February 9, 2023

The use of quantitative data in children’s social care is still in its infancy when compared with other sectors, and this immaturity means that developments in the various methodologies designed to act as an aid to human decision making, and indeed the understanding of whether our interventions work, is suboptimal. The Nuffield Foundation funded Coram to carry out a small scoping project in order to answer the question ‘can data help keep children safe?’ We held an event in December 2018 to share the products of that project.

At the event we showcased a proof of concept that we had developed with the Turing Institute and Kent County Council (see video below). It shows children (as dots) moving through the front door activity of a local authority’s children’s social care service – the node a dot is moving towards (e.g. Referral) indicates the activity taking place and the speed at which a dot moves towards that node represents how long it took to complete that activity. A thematic indicator was added to represent the number of times an individual child had been through the system within the period of data observed (almost four years), with dots starting as white for those going through for the first time in this period, gradating to a deep red as they increasingly experience the system. As it was a proof of concept, this was not an accurate reflection of the service – for example, cases that progressed to Early Help were shown exiting via No Further Action (NFA); and cases that started and completed an activity on the same day cannot be shown in the visualisation, which is why we do not see many cases moving between contact and referral.

When the audience saw several dots getting darker they imagined the distress experienced by vulnerable families repeatedly undergoing assessments. Visualising the data in this way also had an immediate impact on the directors in Kent and led to them looking into how these situations could be avoided.

The emphatic response of our audience to this visualisation encouraged us to run a follow up small-scale project, again funded by the Nuffield Foundation, to build a working prototype tool that would actually take an LA’s data and visualise it in the way demonstrated by this proof of concept.

About the project

The project was an exploratory and developmental undertaking. Our aims were to:

• develop a working prototype;
• explore whether it could be used by several LAs, and what would prevent them from using it; and
• identify what insights such a visualisation could provide to LAs and whether this would improve interventions for children.

We enlisted the support of Timo Hannay, who runs SchoolDash and is a pioneer in analysing and presenting education data, to build the prototype; and partnered with three LAs (East Sussex, Kent and Oxfordshire County Councils) to help us design and test it.

The project started in January 2020 and was originally meant to last a year. However the Covid-19 pandemic and resultant shifting in priorities for Coram and its LA partners led to a delay in completing this project.

How did we do?

We are delighted to say that with Timo’s coding expertise we successfully built a working prototype of the visualisation tool (link), and developed a data processing tool (in Excel) to transform an LA’s data into the format needed by the visualisation tool. Feel free to take a look and try them out. The data processing tool cleans, transforms and anonymises LA data, as well as categorises and segments children into cohorts. The visualisation tool has options for adjusting the speed as well as pausing the presentation, plus the ability to select a child’s dot and follow it through the visualisation.

Due to concerns about data protection, the tool is browser-based and designed so that no data needs to leave an LA’s IT network. The data it uses is based on the Ofsted Annex A return, with the only additional data required being for the outcome of each activity (e.g. did a referral result in a child and family assessment or a Strategy Discussion). This means that, in theory, any LA can use the tool as it uses data that they already collect and produce. However, things were not as straightforward as we had hoped.

What did we learn?

This project aimed to explore whether we could develop re-usable approaches to presenting data in children’s social care that would provide more insight to professionals than traditional methods. However, we have found that individualised approaches are required in the short term, to help develop capability and capacity within LAs and to increase their appetite for this type of resource. Alongside an independent evaluation (report) that was conducted on this project, our main learning points were that:

1. Each of our LA partners operates in a slightly different way, meaning that they required their own versions of the visualisation map and of the data processing tool. If we wanted all LAs eventually to use the tool, this could mean that a different map and data processing tool would need to be created for every LA.
2. Variation in the technological and data maturity of LAs may hinder the use of tools like the one developed in this project. While the project’s LAs had already allocated significant resources to quality assure their data, this project still uncovered some issues and a significant amount of time was spent trying to resolve them. For less well-resourced LAs their data may initially be unusable, however data that is used regularly is more likely to be accurate, so tools like this one can help improve data quality. Technological literacy is also key – the end user needs to be able to operate the tool correctly and properly interpret what it is displaying.
3. There is a mistrust among LAs of tools not created ‘in-house’, even of tools built by other LAs. Many LAs are even reluctant to use powerful open source (and hence free) software like Python or PowerBI to wrangle and analyse data. This is why our data processing tool was built in Excel, despite its limitations. In the end, even though our LA partners had helped us design the data processing tool and visualisation, they still had concerns about them as they had not built the tools themselves.
4. Information governance policies can unnecessarily stymie any innovation in this area. We avoided creating a situation in which our LA partners would have to share their data with us, as getting agreement to that would have signficantly extended the timeframe of the project. This meant trying to build a data processing tool without any real data available to us, and trying to imitate any issues the LAs identified during their testing. It also meant that we could not explore the data available ourselves to see what insights might be gleaned.

What has happened as a result of this project?

The learning from this project has already been applied to many different areas. Firstly our LA partners have been inspired by this project to learn more about how they can present data and make it more accessible to their operational staff. They have since developed their own visualisations and looked at strengthening the capability of practitioners so that they record accurate data and can correctly interpret data analysis. Renuka Jeyarajah-Dent, the project’s Principal Investigator, spoke to colleagues in local authorities and captured some initial ideas for how we might immediately encourage better use of data in children’s social care.

More broadly, we have sought to encourage a transformation in the use of data in children’s social care, meeting with the Cabinet Office’s chief data scientist to discuss how we this can be done; and feeding our learning into both the National Child Safeguarding Practice Panel and the Independent Review of Children’s Social Care, with the two recent reports published by these bodies both recognising the importance of data. In response to these, the Government has now published an Implementation Strategy and Consultation: Stable Homes, Built on Love, within which it identifies the six pillars across which they plan to transform children’s social care, with Pillar 6 focusing on the use of evidence and data.

The project also enabled us to share our learning with several professionals from a variety of sectors. For example Dr Rob Harland, Consultant in General Adult Psychiatry at South London and Maudsley NHS Foundation Trust, showed us his developmental work on data usage in mental health.

Finally, the project inspired Coram to create the Innovation Incubator, to provide the sector with a space to generate, test and scale radical innovative solutions, and to help build its technological and analytical capacity and capability.

What we think needs to happen next

There are too many children for whom distress is buried under layers of information, either across years, among many people in the same agency or across agencies. Discontinuity of staff is a frequent complaint in all sectors that seek to serve children and waiting lists for specialist services like CAMHS and educational psychologists are long – which in turn means that the poorest ‘without voice’ (i.e. access to professional networks via friends/acquaintances or private payments) are left behind. Data can serve to make visible those children hidden in plain sight.  Many suffer from factors that we know through research pertain to risk and, without gathering evidence relating to this, no predictive analytics is possible in any case. The creation of Integrated Care Systems and the levelling‑up agenda provide opportunities in this arena of work that need to be grasped – what does levelling up mean for children, particularly the most vulnerable?

Children’s Social Care is ripe for investment in technology and making better use of data. Data can be used to transform the way that work is allocated and patterns are analysed – this happens in other sectors that deal with complex information so why not children’s social care? Do we not owe it to vulnerable children to try? Their voice and that of their families are often least heard: we believe that amplifying their voice and highlighting their experiences could be a catalyst for attracting the investment of sufficient resources. Everyone the project team met agreed that we do not do enough with the data we have but where to begin to transform its use remains a challenge.

One area to tackle could be the information governance barriers that exist. Self-imposed barriers to sharing data exist in many organisations charged with keeping our children safe, making the task much more difficult than it should be. There needs to be a better balance between data security and enabling data to be used to help children and families in need. If we had not decided to make the visualisation tool web-based, the project could have been delayed by many months just by the three data protection impact assessments we would have had to complete. These barriers make it incredibly difficult for third parties to support LAs in making the most of the data they hold, or even for LAs to share what they have built with other LAs.

Another could be simply increasing the data and technological literacy of the sector – through Coram’s Innovation Incubator we explore the art of the possible and work with our technology partners to share with our LA members what has already been done elsewhere (for example, via our Innovation Collective). Once the data we collect is shown to be useful there will be the incentive to improve its quality and availability, and open up more options for enabling data to help professionals.

Lastly, we would love to turn our prototype visualisation tool into a finished product. To do this we will need to secure further funding. If we could create a tool that enabled a LA to build their own map, then they could potentially create one that represents their entire children’s social care services and see the full experience of the children they support.

Thank yous

The project has been funded by the Nuffield Foundation, but the views expressed are those of the authors and not necessarily the Foundation. Visit www.nuffieldfoundation.org. In particular we would like to thank Rob Street and Ash Patel for their support throughout the project.

The core project team consisted of Renuka Jeyarajah-Dent, Principal Investigator; Timo Hannay (SchoolDash), Matthew Wagner (Kent County Council) and Kevin Yong (project lead).

We are also grateful for the involvement of our other LA partners, in particular Alastair Lee, East Sussex County Council; and James Carter, Oxfordshire County Council.

If you would like to find out more, please contact us at [email protected].

• Categories:
  Coram-i · Data analysis · Sharing Learning
• Author:
  Kevin Yong