Unique property reference numbers are a crucial feature in using data to protect vulnerable people, writes Richard Duffield, head of customer insights at GeoPlace
The response to Covid-19 provided a fresh imperative for sharing public sector data to support vulnerable people, producing examples of best practice and shifting the mindset towards maximising the appropriate use of data.
But significant barriers remain, some coming from longstanding anxieties around data protection and the dangers of identifying individuals when it is not necessary. Others relate to the absence of agreed standards for real time access to the data, with no widespread agreement on which attributes are good predictors of vulnerability and uncertainty over how to find and get them in place.
In response, there is a strong case for a standardised approach to the data, with the use of a key identifier to support the linking of datasets while allaying fears around the privacy of individuals.
The issue provided the focus of a recent UKA Live discussion supported by GeoPlace and involving Paul Davidson, chief digital officer of Sedgemoor Council and director of iStand UK, Jamie Tasker, business intelligence innovation manager at Barnsley Council, and Jo Mann, head of risk and intelligence at Humberside Fire and Rescue Service (FRS).
It highlighted the fragmented approach towards vulnerability data, with organisations often setting up datasets for specific purposes in their own formats and arranging their own sharing agreements. This makes it difficult to match up with those from other organisations and extend their use.
There is currently no agreed list of shared indicators for different types of vulnerability or a formal process for how they should be used, which makes it more difficult co-ordinate work in protecting vulnerable people.
Efforts are being made to identify appropriate indicators, and a key point in the discussion is that data teams should not do this independently but work with service teams on finding the factors that are most relevant and building these into datasets in usable formats.
There is also a significant issue in information governance (IG). It can be possible to provide a sound reason for sharing data with an official in another agency to support a specific service, especially if that agency only wants evidence related to a specific vulnerability – such as fuel poverty or fire risk.
But it can be harder to make the case when the purpose is more speculative or for research. This can create a stand-off with IG teams who want evidence that it can help people but that evidence is not yet 100% clear.
Need for standards and infrastructure
This is all points to a need for standards and a more formalised infrastructure for sharing vulnerability data. While there is sentiment in favour of this coming from central government, iStandUK – which involves local authorities, government organisations and other groups – is taking a lead with the SAVVI (Scalable Approach to Vulnerability via Interoperability) programme.
Paul Davidson described it as a move to develop a common process, fitting attributes such as evidence of poverty, school attendance records and criminal offences, to vulnerability scenarios. It is aimed at developing a set of standards and definitions for how these are used, along with a catalogue of attributes. This could create the scope for data sharing, and enabling public authorities to build dashboards and analysis datasets to identify people and cohorts at risk.
Some local authorities have also been pioneers, among them Barnsley Council, which has created a vulnerability index. This took in 26 data sources with a series of weighted flags to emphasise their relative importance, and was used to contact 14,000 households during the pandemic with advice on issues such as shielding and applying for food deliveries. It is now being used for functions such as identifying households at risk of financial vulnerability.
A key feature of the index is the attachment of unique property reference numbers (UPRNs) – unique identifiers for each addressable location in the country – with a data matching process to ensure it is in place for any households where the systems used did not store it. This makes it possible to identify the issues by household rather than individual, and ensures a system-wide mechanism for checking data integrity and preventing duplicate entries.
This all supports the effort to support vulnerable people, but shields their identities from anyone who does not have to deal with them directly.
Other public sector bodies have made great use of UPRNs. Humberside FRS has developed a fire fatality risk index utilising data science algorithms that combine the identifier with attributes such as people who are over 65, live alone and have mobility problems. This has been greatly assisted by a data sharing initiative between the National Fire Chiefs Council and NHS England. The data helps the FRS identify households in which there are people who are at higher risk from fires and is not shared outside the organisation
Such efforts are demonstrating the UPRN's value as a common standard which can be combined with a risk flag of a possible problem within a household, helping to ensure the relevant attributes are only shared with those in appropriate roles.
Loss of momentum
While its use has been mandated for central government, the discussion acknowledged that they are still not used by many public sector bodies, and many that do are not systematically adding them to all relevant datasets, which makes it harder to break down the barriers in sharing vulnerability data. There was a push to attach them more widely during the pandemic, but anecdotal evidence suggests that without an immediate crisis the momentum has been lost.
It raises the prospect of undermining the response to the next crisis. Sharing data to protect vulnerable people will be significantly harder without the UPRN as a household identifier.
This has to be prevented, and there is a need for vision and commitment during non-extreme times to put in place a systematic approach, supported by robust processes and an organisational mindset, to ensure UPRNs are attached to all datasets to which they can apply. Public authorities have to take this step as intrinsic element of ensuring they have good quality data to be used for the public good.
It needs attention to UPRNs every day to ensure that, when they have an important role to play in a crisis, they will be in place to enable a strong response.