The public sector faces a stiff challenge to ensures its data supports user centred services, writes Tom McCann, director of Sopra Steria
Public services are all about making people’s lives better, and they run on data from a myriad of sources and systems; but too often the quality and construction of the data is not up to the job.
It raises a major question for the public sector in how it can develop a human centred approach to data, ensuring that it is collected, structured and used in a way to help meet the needs of individuals.
This provided the focus of a UKA Live discussion with contributions from myself, Kate Boyle, head of data services at the Police Digital Service, Charlie Boundy, chief data officer of Companies House, Giuseppe Sollazzo, head of data products and services at DWP Digital, and Richard Irvine, chief data officer of Leeds City Council and NHS West Yorkshire Integrated Care Board.
It was based on an understanding that human centred data involves the crucial attributes and details to ensure services can be designed with individual recipients at the centre of how they are delivered. It also provides the scope for a proactive approach to identifying people who need support and sometimes taking action that is preventative rather than reactive.
Complexity and change
This can be relevant to services such as healthcare, protecting the vulnerable and keeping people safe, underpinning their design and supporting public servants in their roles. But it is complicated by the rapid pace of change in the demand for and delivery of services, and organisations are struggling to establish how to achieve the human focus in their data.
As yet there are no definitive answers, but the discussion produced some important insights into making progress.
One – a point made by Sollazzo – is to remember that the quality of data usually reflects the use case for which it is collected. This means it will not always have the attributes for an additional use or may be structured in a way that is not suitable to a different context. Any effort has to begin with an understanding of why the data was originally created and the way in which it can be used.
It also has to be classified effectively according to the new purpose. Boyle pointed out that in the case of vulnerable people, local government, healthcare bodies and police forces will have their own classifications that may be difficult to match, and there is a case for a more consistent definition of different states of vulnerability.
Standardised elements
This comes with a need for more standardised elements to help align the data from different sources. Some progress has been made with the adoption by local government and some other services of the unique property reference number (UPRN) – an identifier for each addressable location in Great Britain – although Irvine pointed out that it has not been widely taken up by the NHS.
Similarly, the health service has a unique identifier for individuals in the form of the NHS number, but this is not shared with other services.
Boundy said it would help for organisations to have a broad picture of where relevant data lays within the public sector, and pointed to the ongoing development of the Data Marketplace as a step towards making this possible.
But any effort to provide identifiers to link data is likely to run into concerns about privacy, and one of the big challenges is to find the right balance in enabling public sector bodies to identify people for legitimate purposes while safeguarding against unnecessary intrusion. On a broad front this is likely to involve transparency about how data is used and a clear ethical framework to undperpin public trust.
Another factor is that, while the public sector now has great experience in quantitative data, it needs to make more use of qualitative data as this provides a stronger understanding of the circumstances of individuals and trends within communities. Technologies such as natural language processing and other types of AI provide great potential, but again they come with a need for safeguards to protect privacy and prevent their misuse.
Data literacy
There is also a need for a good level of data literacy among public servants to understand how to collect and use it to best effect. Most organisations recognise this, but their employees are under continual pressure in their everyday jobs and many will need to be given time and support to develop the relevant skills.
The legal framework is also highly pertinent, and there is currently an unanswered question about how many public servants understand what they will be able to do under the Data Protection and Digital Information Bill – currently going through Parliament – and its imperative to share data to protect vulnerable people.
Such factors will influence the capacity for technology to use the data effectively, and it is becoming more complex with the increasing importance of machine learning and AI. They can do much to target and tailor support for individuals but rely on high quality data that conveys the context and necessary details.
This is an important ambition for the public sector, but it has to come with an acknowledgement that the complexity and pace of change makes it impossible to plan for ever scenario. For the short to medium term there has to be a focus on finding practical solutions for specific challenges.
The discussion produced the suggestion that organisations with similar user journeys should form clusters to understand the details and develop shared approaches. This could lead to positive combined outcomes and, as Boyle described it, “broaden the art of the possible”.
It ties in with the approach of starting small and looking for quick wins, small examples of how human centred data can be brought together across organisations and domains. This will lay the ground for long term success in using data to improve lives and build a better society.
You can download and watch the recording of the UKA Live discussion here:
