That’s something we’re able to draw upon as we’re such a core part of Google.”DeepMind’s co-founder, Mustafa Suleyman, said the company was better placed than any other to handle sensitive medical data, given its long history of securing highly personal information from other fields. He said: “As Googlers, we have the very best privacy and secure infrastructure for managing the most sensitive data in the world. That’s something we’re able to draw upon as we’re such a core part of Google.”
Related: Google's DeepMind shouldn't suck up our NHS records in secretRelated: Google's DeepMind shouldn't suck up our NHS records in secret
He added: “When we developed our information governance toolkit and we submitted that for assessment to the health and social care information centre (HSCIC), which approves these data-sharing agreements, we got 100% for our toolkit. There’s pretty much nobody else who’s been able to get a score as high as that.”He added: “When we developed our information governance toolkit and we submitted that for assessment to the health and social care information centre (HSCIC), which approves these data-sharing agreements, we got 100% for our toolkit. There’s pretty much nobody else who’s been able to get a score as high as that.”
Suleyman said DeepMind’s information governance processes were awarded a level three certification by HSCIC, the highest possible. “The vast majority of other organisations in the country only have a level two, including most of the hospitals.”Suleyman said DeepMind’s information governance processes were awarded a level three certification by HSCIC, the highest possible. “The vast majority of other organisations in the country only have a level two, including most of the hospitals.”
Suleyman was speaking in the wake of a row over tDeepMind’s data-sharing agreement with the London’s Royal Free hospital, published by the New Scientist last week. In the document, the hospital agreed to share five years of historical data on patients, as well as real-time information on their status, which alarmed patient groups. Suleyman was speaking in the wake of a row over DeepMind’s data-sharing agreement with the London’s Royal Free hospital, published by the New Scientist last week. In the document, the hospital agreed to share five years of historical data on patients, as well as real-time information on their status, which alarmed patient groups.
Acute kidney injuries kill 1,000 people a month, because of the difficulty in identifying problems and acting on them quickly enough. Suleyman said the timescale for an intervention is a matter of hours, but if the injury is caught early, “then with really simple and really cheap things like intravenous fluid and antibiotics, you have the potential to save that person’s life”.Acute kidney injuries kill 1,000 people a month, because of the difficulty in identifying problems and acting on them quickly enough. Suleyman said the timescale for an intervention is a matter of hours, but if the injury is caught early, “then with really simple and really cheap things like intravenous fluid and antibiotics, you have the potential to save that person’s life”.
The data-sharing agreement prevents Google from using the information in any other part of its business. It also requires DeepMind to delete its own copy when the agreement expires in September 2017. But the breadth and depth of the data being shared caused some to worry that the company had intentions to use it beyond the stated aims. Sam Smith, of the health data privacy group MedConfidential, said: “This is not just about kidney function. They’re getting the full data.”The data-sharing agreement prevents Google from using the information in any other part of its business. It also requires DeepMind to delete its own copy when the agreement expires in September 2017. But the breadth and depth of the data being shared caused some to worry that the company had intentions to use it beyond the stated aims. Sam Smith, of the health data privacy group MedConfidential, said: “This is not just about kidney function. They’re getting the full data.”
DeepMind said the app at the heart of the data-sharing agreement, Streams, requires the data set to make accurate and timely calls about patients potentially suffering acute kidney injuries. The data is broad, covering conditions seemingly unrelated to kidney function, because kidney injuries are both more common and more serious in patients with multiple comorbidities. And it is deep, going back five years, because historical trends are important for identifying patients who are most at risk. A patient who has had an elevated creatinine level for half a decade requires different treatment to one who is suffering a recent spike.DeepMind said the app at the heart of the data-sharing agreement, Streams, requires the data set to make accurate and timely calls about patients potentially suffering acute kidney injuries. The data is broad, covering conditions seemingly unrelated to kidney function, because kidney injuries are both more common and more serious in patients with multiple comorbidities. And it is deep, going back five years, because historical trends are important for identifying patients who are most at risk. A patient who has had an elevated creatinine level for half a decade requires different treatment to one who is suffering a recent spike.
While DeepMind is known for its focus on artificial intelligence and machine learning, Streams doesn’t incorporate those fields. Instead, the heart of the app’s diagnostic function is based on an algorithm mandated by NHS England. DeepMind is one of many companies building technology that pulls in patient data, processes it according to the algorithm and delivers alerts to doctors and nurses.While DeepMind is known for its focus on artificial intelligence and machine learning, Streams doesn’t incorporate those fields. Instead, the heart of the app’s diagnostic function is based on an algorithm mandated by NHS England. DeepMind is one of many companies building technology that pulls in patient data, processes it according to the algorithm and delivers alerts to doctors and nurses.
Suleyman said DeepMind’s app is different to other systems in how it communicates information to doctors and nurses. Rather than the pager and paper model that some hospitals use, or off-the-shelf software created foraccounting or logistics industries, DeepMind has “done what we know best as technologists, which is to ask the user what they want”, Suleyman said.Suleyman said DeepMind’s app is different to other systems in how it communicates information to doctors and nurses. Rather than the pager and paper model that some hospitals use, or off-the-shelf software created foraccounting or logistics industries, DeepMind has “done what we know best as technologists, which is to ask the user what they want”, Suleyman said.
He added: “As a result, the reaction that we’ve got from people is incredible. Everybody is like ‘I’ve never seen anything like this before, this is so intuitive, this feels like one of my cutting-edge apps in the app store.’”He added: “As a result, the reaction that we’ve got from people is incredible. Everybody is like ‘I’ve never seen anything like this before, this is so intuitive, this feels like one of my cutting-edge apps in the app store.’”
Because Stream doesn’t use any of the machine learning or AI technology DeepMind is famous for, it might feel an odd fit at the company but Suleyman says “in the long term, both of those options are potentially possible”.Because Stream doesn’t use any of the machine learning or AI technology DeepMind is famous for, it might feel an odd fit at the company but Suleyman says “in the long term, both of those options are potentially possible”.
DeepMind is also trying to counter concerns over its motivations with a panel of independent reviewers, which will meet for the first time in June to asses the data-sharing agreements and security measures put in place for its health projects.DeepMind is also trying to counter concerns over its motivations with a panel of independent reviewers, which will meet for the first time in June to asses the data-sharing agreements and security measures put in place for its health projects.
Unlike the AI ethics board, set up when DeepMind was bought by Google, the health review panel is public. It includes the editor of the Lancet medical journal Richard Horton, the NHS’s former “Kidney tsar” Professor Donal O’Donaghue, and the chair of Tech City UK Eileen Burbidge.Unlike the AI ethics board, set up when DeepMind was bought by Google, the health review panel is public. It includes the editor of the Lancet medical journal Richard Horton, the NHS’s former “Kidney tsar” Professor Donal O’Donaghue, and the chair of Tech City UK Eileen Burbidge.
It will issue an annual report outlining its findings, and Suleyman says: “They’re not going to be contracted, they’re not going to be paid, and they’re going to be free to speak publicly about what we’re doing. I’m really proud to be able to say that and to be able to open ourselves up for scrutiny proactively.”It will issue an annual report outlining its findings, and Suleyman says: “They’re not going to be contracted, they’re not going to be paid, and they’re going to be free to speak publicly about what we’re doing. I’m really proud to be able to say that and to be able to open ourselves up for scrutiny proactively.”