PublicTechnology talks to the think tank’s digital and data specialist Eleonora Harwich about how AI could help eradicate disparity in the quality of healthcare and why GDS needs to focus more on the needs of its colleagues in Whitehall
Credit: PA
With a name that invokes change and improvement, it is fitting that Reform appears to be the think tank with perhaps the keenest focus on technology and digital government.
In the last six months, the non-partisan organisation has published research covering areas such as digital borders, what the increasingly internet-based crime landscape means for the police, and how the government could make best use of blockchain technology. Since September, Reform also employs a dedicated head of digital and technological innovation, Eleonora Harwich, whose role covers the use of data and digital throughout policy and public services.
The fountainhead of public sector digital strategy and transformation initiatives – at least in its conception – is, of course, the Government Digital Service. The Cabinet Office-based organisation has been in operation for six years, and continues to build around its core vision of ‘Government as a Platform’. Existing constituent parts include the Notify messaging tool, the Verify identity-assurance platform, and a range of other service-design templates and toolkits.
If AI could be more accurate than a radiologist, that could reduce the disparity in the quality of care – which is a huge issue
While GDS still has significant room for improvement, Harwich tells PublicTechnology, it deserves recognition for the work it has done in creating and maintaining a common online home for all of the public sector – one that remains a leader on the world stage.
“I think GOV.UK is absolutely amazing,” she says. “When you navigate a government website in France, you can never find information. In comparison [with others], the UK is still doing really, really well.”
One area where GDS could improve is in designing services with a view to making them as easy to implement as they are to use, Harwich says.
“One example is Verify; obviously, for an end user it is a simple platform. But I do not think the back-end procedure has been fully taken into account,” she says. “It is very important to understand the full process – it is not just about the citizen. With Verify, they have a resource that no departments are using.”
In the last couple of years ‘data’ has become perhaps the single biggest buzzword – bigger even than ‘digital’ – among public sector technology and transformation advocates. The government has, more than once, expressed somewhat grandiose ambitions to improve how it uses data to design and implement policy. Harwich says that such ambition now needs to turn to action.
“I think it is very difficult – but they could be doing better. They could move away from sweeping statements like ‘data is the new oil’,” she says. “The Government Transformation Strategy, the UK Digital Strategy – they all have a policy [along the lines of] ‘we are going to make data more accessible’.”
Harwich adds: “I am not saying that is easy – there is the same issue in the commercial sector, and a lot of [public bodies] also have the complexity of starting from scratch. But [the government] needs to create a model that makes sense for sharing data across departments. The way that they are solving that problem now is just creating data registers. But they need to really spend time trying to understand a better way of collecting data in the future. “
A typical example of how a new way of gathering data could improve the efficacy with which it can be put to use is in hospitals, where data gathered by different departments typically “does not follow a patient”, Harwich says.
Open season
There has been some suggestion that, in much the same way as GDS was formed to drive digital policy and initiatives across the rest of central government, Whitehall could benefit from the creation of a dedicated centralised data agency. But Harwich believes that this is unnecessary, and that the requisite expertise and skills are already present in existing organisations including GDS and the Office for National Statistics.
Whoever leads the charge, the focus – in central government and the wider public sector – should be on creating and implementing open standards, she says.
“[They need to ensure] that everyone adheres to open standards, then they would actually be able to do service integration,” she says. “Getting data-sharing right is so difficult, but they have to find a way to make this work. It is a huge job, and it is not short-term.”
Harwich picks out the work being done by NHS Digital to, by 2020, create for each citizen a single healthcare record that is digital, interoperable among all health-service entities, and able to be updated in real time, as an example of the kind of projects that could be delivered in an environment where data is freely shared and governed by open-standards principles. The challenge for the healthcare technology agency is to create a plan that works for a wide range of individual organisations with differing levels of digital maturity, she says.
The government needs to create a model that makes sense for sharing data across departments. The way that they are solving that problem now is just creating data registers
The use of artificial intelligence and automation is another area in which the NHS could potentially be among the first to reap the benefits, Harwich believes.
“If AI could be more accurate than a radiologist, [for example], that could reduce the disparity in the quality of care – which is a huge issue,” she says.
But perhaps the biggest obstacle to adoption of AI and robotics is not technological, but cultural, she adds.
“I fear for the first big accident [caused by AI],” she says. “We have to be honest – it is definitely going to happen, and I really wonder what the backlash to that will be.”
Later this week Reform will be publishing a new report on the potential benefits of pitfalls on the use of AI across the NHS. Look out on PublicTechnology for a write-up of the report’s key findings.
