Find out more

Why does this matter?

Some children face a higher chance of experiencing difficulties with their health, education or development. These difficulties can be linked to experiences such as poverty, parental health problems, domestic abuse, disability, or involvement with children's social care. Many children who need extra help are not included in national longitudinal studies, so we know less about how their experiences affect them as they grow up. Following children over time could help researchers understand what happens before, during and after contact with services, and what helps children and families to achieve better outcomes.

What did we want to find out?

We wanted to know whether information already collected by public services could be used to identify children who may be at greater risk of poor outcomes. We also wanted to understand which approaches could be used to recruit these children and their families into a new longitudinal birth cohort study.

What we did

We used data from ECHILD alongside similar data from Scotland and Wales to look at children aged 0–5 who were involved with children's social care, including children in need, children subject to child protection arrangements and children who were looked after. We also linked mothers to their children and looked at maternal health, education, social care involvement and area-level deprivation. We looked at maternal psychosocial health factors such as recorded anxiety and depression, substance use, self-harm and domestic violence or abuse in the year before birth. We also examined maternal education indicators such as free school meal eligibility, special educational needs provision and GCSE attainment.

What we found

Children's social care: Large numbers of young children are involved with children's social care, but the pattern varies between UK nations. Infants were more likely to become looked after, while children aged 4–5 were more likely to be recorded as children in need or receiving care and support.

Maternal health: Mothers with at least one recorded psychosocial health risk factor in the year before birth were more likely to have a child who later became involved with children's social care.

Maternal education: Indicators of educational disadvantage also identified groups whose children were more likely to become involved with children's social care. For example, in England, almost 1 in 5 mothers who had been eligible for free school meals, received special educational needs provision, or had not achieved five A*–C GCSEs had a child who became involved with children's social care before age six, compared with around 1 in 10 for other mothers.

Recruitment: National administrative health records could potentially be used to identify families for recruitment, but some particularly sensitive cases would be excluded for ethical and legal reasons. Recruiting directly from national children's social care records was considered much more difficult. Local authorities may be better placed to approach families because they have existing relationships with children and parents, although this approach would be more expensive and may be harder to make nationally representative.

What this means, and what's next

Administrative data could be an important part of a future longitudinal study, helping researchers include children who might otherwise be missed. A future study would need strong partnerships with data providers, local authorities and trusted professionals, as well as careful involvement of children and families in the design of recruitment materials and methods.

Introduction: 

The study assessed the feasibility of using routinely collected administrative data to define and sample a target population for a new longitudinal birth cohort of children at risk of poor outcomes. Two broad sampling strategies were considered: (1) identifying children directly through involvement with children's social care (CSC); and (2) identifying children at elevated risk through characteristics recorded for their mothers.

Methods: 

Children aged 0–5 with CSC involvement were examined for entry into and movement through different levels of support. For the maternal sampling strategy, we used the mother-baby link in ECHILD to examine maternal health, education, social care involvement and area-level deprivation. Maternal psychosocial risk factors included mental health conditions, substance use, self-harm and domestic violence/abuse recorded in hospital records in the year before delivery.

ECHILD data used:

HES: Admitted Patient Care records (2002-2020)

CSC: Children in Need (2009-2021) and Child Looked After (2006-2021)

NPD: SEN Spring Census (2006-2020) and KS4 attainment (2004-2019)

Mother-baby link

Results:

Across England, substantial numbers of children aged 0–5 were involved with CSC. Entry patterns differed by age and intervention type. Infants had relatively high rates of entry to care, while older preschool children were more often recorded as children in need or receiving care and support. Escalation generally occurred sequentially rather than directly from the lowest to the highest level of intervention.

Among mothers giving birth in England between 2009/10 and 2015/16, 21% of those who had at least one recorded psychosocial risk factor in the year before delivery had a child who subsequently received a CSC intervention before age six. The corresponding proportion among mothers without a recorded health-related risk factor was 8%.

17% of mothers eligible for free school meals, 16% with special educational needs provision and 17% who did not achieve five A*–C GCSEs had a child who became involved with CSC before age six, compared with 11%, 12% and 9%, respectively, among comparison groups.

Conclusions: Administrative data can support the identification and sampling of children at increased risk of poor outcomes. Maternal health, education and deprivation indicators provide potential markers for targeted sampling before a child's own involvement with services. However, direct recruitment from national CSC records is substantially constrained by confidentiality, ethical, legal and practical considerations. Local authority-based approaches may offer a more feasible route for recruiting children already involved with services, although they are more resource-intensive and may be less nationally representative.

A future UK-wide cohort would require harmonisation across nations, clear governance and lawful bases for data use, careful ethical consideration of recruitment from sensitive records, and co-design with families and people with lived experience.

‍Full report: Scoping the feasibility of a new longitudinal birth cohort study of children at risk of poor  outcomes across the UK (2025).