Policy paper

Looked-after children data strategy

Published 17 April 2023

Introduction

In March 2022, the government published Inclusive Britain: government response to the Commission on Race and Ethnic Disparities. Inclusive Britain sets out over 70 actions in response to the Commission’s recommendations, grouped under 3 main themes:

• trust and fairness
• opportunity and agency
• inclusion

Together, these actions set out a race and inclusion strategy for Britain.

Inclusive Britain highlights disparities in adoption rates for ethnic minority children and among those who are looked after, and it sets out a number of actions in this area.

Inclusive Britain action 28

Vulnerable children deserve the best that society and the state can offer, and it is important to bring good quality, coherent and comprehensive data to bear in ensuring the best possible outcomes for children.

There are differences in outcomes between children in different ethnic groups. The independent review of children’s social care^{[footnote 1]} noted that, for example:

• there are ethnic disparities in children’s social care intervention in family life that cannot be explained by deprivation alone
• when families are referred to social care, there is variation in the response they receive depending on their ethnicity
• children and families from some ethnic groups may be less likely to receive support
• children from ethnic minority backgrounds tend to enter the care system later and this impacts their experiences of care
• there are disparities in children’s experiences of relationships in care

Alongside the recommendations in the review, which have the potential to begin to address disparities, the review noted there is more work to do to improve the evidence base and understand these disparities and address them.

Action 28 of Inclusive Britain aims to address this:

To improve the existing evidence base, the RDU [Race Disparity Unit] will work with the Department for Education (DfE) and other stakeholders to develop and publish, in 2022, a strategy to improve the quality and availability of ethnicity data and evidence about looked-after children and their routes out of care.

A more detailed understanding of the demographic profile of looked-after children could inform:

• better planning around providing the right permanence options for children, including adoption, special guardianship, long term fostering or a return to parents
• the recruitment and approval of adoptive parents willing to take children who are waiting and special guardians or foster parents who can meet their needs
• the provision of early intervention policies and strategies to support children and families to minimise the risk that the children do not enter or re-enter care

Scope of the strategy

The focus of this strategy is primarily on the care system (and routes out of care), the remit of DfE. This strategy has been prepared by RDU, working with analysts in DfE and The Children and Social Care Secretariat, Coram-i. The document sets out:

• what data is currently available
• an assessment of the quality of available data
• what the internal and external user needs are for data that are met or partially met
• further work priorities

What data is available?

A summary of this section is provided in Annex A.

Children Looked After (CLA) return

DfE’s main source of data relating to looked-after children is the annual Children Looked After (CLA) return (also known as SSDA903).

The information is collected for each child looked after and includes information about their personal characteristics and information about their placement, legal status and their adoption from care, including dates for each stage of the care and adoption process.

When collecting ethnicity, the child is asked to identify his or her ethnicity. However, for children who are not deemed mature enough to have capacity to consent to sharing their personal data with others, the primary carer of the child is asked to identify which of the ethnicity codes most closely matches the ethnicity of the child

The data is quality assured, and local authorities are able to update their historic data annually.

Historical data is refreshed in the statistical releases each year. This is considered to be the most robust source of information on adoption. It is used to produce the Children looked after in England including adoption statistical release and the Adoption Scorecards.

Data quality of the CLA return

The quality of the data collected in the CLA return is generally considered to be good. Details on this are described in the:

• collection guide
• methodology information

Typically, there is a higher level of unknown ethnicity in the current reporting year, reflecting the fact that some children start to be looked after towards the end of the reporting period.

As local authorities amend historical data, the number of children with unknown ethnicity generally reduces in the subsequent years. For example:

• in 2019, there were 620 children reported with unknown ethnicity (40 refusals and 580 information not yet available)
• in 2020, the figure for 2019 was 410 (30 refusals and 380 information not yet available)
• in 2021, the figure for 2019 was 310 (30 refusals and 280 information not yet obtained)
• In 2022, the figure for 2019 was 270 (20 refusals and 250 information not yet obtained)

270 represents 0.3% of the total 2019 CLA population.

Statistics on the numbers of looked-after children, when disaggregated by ethnicity and other variables such as gender, age group and local authority, are often suppressed. This is to prevent the risk of a child’s identity being disclosed.

Sometimes data for more than one year is combined to reduce the amount of suppression. Also separate disaggregations for looked-after children are shown, such as:

• ethnicity by gender
• ethnicity by age

However, combining data or showing separate disaggregations in this way can make it hard to find patterns in the data.

DfE collections and statistics go through continuous improvement, with user engagement and consultation as standard. For example, DfE holds annual focus groups with local authorities and internal colleagues to understand upcoming user needs. It also runs consultations on specific releases. The most recent ones were:

• Children in Need/Children looked After outcomes release, after which DfE added ethnicity breakdowns to the Key Stage 4 data
• a new stability publication, which will include data on placement stability by ethnicity

Adoption and special guardianship (ASG) quarterly data collection

The Children and Social Care Secretariat (delivered by Coram-I) collects and publishes quarterly information about:

• children and prospective and approved adopters in the adoption process
• children leaving care through a special guardianship order (SGO)

The ASG data return is voluntarily collected but is completed by every local authority, regional adoption agency and voluntary adoption agency in England every 3 months. The data collection is delivered by the Children and Social Care Secretariat, Coram-i under contract to DfE.

Data is collected for individual children where a best interest decision is made for adoption. Ethnicity is collected for all children and is recorded as stated by the primary carer or child.

As well as ethnicity, other data collected includes:

• the child’s characteristics – for example gender, date of birth, disability, and whether they are part of a sibling group
• key dates in the process – for example, when the child entered care and when they were matched
• outcomes of the adoption process

Data is collected at aggregate level for children who have left care through an SGO and includes characteristics of the children including age range, gender, disability, ethnicity, and whether they are part of a sibling group. Data is also collected for special guardians, including the relationship to the child and age range of the special guardian.

Data is also collected for all prospective and approved adopters active in the adoption process. As well as ethnicity, this includes:

• the adopter’s characteristics – including gender, date of birth, religion, sexual orientation, disability, relationship status
• whether the adopter has previously adopted
• the key dates – for example when they registered and were approved, and when they were matched to a child or children
• the outcomes of the adoption process – for example the number of children placed for adoption and a date and reason for leaving the adoption process

Some data is collected by both DfE and ASG. For example, date of birth, gender (although not using the same codeset), ethnicity and key dates in the adoption process.

A summary of the information collected through the 2 main data collections is available in the table below.

Data quality of the ASG data

Data checks are undertaken each quarter for child and adopter data returns. There has been a continued increase in the number of data returns that are received, query and error free or resolved.

There is a 100% return rate for each data collection and the data is usually published within 3 months of the end of each quarter. This makes the data the most up-to-date data available on children and adopters in the adoption process. It accurately reflects adoption activity across England.

What are the key data outputs?

Children looked after in England including adoption National Statistics release

These statistics contain national and local authority level estimates for children looked after in England, including numbers of CLA adopted, care leavers and CLA who were missing.

Some data at the national level contains breakdowns by ethnicity. Ethnicity data is also published at local authority level for CLA as at 31 March. They are derived from the SSDA903 return.

See Annex B for more details.

Adoption scorecards

These statistics contain the number of children who were adopted at local authority level. The data is disaggregated by ethnicity. Data shows the amount of time local authorities take to match and place looked-after children with an adoptive family. These scorecards allow local authorities and other adoption agencies to monitor and compare their performance.

The metrics used in the scorecards are derived from both the SSDA903 return and the ASG data. The metrics are not broken down by ethnicity.

Ethnicity and children’s social care

The analysis in this DfE publication was commissioned by the Independent Review of Children’s Social Care and explores the differences by ethnicity of children referred to or in need of support from social care by their:

• characteristics
• needs
• journeys
• outcomes

This includes looked-after children, as well as children on ‘child in need’ plans or on child protection plans. The publication used data from a number of sources, including the CLA return (SSDA903), the 2011 Census and the Children in Need Census.

The publication includes a number of analyses relevant to looked-after children, including:

• ethnic disparities in routes out of care
• ethnic representation of children in care compared to the wider child population
• ethnic differences in conversion rate from referral to being a looked-after child
• age at first entry of by ethnicity and gender
• average time between referral and entering care by ethnicity
• ethnic differences in routes leading to being a looked-after child
• ethnic differences in foster and residential placements
• ethnic differences in locality of placements
• ethnic differences in placement moves

User engagement following the Independent Review

In the Children’s Social Care (CSC) reform Implementation Strategy, DfE committed to publishing a children’s social care data strategy by the end of 2023. DfE have already begun engaging the sector and undertaking activity related to improving the use and collection of CSC data, including:

• proposing indicators (for consultation) as part of a new CSC Dashboard to support learning and understanding progress towards the outcomes in the new CSC National Framework
• undertaking a pilot to trial a new technological approach to streamline data collection processes
• considering DfE data quality and aggregation

Working with local authorities

Adoption data is not regularly published by local authorities. Some local authorities and Regional Adoption Agencies (RAAs) will provide some figures on adoptions in wider reports (such as their annual reports) that might provide some helpful additional background. However, the level of detail varies too widely to be able to draw a coherent national picture.

As part of implementing recommendations from the Independent Review of Children’s Social Care (CSC), DfE is developing a new CSC Dashboard and has committed to publishing a CSC data strategy. DfE is working with the sector on the indicators to include in the Dashboard, and the content for the data strategy, including what data should be captured.

As part of work to improve the use and collection of CSC data, DfE has begun a small-scale trial collecting CSC data directly from participating local authorities and fed back to them through dashboards. Ethnicity is being considered for collection and the data will be able to be linked to other data sources, including ethnicity data.

DfE already produce the following data tools to help influence local authority practice:

• adoption scorecards
• local authority interactive tool
• unrounded, unsuppressed data supplied to local authorities from ‘Get Information About Pupils’
• outcomes data supplied annually to Virtual School Heads and National Consortium for Examination Results (NCER)
• data supplied to Data 2 Insight for their benchmarking tools
• data supplied to Ofsted (including ASG quarterly data and data on Children Looked After by Local Authorities)

DfE works with local authorities to use these tools and intervene where necessary, for example, through RAAs. DfE works closely with RAAs and challenges them regarding their quarterly data and what they can do to target any issues as highlighted.

Additionally, the adoption scorecards are a valuable monitoring tool and allow for comparison of performance across authorities and years. DfE also has regular conversations with the RAA leaders group about data including to monitor progress against the adoption strategy and to target recruitment and matching.

Data quality areas and user needs

A summary of this section is provided in Annex C.

Characteristics

The Ethnicity and children’s social care report presents an analysis of the age profile of children who enter care for the first time, by ethnicity and gender. However, further analysis could be undertaken, especially around the profiles of children entering, leaving and remaining in care. This could meet needs for more data on routes out of care and time spent in care across different ethnic groups.

There is also interest in data on looked-after children with an education, health and care (EHC) plan and children formerly looked after with an EHC plan. This would help inform policy on planning support for the most vulnerable children including post-placement support.

It would also be good to consider the data sources and ensure that they provide as much ethnicity disaggregation as possible and are aligned with the Government Statistical Service (GSS) harmonised standard for ethnicity. Further disaggregation will allow analysis of outcomes for children with some ethnic groups with smaller populations including the detailed mixed, other and other white ethnic groups. More suppression might be required.

Whether and how ethnicity classifications are aligned might be part of wider discussions between RDU, ONS and DfE about harmonised categories.

Details about adopters and guardians

More information on the ethnicity and religion of all prospective and approved adopters, and the religion of the child would be useful. Adopter and family preferences may play a role, so results would need to be interpreted with caution.

Collecting data on the religion of young children would be difficult. For example, children could be seen to have ‘inherited’ the faith of their birth parents but it could be questioned whether it is their faith. An older child may have a faith which might not be the same as their birth parents. There are also further complexities about whether a parent has disclosed their own or the child’s religion. Further work on these complex issues would need to be done to determine whether this could be collected in a way that would provide accurate information before being used in analysis.

It is not currently possible to analyse whether children and their adoptive families have different or the same ethnicities, and religions, even if that data was also collected for children. This can make analysis on interracial adoption difficult. Data is collected on child and adopters’ ethnicities but the child and adoptive family are not linked within the ASG datasets. There is currently no reliable way to match the 2 records. There have been initial discussions with local authorities, RAAs and Voluntary Adoption Agencies about how the data could be matched. For example, this could be done by using  consistent child and adopter reference numbers across all agencies. This would need further investigation to work out how much of a burden this would be for agencies.

As an alternative, it might be possible to produce a chart showing the percentage of children waiting from different ethnic groups by the percentage of adopters that are waiting. The data would not be linked but it would show the mismatch. Further consideration should be given to whether this is robust enough to be published on a regular basis.

Even with linked data, comparing the ethnicity and religion of a child and available adopters can sometimes be difficult. Adoptive families can be made up of one individual adopter or 2 adopters. If the ethnicities (or religions) of the two-person adoptive family are different it becomes problematic when trying to determine the ethnicity and religion of the household and therefore working out if there are enough adoptive families to match with the ethnicity and religion of the children waiting.

Sibling placements

Information about sibling placements would be useful because siblings are considered harder to place. This suggests that information about children with other characteristics, such as disability, might also be useful in this area.

It is not currently possible to analyse the ethnicities of sibling groups using the ASG data. Data is collected about:

• whether a child is part of a sibling group
• the number of children placed, or planned to be placed, for adoption together as a sibling group
• the number of siblings placed, or planned to be placed, for adoption separately from the child

But the data does not link the children in a sibling group. Linking sibling groups would enable analysis of the characteristic makeup of sibling groups including whether the siblings have the same or different ethnicities.

This information is known at a local authority level and there have been initial discussions with local authorities about introducing a sibling identifier reference to the data collection. However, this would require development work to be done with authorities and their systems providers to generate the sibling references and incorporate it into the reporting systems. This would make the addition of this data field extra work for local authorities.

Other reasons for each child leaving care

There are other reasons that children might leave care than through adoption. These include:

• special guardianship orders (SGOs)
• child arrangements orders (CAOs)
• return to birth parents

It would also be important to look at long-term fostering, although this is not a route out of care. This can be a positive outcome if considered the best option for a child.

Current available data suggests that children from Asian, black and any other ethnic groups are more likely to return home or go into independent living, compared with children from white, mixed and unknown ethnic backgrounds. However, a fuller picture of the whole process is needed, and there is interest in knowing whether children from different ethnic backgrounds are more or less likely to leave through a particular route.

DfE holds data on the ethnicity of children who leave care under an SGO or CAO. The ethnicity of children leaving care through a SGO is already published nationally in the CLA statistics release, but at the local authority level there would be too much suppression for the data to be useful.

Durations

DfE has the information to be able to produce different analyses for the child, for example:

• how long different outcomes take
• whether certain processes take longer for ethnic minority children, adopters or guardians

Some data is available to answer these questions but the main issue is the level of detail that users would find valuable.

DfE also has some data on entries into care and ceasing care by ethnic groups.

Outcome indicators

DfE publishes information for a range of key outcome measures for children in need. They cover:

• special education needs
• educational attainment (at key stage 4)
• destinations from school
• absence from school
• permanent exclusions and suspensions from school
• free school meal eligibility
• type of school attended

Key stage 4 data is also published at national level for children who were previously looked after who left care through an adoption, SGO or CAO.

Changes to the publication for the year ending 31 March 2021 include the addition of an ethnicity breakdown for key stage 4 data at national level. DfE is also looking at how to best publish ethnicity breakdowns for future releases. Small numbers are likely to be an issue for some measures and they will need to be suppressed.

A wider range of outcome measures disaggregated by ethnicity for looked-after children would be useful. It would help local authorities to explore whether there are specific aspects of the education of looked-after children that might be improved, and might inform the work of other service providers. For example, information about what happens to each child who leaves care beyond the age of 18 might be important, such as:

• employment
• training
• contact with the criminal justice system
• housing
• healthcare

Certain outcomes in childhood or early life may lead to a cumulative disadvantage over the lifetime. It would be important to know if this is the case for looked-after children or adopted children.

Further explanatory or contextual secondary analysis

Further analysis could explore the relationship between different aspects of looked-after children and their circumstances, to identify possible causality.

For example, the Rees Centre at Oxford published research in 2015 which suggested that special educational needs was a key factor in predicting outcomes for looked-after children. It would be useful to see whether ethnicity is a key predictor in adoptions or SGOs, or whether there are other factors (such as EHC plans) that play a stronger role.

Further secondary analysis might be able to explore the presentation of age-standardised rates alongside unadjusted rates of adoptions. Age-standardised rates are used to allow for comparisons to be made between populations that may contain different overall population sizes and proportions of people of different ages. They are frequently used in health analyses, for example. In the case of adoptions, the use of age-standardised rates could help with comparisons between ethnic groups and over time. The rates would take account of the varying age profiles between different ethnic groups – people in ethnic minority groups tend to be younger on average, for example.

Applying national results to the local authority level

Implicitly throughout this paper it has been noted that, in order to prevent disclosure, the level of suppression required may limit any analysis at local authority level. However, insightful analysis and discussion by experts might demonstrate value in the use of national data to influence local authority practice.

Priorities

DfE will work on the following.

Work to understand the important metrics required for a new CSC Dashboard to help support learning and improvement. This was one of the main data recommendations from the Care Review (consultation began in February 2023).

Understanding what real-time data DfE and sector would benefit from collecting directly from management information systems (ongoing).

Review what data is published that is not currently broken down by ethnicity, particularly focussing on outcomes. For example, since the report of the Commission on Race and Ethnic Disparities, DfE have added ethnicity breakdowns to key stage 4 outcomes data. The plan is to include these for key stage 2. A priority for the upcoming release is to restart the series following its pause during the pandemic so additional breakdowns may likely be after the 2023 update.

Include extra breakdowns on stability, including by ethnicity, in the main looked-after children statistical release (placement stability by ethnicity in November 2022, school stability from Spring 2023, social worker stability from Spring 2025).

Review whether instead of major ethnicity categories (Asian, black, mixed, white, ‘other’), DfE can include breakdowns by more granular categories. This will need to be considered alongside the DfE’s confidentiality policy. For CSC data values fewer than 6 are suppressed and round national figures to the nearest 10, and therefore, would need to consider whether these breakdowns are useful (November 2023).

Explore with Coram-i what further data they collect that does not form part of the regular ASG outputs to be included in their publications. For example, DfE has previously released the breakdown of waiting times for children by ethnicity when asked. Any developments would need to be considered within the capacity of the wider contract DfE has with Coram-i (ongoing).

Annex A: Summary of data sources

This section summarises the 2 main data collections for looked-after children.

Information	Children Looked After (CLA) data collection (SSDA903)	Adoption and Special Guardianship Quarterly Data Collection (ASG Data)
Collecting authority	Department for Education	Children and Social Care Secretariat, Coram-i under contract to DfE
Type of data collected	National and local authority level estimates for children looked after in England, including numbers of CLA adopted, care leavers and CLA who were missing	Information about children and prospective and approved adopters in the adoption process, and children leaving care through a Special Guardianship Order (SGO)
Statutory or voluntary	Statutory	Voluntary
Frequency of collection	Annual	Quarterly
Data supplying organisations	Local authorities	Local authorities, regional adoption agencies and voluntary adoption agencies
Ethnicity collected	For individual children, based on 2001 Census categories, with addition of Traveller of Irish Heritage and Gypsy/Roma	For children and (prospective adopters), based on 2001 Census categories, with addition of Traveller of Irish Heritage and Gypsy/Roma. For aggregate SGO information, aggregate 5 groups (white, Asian or Asian British, black or black British, mixed, ‘other’)

Other data available

Children Looked After (CLA) data collection (SSDA903)

Individual information about children – see full list in Annex B.

Adoption and Special Guardianship Quarterly Data Collection (ASG Data)

Individual information about children and prospective adopters including:

• gender
• date of birth
• disability
• whether children are part of a sibling group
• adopters religion, sexual orientation and relationship status
• whether an adopter has previously adopted
• key dates in the process – for example, when the child entered care and when they were matched
• outcomes of the adoption process

Aggregate level information on all looked-after children who have left care through a SGO including:

• age range
• gender
• disability
• whether they are part of a sibling group

Aggregate level information collected for special guardians including:

• relationship to the child
• age range of the special guardian

Summary of other information presented in this strategy

Children Looked After (CLA) data collection (SSDA903)

Ethnicity categories could be aligned to the GSS harmonised standard.

Level of suppression of data that is necessary can limit analysis for local authorities.

Adoption and Special Guardianship Quarterly Data Collection (ASG Data)

Ethnicity categories could be aligned to the GSS harmonised standard.

There is currently no reliable way to match the child and adopter records. This can make analysis on interracial adoption difficult.

Religion data is not collected (and difficult to collect) for the child.

Adoptive families can be made up of one individual adopter or 2 adopters. This can make analysis more complex.

Children are not linked in sibling groups.

Annex B: Details of the ‘Children looked after in England including adoption’ national statistics

These statistics contain national and local authority data. Some data at the national level contains breakdowns by ethnicity, as indicated.

For children looked after on 31 March:

• characteristics of children looked after (includes ethnicity breakdown)
• reasons for being looked after
• legal status of children looked after
• placements of children looked after
• placement stability (includes ethnicity breakdown)
• locality of placements
• unaccompanied asylum-seeking children (UASC) (includes ethnicity breakdown)

Children starting to be looked after (includes ethnicity breakdown)

Children who were looked after during the year (includes ethnicity breakdown)

Health outcomes for children looked after for at least 12 months on 31 March:

• offending rates
• substance misuse
• health and development outcomes
• emotional and behavioural health

Children looked after who were missing

Children who ceased to be looked after in the year: (includes ethnicity breakdown)
- reasons for ceasing to be looked after - looked-after children who were adopted during the year (including ‘children for whom the local authority has made the decision that the child should be placed for adoption’) (includes ethnicity breakdown) - looked-after children who left care through a special guardianship order (includes ethnicity breakdown) - children for whom the local authority has made the decision that the child should be placed for adoption, those placed and waiting for adoption and children where the decision to be placed for adoption has been reversed (includes ethnicity breakdown)

Former care leavers:

• ‘in touch’
• activity of former care leavers
• accommodation of former care leavers
• accommodation suitability
• ‘staying put’
• former care leavers who were unaccompanied asylum-seeking children

Annex C: Summary of other data quality areas

In addition to the priorities listed in the strategy, other data quality areas that are mentioned in the data needs section include:

• aligning categories to GSS harmonised standards for ethnicity
• understanding more about sibling groups and placements
• more understanding of the ethnicity and religion of all prospective and approved adopters, and the religion of the child – the strategy notes the numerous challenges involved in this work
• further analysis that could explore the relationship between different aspects of looked-after children and their circumstances, to identify possible causality
• developing age standardised rates for comparisons between children in different ethnic groups
• understanding a wider range of outcome measures disaggregated by ethnicity for looked-after children – this could include understanding outcomes for children who leave care beyond the age of 18 in areas such as, employment, training, housing, healthcare and contact with the criminal justice system

1. The independent review of children’s social care Final report (PDF) and Racial and ethnic disparities in children’s social care (PDF) ↩