Common Data Model for Children's Services Statutory Returns
The problem: Every day, our Children’s Services Departments (CSDs) make key decisions deciding the future of vulnerable children without timely and relevant comparable data. Without this, they cannot accurately assess what works and what doesn’t or make well-informed commissioning decisions. Statutory returns processes specifically are very inefficient and cost-intensive. Although each of the 153 CSDs in England faces this problems, they’ve not been able to solve it . Each CSD spends upwards of three months creating comparable datasets (SSDA903 and CiN Census) on vulnerable children. The great amount of time and money (~£50k/authority) that this demands means these datasets are prepared only once yearly. Nationally, the cost is ~£7.6m, equal to the total annual pay of hundreds of analysts.
Our objectives: We want to improve this process by giving CSDs access to comparable, data on children in care (eg, needs, services, outcomes, etc) which will in turn enable them to make better, more appropriate commissioning decisions so that vulnerable children achieve the best outcomes possible. Our discovery will focus on developing a standardized reporting approach for the SSDA903 statutory returns, and identifying what the unmet data needs stemming from the current process are. We will develop a deep understanding of user needs (UNs) and the existing statutory returns processes, before designing a potential solution to increase efficiency and improve child outcomes.
To achieve our goals, we will work with Social Finance (a non-profit with 11 years’ government data experience). Our agile and iterative approach will follow GDS Service Standards. We will check our progress against our objectives with internal “check and challenge” sessions and face-to-face meetings with LAs such as show and tell sessions.
Inception (Sprint 1):
- Kick-off meeting with LAs to identify key objectives,
- Draw up a roadmap and research questions
- Identify users to be interviewed.
User Research (Sprint 2):
- Conduct interviews and workshops with service users and providers.
- Map stat returns requests against current reporting capabilities
- Identify detailed user personas summarising the needs and pain points for generalised roles, and a
- Identify a longlist of UNs, and potential ways to improve efficiency and outcomes.
- Submit data request and conduct data diagnostic
Validation (Sprint 3):
- Test UNs through desktop analysis and workshops
- Get feedback to validate and prioritise UNs.
- Create user journeys and epics to further evidence our findings.
Solution Design (Sprint 3):
- Work with LAs on solutions for the prioritised UNs.
User Testing and Feasibility Assessment (Sprint 4):
- Hold workshops with users to test technical feasibility of potential solutions
- Select the final options to develop a business case for prototype.
- Decide what a solution would look like, what data it could draw on, and analyse its technical requirements.
- Align IG so that we have the data required for alpha
Polished Outputs (Sprint 3-4):
- Prepare a business case explaining the current cost of the problem and the potential for local and national savings.
- Produce a user research report summarising findings from the user research. Write a concluding report detailing a plan for alpha.
- Publish outputs via Github (com/SFDigiLabs).
Our discovery will last 8 weeks so that the outputs are delivered by the end of March 2019.
Current cost: Preparing comparable datasets on vulnerable children currently costs each LA around £50k, amounting to £7.6million nationally. Budgets for CSDs are being cut, yet demand is going up.
Impact: Giving CSDs in GM access to the data they need to complete statutory returns quickly and in appropriate formats will enable better operational, commissioning, and strategic decisions. Other LAs have successfully implemented common data models, that have created positive social and financial impact. The Greater London Authority has used Social Finance’s Edge of Care Analysis Module to conduct analysis required for commissioning decisions 7x faster to identify children most in need of support. This means 450 children will access services and stay with their families, which is expected to deliver savings of £21.6m over 7 years. However, this was built on annual stat. return data and does not support the more regular decisions CSDs want to make using data, which we hope to do.
Efficiency gains: More efficient, standardised stat. returns processes could make significant savings by speeding up analysts’ work. A trial of the Analyst Workbench (co-designed with Social Finance) in GM showed that data analysis can take minutes instead of hours, so analysts could produce key charts upwards of 15x faster. This fundamentally changes the role of analysts, so they save time on reporting, which can be utilized to deliver insights on what causes certain trends to better inform key decisions. We hope to realize similar efficiency gains.
Better strategy: Improving communication and links between LAs will enable them to establish common metrics to better understand ‘what works’ across different practice models, co-commission services, align strategies, and pool data to leverage advanced data science techniques.
We believe this project could make million-pound national savings and enable appropriate evidence-backed, fast decisions to be made in the support and care of vulnerable children
Steering committee: GMCA is an active participant of a working group of 12 UK LAs who are discussing how to make better use of data in their commissioning and service delivery. GMCA has provided feedback on discovery work that has been done in other LAs, including two that are part of GM. We will leverage this working group to give input and feedback via workshops, ensuring our work is relevant to the wider market. Other LAs have already shown interest in working with us, namely Islington and Surrey.
Sharing outputs: The working group provides a pipeline of potential alpha partners, allowing us to quickly test any solutions developed at scale. Sharing data and best practice in this group will ensure our learnings from both the technology development and service delivery sides go beyond only our project. We will make these learnings available to other councils by publishing materials and research on GitHub and other online forums, including Twitter and the LocalGov Digital Slack thread
We will complete our user research in 4 weeks, leaving a month to test potential solutions and draw up and submit the following outputs by the end of March 2019.
Business case: Our business case will give a detailed baseline explaining the current cost of the problem to LAs both in GM and nationally, and the potential for savings in both cases. It will assess inefficiencies in CSDs’ statutory returns processes and the resulting costs, look at potential savings to be made via faster analysis, and conduct Cost Benefit Analysis (CBA) of the outcomes for vulnerable children. It will also assess monetary and outcomes benefits of sharing best practice, digital tools and data across LAs and services. We will engage with budget holders early on to standardise the CBA output across LAs so that we agree what is needed to make key decisions.
User research report: From our agile and iterative user research process we will develop user personas, user journeys, user stories, epics, and a longlist of UNs, all of which will be tested and validated in workshops with users. These will form the body of the user research report and inform the business case and concluding proposal. Our user research report will highlight the key UNs identified and discuss how they compare across LAs both within and outside of the working group. We will also detail the potential solutions developed over a series of iterations.
Concluding proposal: If there is a clear business case, we will present a proposal for alpha to assess the impact and feasibility of a common data model or other solution. This would include a list of LAs and key people who would work on the alpha version, a rough estimation of the costs expected, a preliminary journey map, technical requirements, dependencies, prerequisites, a discussion of the potential for upscaling the data model or service, and a more detailed idea of how the sharing of best practice and digital tools between LAs would be improved
The key user groups that we have identified for user research in discovery are:
- Data and systems leads – including data analysts, IT leads, data owners, and application support. These users are key for testing technical feasibility, data extraction processes, and integration with existing systems (e.g. APIs).
- Information governance leads – including IG lead and SRO. These users are key for understanding data, and decision-making process within the Council, as well as data protection procedures and relevant statutory gateways
- Operational leaders and managers – These users are key and understanding key decision points and the information needed to take better decisions.
We will use a mix of workshops, and semi-structured interviews to conduct our user research. We will use these interviews and workshops to develop a list of pain points, and user needs, which we will feedback and iterate with our steering groups. We will take a similar approach with solution design.
Our proposed user research objectives are:
- Understand what data is currently available to CSDs through a data diagnostic
- Understand what leadership within these teams want in order to make more informed decisions
- Understand the key challenges in the current statutory returns processes, including how individual LAs are adapting the return process to meet their own reporting and business intelligence needs
- Produce a list of key, prioritised leadership UNs
- Understand what current solutions are available
- Understand how we can improve data processing with a common data model
- Assess the feasibility of meeting the UNs with the current data available
- Identify a minimum viable common data model
- Design potential solutions
We would welcome the support of the Local Digital Collaboration Unit to:
- Help further share results
- Engage with more LAs
- Provide check and challenge on process findings
- Engage with GDS
GMCA has not been granted funding for this project in the past and has not previously applied for any funding to fund this project