Where are we now? The use of administrative data to inform children’s social care research

This blog, written by the ADR England Research Community Catalyst: Children at risk of poor outcomes team, outlines key insights from recent research into how administrative data can be better used to support children’s social care (CSC). Drawing on three new reports launched in January, the authors highlight current evidence gaps, emerging priorities, and opportunities to strengthen the research infrastructure that informs policy and practice for children and families.

There is consensus that we need to do more to prevent poor outcomes for children. Part of ‘doing more’ requires a better understanding of both evolving risks to children and what kinds of timely help lead to better outcomes. Researchers and analysts can contribute to this agenda by generating new knowledge and evaluating a range of child- and family-focused interventions.

One type of data that is particularly useful for generating new knowledge is data produced routinely by organisations directly involved with children and families – we call this kind of data  administrative data. This data is particularly useful when it can be linked across sectors such as health, education, welfare, and justice, and when combined with demographic information. When made available to researchers in anonymised formats, it enables important questions to be asked about children’s needs, their interaction with services, their pathways through these services, short- and long-term outcomes, and how these vary depending on where they live.

To accelerate the use of this valuable large-scale data, ADR England, together with Foundations: The National What Works Centre for Children and Families, has invested in a team of leading researchers to establish a Community Catalyst. This initiative aims to build capacity among a wide and diverse network of researchers and analysts. The Community Catalyst focuses on administrative data that can help us better understand the needs of children at risk of poor outcomes and identify effective solutions.

What we don’t know yet – and why it matters

To date, we are not making sufficient use of children’s social care data, despite a substantial population of children in need. Concerns about the national children’s data infrastructure were firmly recognised by the recent Independent Review of Children’s Social Care. Several developments are underway in response, including the Department for Education’s publication of a data strategy (2023).

In January, our Catalyst team, led by Professors Lucy Griffiths and Karen Broadhurst, launched three new reports designed to help researchers understand evidence gaps and priorities for new research in the context of evolving risks to children. This work identified key research agenda themes. Below, we outline suggestions for how researchers can advance work in these areas, and where data challenges remain:

1. Early intervention and support for children in need: Tracking service pathways

There are significant gaps in our knowledge about children in need and those subject to child protection. This mismatch contrasts with the policy emphasis on early intervention, prevention, and family support. We need a clearer view of children’s longitudinal trajectories through children’s social care services (CSC), health, education, and criminal justice services to understand what support they receive. This is not easy to capture, however, ADR UK flagship datasets allow for critical linkage across data sources. In particular, the Education and Child Health Outcomes from Linked Data (ECHILD) dataset holds considerable promise for advancing our knowledge about children at risk.

2. Understanding household-level hardship beyond area deprivation

Beyond area-level measures of deprivation, there is a need for research that focuses on household-level hardship. Free school meals (FSM) can be used as a proxy for household deprivation, as eligibility is based on benefit claims. Another approach involves using data from the Census 2021, which includes an indicator of household deprivation across four dimensions: employment, education, health and disability, and housing. However, this does not directly reflect a household’s income or wealth.

Details of family income and background are not yet available through ADR UK’s administrative datasets. A promising development is the DfE’s Pupil Parent Matched Dataset, expected to feature in future iterations of the Longitudinal Education Outcomes (LEO) dataset. This links parents’ tax records to their children’s educational outcomes via data held by the Department for Work and Pensions (DWP).

3. Whole-family approaches to mental health

Most mental health research focuses on either children or parents, but few studies have explored the burden of mental health at the family or household level. This is likely due to challenges in linking individuals within households. Ongoing work by ECHILD to link individuals to residences using the Unique Property Reference Number, and the creation of a new mother-baby cohort (ECHILD-MB), are important steps forward. Hospital records in ECHILD offer some insights, but research is limited by the lack of linkable national GP or community mental health service data.

4. Linking Special Educational Needs and Disabilities (SEND) and long-term outcomes

Research often uses the National Pupil Database to identify children with SEND, helping to understand their prevalence relative to the general population. However, significant gaps remain in understanding how SEND affects outcomes such as school disengagement and post-16 employment. Administrative data from across different sectors, such as the linkable health and education data available in ECHILD, could offer a more coherent picture of how children access and move between services. Work is also underway to link local authority CSC data with primary care data in England.

5. Spotting signs of disengagement from education

Data gaps make it difficult to assess whether children at risk are engaged in education. Proxy measures like absenteeism, exclusions, or attendance help, but don’t provide a full picture. Moreover, there is no national data collection on children not in school. Although it is possible to identify children who disappear from school rolls using the National Pupil Database, it is impossible to determine the reason for their disappearance. We welcome the Children Not in School Bill, which proposes a mandatory local authority register of children not in school.

6. Understanding services and workforce capacity

Administrative data can help us understand the cost, effectiveness, and workforce of children’s social services. However, current data provides limited insight into what makes a service effective from a cross-sector perspective. Existing resources include the DfE’s annual local authority expenditure statistics and the children’s social care workforce survey. While these may provide some useful insights, new in-depth data from service providers, local authorities and agencies is needed. We also recognise the lack of service information collected within the Children in Need census. The focus is on referrals and periods of being a child in need, but not the type, frequency, or duration of services provided.

7. Dual system involvement: Criminal justice and welfare

The Ministry of Justice and Department for Education (MoJ-DfE) linked dataset offers new opportunities to explore dual system involvement. It includes data on prisons, courts, Police National Computer, children looked after and children in need (National Pupil Database). This linkage enables analysis of relationships between childhood characteristics, education, and offending.

8. Addressing domestic violence and abuse (DVA) through better data

DVA continues to be a neglected topic within the administrative data community due to sparse and fragmented information. The MoJ-DfE dataset includes Police National Computer extracts that could be used for further study. However, the sensitive nature of DVA can make it difficult to obtain accurate information. Crime data only gives a partial view of DVA because it is often underreported. Administrative data from various service providers, such as emergency departments, women’s refuges and accommodation services, could offer a more comprehensive view of victims' experiences.

9. Experiences of ethnic minority groups

Ethnicity data in the Children in Need and Children Looked After censuses is limited to five broad categories, which can obscure important within-group differences. Where possible, we recommend researchers use more granular ethnic classifications, such as those in the census or ECHILD. Religion is not recorded in CSC data, and while the census collects it, for younger children this may only reflect the parent’s beliefs.

10. Capturing risks beyond the family home

Data on extra familial harms (risks that arise in the community or peer group such as sexual or criminal exploitation) is currently limited. Traditional social work has focused on risks within the family home, but broader influences also matter. The Children in Need census does not yet include key data items related to extra familial harms. Data linkages across services and agencies are needed to build a more accurate picture.

11. Evaluating what works in policy and practice

Using administrative data to evaluate policies and practices is vital to ensure strategies are evidence-based, impactful, and aligned with the needs of children and families. These research priorities aim to drive meaningful improvements in CSC services and outcomes.

The road ahead: Unlocking the full potential of administrative data

Many of the knowledge gaps identified here align with the DfE’s Areas of Research Interest (2025). Within the context of the Opportunity Mission, the DfE highlights the need to better understand children’s service journeys, the long-term impact of early intervention, and the effectiveness of multi-agency responses to vulnerability and risk - core themes that underly the work of this Community Catalyst. Key research priorities where further evidence is encouraged include SEND, family support, health and wellbeing, educational attainment and attendance, and school engagement. Addressing these gaps will enable researchers to contribute to national strategy and deliver evidence-based insights for better policy and service delivery.

ADR UK continues to support this work by investing in data access and research infrastructure. Significant progress has been made in making administrative data accessible to researchers across the UK via trusted research environments. These datasets are vital for exploring why and which children become involved with CSC. However, several challenges remain:

• Limitations to linking datasets due to the lack of personal identifiers
• Specific to UK CSC research, insights are limited to school-aged children due to identifier availability
• Inconsistent data recording due to poorly standardised definitions and coding
• Underuse of regional and local datasets, which offer more granular insight
• Lack of data on the quality and nature of service inputs - an urgent issue following the Independent Review of Children’s Social Care’s recommendations
• Need for better integration of qualitative evidence to support co-produced research

Find out more

You can read more about the full and summary findings. The stakeholder consultation and research agenda can also be explored.

Team involved: Professor Lucy Griffiths, Swansea University; Professor Karen Broadhurst, Lancaster University; Professor Katie Harron, University College London; Professor Jenny Woodman, University College London; Dr Dougal Hargreaves, Imperial College London; Professor Lisa Holmes, Sussex University; Dr Grace Bailey, Swansea University; Kat Tranter, Lancaster University; Dr Eliazar Luna, University College London; Marie Greaves, Swansea University.