The statistics are based on information on looked after children collected by the Department for Education in the SSDA903 return. This is completed annually by local authorities in England. Data is required for two groups of children:
every child who is looked after by the local authority at any time during the year
care leavers (who are children who have been looked after for at least 13 weeks, which began after they reached the age of 14 and ended after they reached the age of 16, who are now aged 17, 18, 19, 20 and 21)
History of the collection
The SSDA903 collection began in 1992 and prior to 1998 it covered all looked after children.
Between 1998 and 2003 the SSDA903 return covered a one third sample of children, those with a date of birth divisible by three. There was also an aggregate return, the CLA 100 return, which was used to provide ‘fast track’ data and to gross the sample data.
In 2000, the OC1 (educational qualifications of care leavers) return was introduced, followed by the AD1 (children adopted) returns in 2001 and the OC3 return (19th birthday) in 2002.
From 2004, the CLA100 return was discontinued and the SSDA903 reverted to cover all looked after children. Becoming a web-based data collection for the first time, the SSDA903 return also incorporated the AD1, OC1 and OC3 returns.
Additional information on offending, health promotion checks, immunisations, dental checks, health assessments and substance misuse problem identification and intervention is also collected through the SSDA903. This is collected for children who have been looked after continuously for at least 12 months at 31 March. This information was previously collected via the Outcome Indicators (OC2) return which contained aggregated local authority level data. In the 2008 to 2009 collection this information was added to the SSDA903 and is now routinely collected at child level.
A more detailed timeline of changes to the SSDA903 collection is given in the Annex.
Historical revisions to the data
The SSDA903 data is collected in a longitudinal database, with one record for every episode of care. Local authorities update the database every year, including making amendments to previous years’ records where there have been changes.
The SSDA903 collects information about the child - for example gender, date of birth and unique pupil number - and details about the child’s episodes in care. When a change in legal status or placement (or both) occurs, a new episode of care is started and the date and reason for the change are recorded.
Revisions to previously published data occur because of corrections made by local authorities to their historical data. Local authorities can directly amend records for the years since 2004 in the system. Common revisions include correcting episode information and including missing records.
The table below gives an indication of how the data for 2019 has changed due to historical revisions between last year’s 2019 publication and the current 2020 publication:
Counts of children looked after in the year ending 31 March 2019, in the 2019 and 2020 statistical releases
Count of children
Reported in 2019 release
Reported in 2020 release
Looked after children at 31 March 2019
Children starting to be looked after during the year ending 31 March 2019
Children ceasing to be looked after during the year ending 31 March 2019
An audit of 2018 to 2019 records which were changed by local authorities in 2020 shows that:
870 child records were deleted
120 child records were added
1,880 master changes were made - these are changes to a child’s date of birth, gender or ethnic group.
A correction to legal status or a date change in any episode of care may mean that the child will no longer be included in a particular set of figures. For example, if a date episode ceased is changed it may mean the child is no longer included in ceased figures for 2019 but is included in ceased figures for 2020 instead. This would mean that the child is counted as looked after at 31 March 2019.
Further information on the data items collected in the SSDA903 return can be found in the collection guide. This and other documents useful for the completion of the SSDA903 return can be found on the DfE children looked after collection website.
A number of automated data validation checks are carried out at the point of data entry to identify and remove:
unlikely or impossible combinations of legal status
unlikely or impossible sequences of dates
information which contradicts data already held about the child, to maintain consistency with data reported in previous years
information which is contradictory within a return, for example OC2 data reported for a child not looked after for 12 months at the end of the year
Any record which fails the validation checks is highlighted and must be corrected. An explanation of each validation check and guidance on how corrections can be made are documented in the list of validation checks available on the DfE children looked after collection website.
Some aggregate return level checks are performed in the system to highlight large changes in return level data compared to the previous year. These are there as prompts for local authorities to check their data. Any outstanding aggregate checks need to have explanatory commentary added to the system before the local authority can sign off their return.
Production of snapshot tables
The annual data reported by each local authority is linked (by a unique identifier for each child) to data supplied by the same local authority in previous years (from 1992 onwards).
At the end of the data collection each year a ‘snapshot’ of the database is taken. This includes all the historic data for each child. These snapshot tables are checked to make sure:
they have been produced correctly from the database
the data is consistent between years
sample checks show they accurately reflect the live database
previous years data is consistent with the previous years' snapshot
These snapshot tables are then processed to produce the data for this statistical release.
Creation of National Pupil Database (NPD) input files for matching to other data
One of the snapshot tables is an ‘Episodes’ table, which contains a row for each episode of care during a child’s period of care. This table is the basis of the input file which is sent to be matched in the NPD.
A file including all children who have been looked after during the year is created for each year beginning 1 April and ending 31 March. Historically, we did this for the most recent six years which allowed us to produce an updated five-year time series in the Outcomes statistical release - six years are needed due to the timing of exclusions data.
In 2019 we reduced the number of years to five - in line with the main statistical release - as we no longer refresh earlier years data for the Outcomes release. Note in 2018 we only refreshed one year - 2017 - in the Outcomes publication.
For a record to be eligible for matching, they must have a valid unique pupil number (UPN) which is the main data item used to match data the SSDA903 data to the NPD.
A number of indicators are created on the child level input files which are sent for matching – these indicators don’t include children that have been looked after in respite care only. The indicators are:
child is looked after on 31 March
child is looked after continuously for at least six months
child is looked after continuously for at least twelve months
child is eligible for Pupil Premium
other variables including child characteristics and placement information.
Some children are looked after in respite care under a series of short-term placements to provide parents or full-time carers with a break. This pattern of respite care is unlikely to impact on the outcomes for a child so children who are looked after in respite care only are excluded from the data.
Before the input file is sent for matching a number of quality assurance checks are completed. These include:
the analysis being dual run independently by two members of the team
checks of summary figures against previous years for consistency
checks against data published in the ‘Children looked after in England (including adoption) year ending 31 March’ statistical release.
checks on the volumes of UPNs by local authority and age.
The UPN allows the looked after children data to be matched in the NPD. Other information such as date of birth, gender, ethnicity and the local authority responsible for the care of the child may also be used to aid matching. It is the responsibility of local authorities to ensure that the UPN and child information provided on the SSDA903 return is accurate, as incorrect data could lead to a mismatch of data in the NPD. Local authorities can amend UPNs for previous years during the collection.
Since 2013, a valid UPN should be returned for all looked after children aged 4 or over at 31 August within the collection year (with the exception of respite cases). Previously UPNs were only required for looked after children aged 6 to 15 years old at 31 August. UPNs can be returned for younger children if they have been assigned one.
There are several reasons why a looked after child may not have a UPN, for example if sources collating the UPNs reflect discrepancies for the child's name and/or surname or date of birth preventing reliable matching or if the child:
is not of school age and has not yet been assigned a UPN (some children not of school age have a UPN depending on the type of setting they attend)
has never attended a school where UPNs are mandatory (for example. some unaccompanied asylum-seeking children or children that have only ever attended an independent school)
is educated outside of England
is newly looked after (from one week before the end of the collection period) and the UPN was not yet known at the time of the SSDA903 return
The outcomes publication does not use an age filter when presenting attainment information - we publish data for children at the end of the key stage so they can be of varying age. However, the matching rates in the tables below use the following ages (at 31 August) to ascertain the degree of matching for each key stage:
age 6 for key stage 1
age 10 for key stage 2
age 15 for key stage 4
This is not a precise measure of data quality as some of these children may have repeated a year and therefore wouldn’t be eligible for key stage exams. Conversely, there will be other children who are eligible for exams who are not aged 6, 10, or 15 years old but will not be included in the key stage percentages in the tables below.
Percentage of children looked after continuously for at least 12 months at 31 March that have a UPN submitted in the SSDA903 return
Note in the table: Data for 2019 is based upon 2019 SSDA903 data at 31 March 2019. Data for 2017 and 2018 are based on 2018 SSDA903 data at 31 March 2017 and 2018. Data for 2015 and 2016 are based on SSDA903 data at 31 March 2015 and 2016.
Children aged 4-15 at 31 August (of previous year)
Children eligible for key stage 1 teacher assessment
Children eligible for key stage 2 tests
Children eligible for key stage 4 tests
The UPN allows the data to be matched to attainment information held in the NPD, which is collected from awarding bodies. The percentage of successful matches is summarised in the table below. Annual processing changes are:
up to and including the 2017 publication the matching process was re-run for historical data each year.
in the 2018 publication the matching process ran for the current and previous year only
from 2019 onwards the matching process ran for the current year only as historically changes to earlier years have been small.
Percentage of children looked after continuously for at least 12 months at 31 March with a UPN that have successfully been matched to attainment data in the NPD
Note in the table: Data for 2019 is based upon 2019 SSDA903 data at 31 March 2019. Data for 2017 and 2018 are based on 2018 SSDA903 data at 31 March 2017 and 2018.
CLA aged 6 that were successfully matched to key stage 1 data
CLA aged 10 that were successfully matched to key stage 2 data
CLA aged 15 that were successfully matched to key stage 4 data
Information on attainment, special educational needs (SEN), absence and exclusions are based on CLA data that has been matched in the NPD to attainment, school census, alternative provision census and where applicable pupil referral unit (PRU) census data.
CLA data is matched to provisional key stage 1 data and revised key stage 2 and key stage 4 data.
School census data
CLA data is matched to the spring school census data which is a snapshot of pupil information in schools as at the third Thursday in January. Pupil level data has been collected via the school census since 2002 (annually). A termly school census was introduced in 2006 for secondary school and since 2007 it has been returned for all other schools (nursery, primary, secondary and special); it has also been returned by pupil referral units since 2014 (the pupil referral unit census was introduced in 2010, but discontinued from 2014 onwards as these institutions now complete the school census). CLA data is also matched to the pupil referral unit census for January 2013. Absence and exclusions data is matched to the school census data in the NPD before the CLA-school census matching is carried out.
Alternative provision census data
CLA data is matched to the alternative provision census which is a snapshot of pupil information in alternative provision at the third Thursday in January. Pupil level information has been returned via the census since January 2008. The alternative provision census covers pupils attending a school not maintained by a local authority for whom the authority is paying full tuition fees, or educated otherwise than in schools and pupil referral units, under arrangements made (and funded) by the local authority.
The Department currently publishes two annual statistical releases:
Children looked after in England (including adoption) (December)
Outcomes for children looked after by local authorities (March/May)
The UK Statistics Authority has designated these statistics as National Statistics, in accordance with the Statistics and Registration Service Act 2007 and signifying compliance with the Code of Practice for Official Statistics.
Designation can be broadly interpreted to mean that the statistics:
meet identified user needs
are well explained and readily accessible
are produced according to sound methods
are managed impartially and objectively in the public interest
Once statistics have been designated as National Statistics it is a statutory requirement that the Code of Practice shall continue to be observed. Since reconfirmation, we have continued to align the statistics to the Code of Practice, for example we have:
continued to facilitate local authority user groups annually to discuss changes to the production of these statistics
improved the commentary supporting the statistics, through the introduction of this CLA statistics guide
minimised the number of releases by releasing all information together at the earliest opportunity. For example, in 2018 we released the main publication data altogether (rather than a main/additional table release) and in 2015 we moved the OC2 outcomes to the main publication rather than delaying its release until the later outcomes publication
developed the outcomes publication to include more comparator information, for example including figures for non-looked after children and children in need 13 alongside the looked after children figures, where comparable figures are available
provided local authority data in machine readable, tidy data format.
Children looked after in England (including adoption) statistical release
The ‘Children looked after in England including adoption’ statistical release provides information - for the year ending 31 March - on:
children looked after at 31 March in each year, including unaccompanied asylum seeking children
children looked after at any time
children who started to be looked after
children who ceased to be looked after
children looked after who were adopted
care leavers now aged 17, 18, 19, 20 and 21 years old
children who went missing or away from placement without authorisation
All figures are based on data collected from the annual SSDA903 return. The longitudinal nature of the collection means historical information can be amended each year, so we update all time series figures published in each release.
Prior to 2013, grossing factors were applied to the data in order to compensate for records with errors in their episodes. These were calculated by comparing the number of valid looked after children records reported in the return with the number of invalid records in the system. In 2013, changes to validation checks on the data submitted improved the quality and grossing factors were close to 1. After a review grossing was discontinued from 2013 onwards.
Up to 2019, the information was published as a series of data tables. From 2020 the information has been published as a series of datasets.
Prior to 2019, some of the statistics were published as experimental as they were based upon recent additions to the collection and local authorities had highlighted some data quality concerns. In 2019 we removed the ‘experimental statistics’ label from these statistics and instead published them as official statistics.
This publication is the main information source on the numbers of looked after children in England; it is used to inform policy decisions to improve outcomes for this vulnerable group of children. It is also used extensively by other organisations interested in this subject.
Outcomes for children looked after by local authorities statistical release
The ‘Outcomes for children looked after by local authorities’ publication provides national and local authority level information on the outcomes for children who have been looked after continuously by local authorities in England for at least 12 months at 31 March and has been published since 2010.
In the ‘Outcomes for children looked after by local authorities’ publication the definition of a ‘looked after child’ is a child who has been continuously looked after for at least 12 months up to and including 31 March.
This definition has been used because 12 months is considered an appropriate length of time to gauge the possible association of being looked after on educational attainment. However, a child may not have been in the care of a local authority for the whole of a key stage period. It is also the cohort of children for whom information on outcomes such as health, wellbeing and offending are collected through the SSDA903
In the ‘Outcomes for children looked after by local authorities’ publication the definition of a ‘non-looked after child’ is a child who has not been looked after continuously for 12 months as at 31 March. This will include both children who have never been looked after and also those who have been looked after but who have not met the 12 months criteria.
Figures published are based on data from the SSDA903 returned matched to data in the NPD. Information is provided on:
educational attainment at key stage 1, key stage 2 and key stage 4
special educational needs (SEN)
absence from school
exclusions from school
From 2015, the publication includes information on educational attainment at key stage 2 and key stage 4 for previously looked after children who left care through an adoption, special guardianship order or child arrangements order.
This publication is the main source of information on the outcomes for looked after children in England and information on the attainment outcomes of previously looked after children who left care through an adoption, special guardianship order or child arrangements order. It is used to inform policy decisions to improve the life chances for this vulnerable group of children. It is also used extensively by other groups interested in this subject and by local authorities for benchmarking purposes.
Throughout the Outcomes release comparisons are made to children in need and non-looked after children where similar figures are available.
Data refreshes - In previous releases we have refreshed earlier years data within each time series for looked after children. This year we have assessed whether this annual refresh is still necessary and concluded that, given there is minimal change (if any) on the figures for earlier years from one release to the next and balancing that with the resource required to do the refreshes that we would no longer update previous years data
Key stage 1 - Changes to the reading, maths and science teacher assessment (TA) frameworks in 2019 means that attainment is not comparable to previous years. So only writing TA can be compared to 2018 (and this changed in 2018, so can’t be compared prior to 2018). As a results we have only published 2018 and 2019 data
Destinations - For the first time this year, data has been included at national level on the destinations of looked after children after key stage 4. This is 2017/18 destinations for the 2016/17 cohort.
Data was refreshed for 2017 only (instead of the whole time series)
At key stage 1, changes made within the 2017/18 writing teacher assessment frameworks mean that judgements made in writing in 2018 are not directly comparable to those made using the previous interim frameworks in 2016 and 2017
At key stage 2, changes made within the 2017/18 writing TA frameworks similarly mean that judgements in 2018 are not directly comparable to those made using the previous interim frameworks in 2016 and 2017. At key stage 2 this also means that figures for reading, writing and mathematics combined in 2018 are not comparable to previous years
In 2017, pupils sat reformed GCSEs in English language, English literature and maths for the first time, graded on a 9 to 1 scale. New GCSEs in other subjects 16 were phased in, first being taught from September 2016 to 2018. Of the reformed GCSE subjects introduced in 2018, only science had a significant change in the structure of the qualifications available; combined science was introduced replacing core and additional. Combined science is a double award GCSE and is graded from 9-9 to 1-1. When comparing the 2018 revised headline and additional measures to the equivalent revised data from 2017, it is important to note any changes to methodology or data changes underpinning the 2018 measures as a result of the introduction of further reformed GCSEs graded on the 9-1 scale which have a higher maximum score than unreformed GCSEs
Since 2018, Attainment 8 had a maximum point score of 90, compared to a maximum of 87 in 2017, as a result of the phased introduction of reformed GCSEs. This difference should be taken into account when considering any change in Attainment 8 score between prior to 2018 and beyond. Due to the introduction of reformed GCSEs and the 9-1 grading scale, this measure will not be comparable between years until all reformed GCSEs are introduced into performance tables in 2020
As announced in July 2017, from 2018 the headline EBacc attainment measure is the EBacc average point score (EBacc APS). This replaces the previous threshold EBacc attainment measure. EBacc APS measures pupils' point scores across the five pillars of the EBacc. This ensures the attainment of all pupils is recognised, not just those at particular grade boundaries, encouraging schools to enter pupils of all abilities, and support them to achieve their full potential.
In 2017, the following changes were made to the publication:
Local authority level data was published as underlying data instead of formatted tables, for the most recent year. This follows the department’s new guidelines for improving the way we produce and publish data for users
At key stage 2, the progress methodology was further refined to allow a greater number of prior attainment groups and better comparisons between pupils working below the standard of the tests
Local authority key stage 2 attainment data was updated to align with the main key stage 2 statistical release. The national level data remains the same as in previous years and excludes pupils with missing or lost test results and pupils where results are suppressed pending the outcome of a maladministration investigation. The local authority level data excludes pupils with lost results but includes those with missing results or pending maladministration. It also excludes children that have recently arrived from overseas. As a result, national level England figures don’t match national figures from local authority data
In 2017, pupils sat reformed GCSEs in English language, English literature and maths for the first time, graded on a 9 to 1 scale. The headline threshold 17 attainment measures use a grade 5 for reformed English and mathematics. When comparing 2017 headline measures to the equivalent revised data from 2016, it is important to note the changes in methodology underpinning the 2017 data. Additional measures were provided using a grade 4 threshold in order to allow comparison to 2016
There were two new tables for looked after children at national level at key stage 4; both provide additional breakdowns of the Progress 8 measure (school type and length of most recent period of care)
Local authority key stage 4 attainment data was updated to align with the main key stage 4 statistical release. The local authority attainment data has children recently arrived from overseas removed. As a result, national level England figures don’t match national figures from local authority data.
In 2016 - changes were made to the publication to bring it in line with the new educational outcome measures, following the primary and secondary school reforms:
Children at the end of key stage 1 and key stage 2 were assessed under the new national curriculum using new tests and interim frameworks for teacher assessment where the expected standard has been raised. The headline measures were changed to reflect the new curriculum and the data is not comparable with that provided for previous years; therefore, no information is provided for years earlier from 2016.
In October 2013, it was announced that a new secondary school accountability system would be implemented from 2016. It includes two new headline measures: Attainment 8 and Progress 8 and the English Baccalaureate measure was published for looked after children for the first time.
In addition at key stage 4, the department has made changes to how English counts in two of the headline measures in 2016: achieving A*-C in both English and Mathematics and achievement of the English Baccalaureate, to align more closely with Attainment 8 and Progress 8.
In September 2015 the definition of a persistent absentee changed. Therefore, from 2016 the publication includes figures based on the new methodology for all years to provide a time series. This is similar to what has been done in the pupil absence in schools publication.
Prior to 2016, the percentage of children looked after continuously for twelve months at 31 March who had a valid unique pupil number (UPN) recorded in the SSDA903 return, was published in the local authority tables for relevant ages. This information was originally published in 2012. The percentage of children with a UPN for these tables is over 90% for the majority of local authorities and so these rates were not published from the 2016 publication onwards.
The 2015 publication was released later (moved from December to March) to allow the use of revised key stage 2 and key stage 4 data, and to include more comparative and contextual data, and additional information on attainment of children who have been adopted from care or left care with a special guardianship order or child arrangements order.
Health and offending information previously collected via the Outcome Indicators (OC2) return which contained aggregated data at local authority level was added to the SSDA903 return in 2008-09 and collected at child level. From 2010, the OC2 data was included in the ‘Outcomes for children looked after by local authorities in England, 31 March’ publication. However, in 2015, it was moved to the ‘Children looked after in England (including adoption)’ publication.
Coverage and duplicates
The numbers of pupils at the end of each key stage in the publication can be different to that reported in the main attainment publications. Some of this difference is due to the coverage within the releases. Generally the coverage for looked after children is greater than the main attainment publications as we aim to include as many looked after children as possible.
Duplicates within the looked after children matched attainment data have been removed by taking the best attainment result for that child.
Historically underlying data has been published alongside national and local authority tables in each publication. Alongside the underlying data there is an accompanying metadata document.
In 2020, the publication has consistent of data sets being released in the Explore Education Statistics platform in a new format. Previously published national and local authority tables have not been produced.
Recent changes to the releases have included:
in 2017 some local authority tables were discontinued in the ‘Children looked after in England (including adoptions)’ release and the data was instead moved to the underlying data
in 2018, some local authority tables were discontinued in the ‘Outcomes for children looked after by local authorities’ release and the figures were moved to the underlying data.
Rounding, suppression and grossing
Rounding and suppression is applied to the data. The National Statistics Code of Practice requires that reasonable steps should be taken to ensure that all published or disseminated statistics produced by the DfE protect confidentiality. The publications follow the DfE policy statement on confidentiality.
Rounding suppression and grossing for 2020 release onwards.
Note: rounding and suppression for the outcomes release will be updated in Spring 2021 when it is published on the Explore Education Statistics platform.
National and regional figures have been rounded to the nearest 10. Local authority figures are unrounded.
Where any number is shown as zero, the original figure submitted was zero.
The following symbols have been used in the releases (updated to align with GSS standards):
‘c’ to protect confidentiality. Secondary suppression may be required
‘z’ for not applicable
‘:’ for not available
‘~’ for negligible – also used for a percentage (<0.5%) that would round to 0 but is not zero
where the numerator or denominator is small then the figure is replaced by ‘c’
they may not sum to 100% due to rounding
they are rounded to whole numbers
In the ‘Children looked after in England including adoption’ publication:
the average duration for adoptions (formerly table E2) are rounded to the nearest month
the average number of missing or away without authorisation incidents (formerly in table G1) have been rounded to one decimal place.
Rounding suppression and grossing for 2019 releases and earlier
National and regional figures have been rounded to the nearest 10.
From 2018 local authority figures are unrounded in statistical tables. Prior to this local authority figures were rounded to the nearest 5 in tables but unrounded in underlying data tables.
Where any number is shown as zero, the original figure submitted was zero. The exception to this is for CLA data matched with NPD data where zero numerators and percentages are also suppressed.
The following symbols have been used in the releases:
‘c’ (prior to 2019 this was a ‘x’) to protect confidentiality. Secondary suppression may be required
‘.’ for not applicable
‘..’ for not available
‘-’ for negligible – also used for a percentage (<0.5%) that would round to 0 but is not zero
where the numerator or denominator is small then the figure is replaced by ‘c’ (prior to 2019 this was a ‘x’)
they may not sum to 100% due to rounding
they are generally rounded to whole numbers but there are some exceptions, for example key stage 4 percentage and exclusions percentages which are rounded in line with the main statistical releases in these subject areas
In the ‘Children looked after in England including adoption’ publication:
the average duration for adoptions (table E2) are rounded to the nearest month
the average number of missing or away without authorisation incidents (table G1) have been rounded to one decimal place.
In the ‘Outcomes for children looked after by local authorities’ publication at:
key stage 2 and key stage 4 attainment averages have been rounded to one decimal place
key stage 4, progress measures are rounded to two decimal places.
Children looked after in England (including adoption) release
Whilst validation and consistency checks are in place to ensure the information published is of high quality, the following should be taken into account when reviewing the statistics on looked after children.
31 March, starts and ceased comparability
The number of looked after children at 31 March each year does not always equate to the number at 31 March the previous year plus the number who started minus the number who ceased to be looked after in the current year. This is because a child is only ever counted once as starting to be looked after during the year, and once as ceasing to be looked after during the year, but in some circumstances a child may enter care and/or leave care more than once during the year. There are also a small number of children who move in and out of respite care during the year for whom the same will apply.
Long-term fostering placements
We are continuing to discuss recording of long-term fostering placements with local authorities at our focus groups - to work towards improving the quality of this data. It is likely the current figures are an under-count of the true figures.
From 2016, all children in an ongoing long-term foster placement on 1 April 2015, or those who change to a long-term foster placement after this date, should have been recorded as such.
The definition of a long-term foster placement came into effect from 1 April 2015 in The Care Planning and Fostering (Miscellaneous Amendments) (England) Regulations 2015 and is defined as:
A “long-term foster placement” means an arrangement made by the responsible authority for the child to be placed with a foster carer where:
(a) the child’s plan for permanence is foster care,
(b) the foster carer has agreed to act as child’s foster parent until the child ceases to be looked after, and
(c) the responsible authority has confirmed the nature of the arrangement to the foster carer, parents and the child.
Local authority level analysis of this data highlighted large differences in the proportion of foster placements reported as being long-term foster placements. In earlier years, quality assurance checks identified a degree of misreporting at a level that has affected our confidence in the figures and so figures for long-term fostering were not included in the main release tables. We are continuing to discuss with local authorities at our focus groups to work towards improving the quality of this data. The issues around accurately capturing this included differences in recording practices, short-term placements move into being a long term placement and system issues (for example being able to ‘roll back’ the system to update, or difficulties due to system links to financial payments).
Children reported in the collection under the long-term fostering category are included in the categories ‘With other foster carer not FFA/concurrent planning’ or ‘With relative or friend, not FFA/concurrent planning’ as appropriate, in the statistical release data tables for 2016 to 2019. The latest time series of the figures are shown below although it's likely they are an under-count of the true figures:
Numbers of children looked after in a long-term fostering placement
Years ending 31 March 2016 to 2020 *
CLA at 31 March in long-term fostering placements
Inside council boundary
With relative or friend
With other foster carer
Outside council boundary
With relative or friend
With other foster carer
Children starting to be CLA in long-term fostering placements
Inside council boundary
With relative or friend
With other foster carer
Outside council boundary
With relative or friend
With other foster carer
Children ceasing to be CLA in long-term fostering placements
Inside council boundary
With relative or friend
With other foster carer
Outside council boundary
With relative or friend
With other foster carer
* Note Figures for 2016 and 2017 relate to figures published and updated in 2019, Figures for 2018 to 2020 are based upon revised figures produced in 2020.
Local authority of placement
There are a minority of children in each year whose local authority of placement is outside England or is not submitted due to confidentiality purposes. The local authority of placement for these children are in the table below. These children are not included in figures in data tables for the number of children externally placed in other local authorities, or figures for local authority net gain of children.
Children looked after at 31 March 2020 who were placed outside England or whose local authority of placement was confidential
LA of placement confidential
Placed outside England
Children freed for adoption
The 2018 data showed that there are no longer any children subject to freeing orders. Applications for freeing orders could not be made on or after 30 December 2005 as they were replaced by placement orders.
Effect of collecting provider unique reference numbers (Ofsted URN) on placement type
The introduction of Ofsted URN for placements on or after 1 April 2015 means users should interpret changes to placement type over time with caution.
Ofsted unique reference number (URN) was collected for settings subject to Ofsted inspections for the first time in the collection year 2015 to 2016. For example, for children’s homes this would be the URN of the individual home, and for foster or adoptive placements this would be the URN of the relevant service or agency providing the placement. This data - which covers around 90% of placements - is required for any relevant placement that was ongoing on 1 April 2015, or for any new placement after 1 April 2015.
The in-built system validation checks the placement type submitted by local authorities corresponds with the placement type for the given URN. This meant there were some corrections of placement types reported by local authorities in 2016 and 2017. In particular, there were cases where children previously recorded as being in residential schools in 2015, were recorded as being in children’s homes in 2016 and 2017 - in many cases this was because the provider may have been dual-registered.
Whilst there are always small annual revisions to previous year figures, it's possible that changes by placement type in these years may be a result of improvements in the accuracy of reporting placement types, due to collecting URN. Older placements are more likely to be misreported - a placement which was coded with the incorrect URN will not have been corrected if it ended in 2016 to 17. So, any decrease, or increase in placements over time should be interpreted cautiously.
Reason for placement change
For episodes that cease due to a change in placement, from 1 April 2015 the SSDA903 collection records the reason for the placement change. A full list of valid codes can be found in the guide. Nationally 13,610 placements (or 24%) were recorded in the ‘Other’ category in 2020 which is similar to the 24% in 2019, but a slight reduction from 25% in 2018, 26% in 2017*and 34% in 2016*. (*Note that percentages refer to the original figures published for each year in question).
Feedback from local authorities suggested that placement changes due to custody were being reported within the ‘other’ category. We introduced a change for the 2019 collection to allow the identification of children changing placement due to custody, however this has had little effect on the numbers reported as ‘other’.
Reason episode ceased 'other'
The ‘reason episode ceased’ codes collected in the SSDA903 record the circumstances when a child ceases to be looked after. If a child ceases to be looked after because he/she has turned 18, this can already be deduced from his/her date of birth. Therefore, the appropriate code is used to capture the destination of the young person.
In 2014, 5,000 episodes (16%) were reported as ceasing with the code ‘Period of being looked after ceased for any other reason’. After discussion with local authority data contacts three new categories were introduced in 2015 to try to capture the reasons for these children leaving care. These categories were ‘Accommodation on remand ended’, ‘Age assessment determined child aged 18 or over’ and ‘Child moved abroad’. Despite the introduction of these new categories, in 2020 there are still 5,910 children ceasing to be looked after (20%) categorised as ‘Care ceased for any other reason’. 4,100 (69%) of these were aged 18 years or over so it’s likely they ceased to be looked after as they reached 18 years of age and the more detailed code has not been provided.
Children missing from care
Information was collected for the first time in 2015 on children who were missing or away from their placement without authorisation for any length of time. Previously, information had only been collected on children who were missing or absent for at least 24 hours.
Given the changes in the way this data has been reported and the problems local authorities have identified in consistently reporting this data, comparisons over time or between local authorities should not be made.
The definitions used reflect the definitions in the statutory guidance on children who run away or go missing from home or care and are:
Missing from care – a looked after child who is not at their placement or the place they are expected to be (e.g. school) and their whereabouts is not known.
Away from placement without authorisation – a looked after child whose whereabouts is known but who is not at their placement or place they are expected to be and the carer has concerns or the incident has been notified to the local authority or the police.
Information is collected on the start and end date of the missing incident. If a child goes missing or away from placement without authorisation more than once in the same day, each separate incident is counted. It is possible that children who were missing were also away from placement without authorisation during the year, and vice versa, so the figures should not be summed to give a total number who were not at their placement during the year.
Users should exercise caution when summing the number of missing periods and away from placement without authorisation periods together, this is because an away from placement without authorisation period could become a missing period if the child’s whereabouts becomes unknown. These will appear as separate incidents in the data, but could be one continuous period where the child is not at his/her placement.
This information was initially published in 2015 as ‘experimental statistics’ given it was the first year in which the data had been collected and, therefore, local authority data collection systems and recording practices were still bedding in. In 2019 we removed this label and instead publish information on children who have gone missing as Official Statistics.
Feedback from local authorities suggests the figures are becoming more robust and some local authorities have continued to report some improvements in recording. However, since 2017 a growing number of local authorities are informing us that they do not record incidents as ‘away without authorisation’ but instead report all incidents as ‘missing’ as their primary source of this information is the police. We estimate this could mean that at a national level there is an overestimate of the number of children with missing incidents of up to 13% in 2020, and an undercount of the number of children with away without authorisation incidents of up to 29%. There is some variation across the country in the way these incidents are reported and so robust comparisons between local authorities are and tot possible. Some of these local authorities submitted some ‘away without authorisation’ information and this has been included in the tables.
Local authorities who have informed us of these difficulties are footnoted in the missing dataset and we continue to work with local authorities to improve the quality of this data. Given the changes in the way this data has been reported, comparisons over time should not be made.
Information on children missing from care is collected separately by police forces. Details of all missing or absent incidents are provided to the UK Missing Persons Unit and they publish annual information. This information is not directly comparable with the latest DfE statistics due to the different collection methods and definitions.
The missing dataset includes figures for the average duration of being missing. Around 9 in 10 of missing and away from placement without authorisation incidents are for 2 days or less. In each case, there are a small number of particularly long durations of being missing reported by local authorities in the data, which affect the value of the mean duration. Both the mean and median have been presented in the table in order to show a better descriptor of the true average value.
Average number of days per missing incident
Average number of days per away from placement without authorisation incident
Note, the exact times a child went missing or away without authorisation or returned are not recorded, only the date, so the duration of missing incidents are calculated to the nearest whole day.
National figures exclude those who were looked after under an agreed series of short-term placements, those who have died since leaving care, those who have returned home to parents or someone with parental responsibility for a continuous period of at least 6 months and those whose care was transferred to another local authority.
The numbers of young people recorded as having returned home for at least 6 months, or having died after ceasing to be looked after in each year, are in the table - note a young person may be recorded in more than one year.
Care leavers who have returned home for at least 6 months, or have died after ceasing to be looked after, 2018 to 2020
2018 returned home for at least 6 months
2019 returned home for at least 6 months
2020 returned home for at least 6 months
c - suppressed data
Suitability of accommodation
It is not possible for local authorities to determine the suitability of the accommodation where the young person was deported, had gone abroad, or their residence was not known. These accommodation types have been removed from the suitability calculations.
Care leavers for whom information is not known
The percentage of known information about activity and accommodation remains high. In 2020, for 17- to 18-year olds the figure was 94%and for 19 to 21-year olds the figure was 91%. It can be difficult to assess the year-on-year changes because of the number of care leavers for whom the information is not known. Tables F1 and F2 present the percentage of care leavers in each activity or accommodation category as a percentage of all young people. The table below shows proportions of 19 to 21 year old care leavers in each category, based on only those where the information is known.
Activity of 19 to 21-year-old care leavers
In higher education, i.e. studies beyond A level
In education other than higher education
In training or employment
Not in education, training or employment owing to illness or disability
Not in education, training or employment owing to pregnancy or parenting
Not in education, training or employment owing to other reasons
Accommodation of 19 to 21-year-old care leavers
With parents or relatives
Semi-independent, transitional accommodation
Residence not known
No fixed abode/homeless
Bed and breakfast
With former foster carers
~ a percentage less than 0.5
17 and 18-year-old care leavers
The cohort for former care leavers was extended in 2016 to include care leavers aged 17 and 18 - in addition to those aged 19, 20 and 21 - who were looked after for a total of at least 13 weeks after their 14th birthday including some time after their 16th birthday.
These figures were initially released as ‘experimental statistics’, however this label was removed in 2019. Analysis suggests that in the first year of collecting the information (2016) there was under reporting by some local authorities for 17 and 18-year-olds who left care on their birthday. We believe under reporting was very small for 17-year-olds but for 18-year-olds we believe we could have been missing data for around 1,000 - 11% of young people. However, whilst we expect under reporting to remain a risk, the same level of under-reporting was not experienced in 2017 and we do not expect this to be an issue going forwards.
Outcomes for children looked after by local authorities
Denominators for educational attainment, special educational needs (SEN), absence and exclusions only include those children for whom a UPN has been supplied and have been matched to the NPD. This methodology was introduced in 2012.
Prior to 2012, percentages were given based on the total number of looked after children with and without UPNs.
Between 2012 and 2016 the percentage of children looked after continuously for twelve months at 31 March who had a UPN recorded in the SSDA903 return was published in the local authority tables. The percentage of children with a UPN was over 90% for the majority of local authorities in 2016 and so from 2016 these rates were discontinued.
Attainment data is given for looked after children in any educational setting who are at the end of the key stage and have key stage data available. Recent changes to attainment data that affect the publication are summarised in the recent history section of this document.
At key stage 4, 2014 data onwards cannot be compared to earlier years due to two major reforms that took place previously – a full breakdown of the reforms and their impact is given in SFR41/2014.
In 2015, SEN breakdowns were provided at national level for key stage 2 and key stage 4 for the first time in the publication. The SEN information was taken from matched school census data. From 2016, SEN information is taken from both the school census and the alternative provision census matched data.
In 2016, the new progress measures at key stage 2 and key stage 4 were based on models that include children in state-funded mainstream schools. However, children in 29 some settings outside of the model have progress scores. In 2017, the model for the key stage 2 progress measures was expanded to include all state-funded schools.
In the national key stage 2 publication, progress measures are published for state-funded schools which are the schools included in the new national model (from 2017). However, within the CLA outcomes publication we include progress measures for all schools that have been included in the progress measure calculations which in addition to state-funded schools includes non-maintained special schools, as we want to cover as many looked after children as possible. We publish this by SEN.
In the national key stage 4 publication, the Progress 8 measure breakdowns are generally published for state-funded schools which is state-funded mainstream schools (those schools in the national model) plus state-funded special schools. However, within the CLA outcomes publication we include progress measures for all schools that have been included in the progress measure calculations which in addition to state-funded mainstream schools includes state-funded special schools, non-maintained special schools and alternative provision, as we want to cover as many looked after children as possible. We publish this by SEN. In addition, we’ve provided a total for state-funded mainstream schools for comparison.
The SEN provision tables relate to looked after children who have been matched to the school census and the alternative provision census. It is possible for a child to be matched to both of these census types if they attend more than one kind of setting. Such cases will be included in the figures but will only count once. In these cases, the highest level of SEN has been counted.
In September 2014 the special educational needs and disability (SEND) reforms came into effect as part of the Children and Families Act 2014. See the SEND code of practice: 0 to 25 for more detailed information on the reforms.
Further national and local authority figures on special educational needs can be found under ‘Special educational needs’ in Explore Education Statistics.
The school census started to collect absence information on a termly basis in 2006 from maintained secondary schools, city technology colleges and academies. From 2007, this was extended to cover maintained primary schools and special schools.
Absence information is collected in arrears and the latest available data is for the 2018 to 2019 school year.
In September 2015 the definition of a persistent absentee changed. In the 2017 Outcomes statistical release, figures based on the new methodology were calculated for previous years, in order to provide a time series. This is similar to what was done in the 2016 pupil absence in schools publication.
From 2017 onwards, figures are based on the number of children looked after continuously for at least 12 months at 31 March who were aged 5-15 at the start of the academic year (31 August) and were matched to the school census. Figures for previous years will be different from previously published figures as these were based on children aged 4-15, and local authorities may update information for past years.
Exclusions data was first collected for the school year 1994 to 1995. The school census started to collect exclusions data on a termly basis in 2006 from maintained secondary schools, city technology colleges and academies. From 2007, this was extended to cover maintained primary schools and special schools.
The school census collects exclusions information two terms in arrears. For example, the summer school census collects information on exclusions occurring during the autumn term. The latest available exclusions data is for the 2017 to 2018 school year.
Adoptions, special guardianship orders and child arrangements orders
The figures on attainment for previously looked after children who left care through an adoption, special guardianship order (SGO) or child arrangements order (CAO) were published as additional tables to the ‘Outcomes for children looked after by local authorities’ publication in 2016. This information was published for the first time in 2015.
The coverage is partial as the cohort information is taken from the school census and alternative provision census, where children are reported as having previously left care due to adoption, a SGO or a CAO only if their parents or carers have declared this information to the school or alternative provision setting. Therefore, the data is reliant on self-declaration from parents and is partial rather than a full count. Users are advised to exercise caution when interpreting these statistics as there can be issues with low levels of coverage which may introduce bias and changes in coverage and attainment measures mean year on year comparisons should not be made. In the 2015 publication, only information from the school census was published.
Since 2016, information from the school census and the alternative provision census have been published. It is not possible to obtain a complete match of the children looked after data (from SSDA903) for those that have left care due to an adoption, a SGO or a CAO with attainment data. Some adopted children do not change their UPN so records could be matched; however, in most cases where children leave care for adoption they have not reached school age and have not been allocated a UPN. In addition, in some instances when some children are adopted their UPN changes in order that no link can be made between the pre-adoption record (with the original UPN) and the post-adoption record (with the new UPN).
Comparing figures reported on the school census and alternative provision census with those from the SSDA903, we calculate in 2019, nearly three-quarters of these children are covered at key stage 2 and just less than half at key stage 4.
Calculated percentage coverage of children who left care through an adoption, SGO or CAO, 2018 to 2019
Only children with a valid UPN that have been matched to attainment data are included in the coverage calculations above. Some children may have been counted more than once in the SSDA903 data if they have left care under these circumstances more than once.
Information on children that left care due to adoption, a SGO or CAO is reported via the school census and alternative provision census for previous years and may not represent the current situation for a child.
The ‘Children looked after in England including adoption’ and the ‘Outcomes for children looked after by local authorities’ publications are produced using a final version of the dataset. We do not plan to make any revisions to the publications. If we later discover that a revision is necessary, this will be made in accordance with the DfE statistical policy statement on revisions.
Information on care proceedings in England: The Children and Family Court Advisory and Support Service (CAFCASS) publish information on care proceedings applications made by local authorities where CAFCASS has been appointed by the court to represent the interests of children.
Local authority children in care and adoption statistics: Children in care and adoption performance tables (published within the local authority interactive tool) and adoption scorecards show how local authorities and looked after children’s services compare with others in England.
Local authority expenditure: Information on local authority planned expenditure on schools, education, children and young people’s services is available. This data is returned to DfE by local authorities via the Section 251 Budget Return.
Ofsted statistics:Ofsted publish inspection outcomes for local authority children’s services and children’s social care providers in England as well as data on the placements of children looked after and fostering in England.
Adoption and Special Guardianship Leadership Board (ASGLB):Quarterly reports are available here.
Ministry of Justice statistics: The Ministry of Justice publish quarterly family court figures which include information on public law cases and adoptions.
Stability Index: The Children's Commissioner publishes the annual Stability Index which is an annual measure of the stability of the lives of children in care.
Data is collected and published independently by each of the four countries in the UK. Although there are similarities between the data collected by the four nations, there are also differences which may be down to different policies and legislation, and differing historical data collections.
Scotland: Statistics on children’s social work (including children looked after) in Scotland and statistics on the outcomes of looked after children in Scotland are available on the Scottish Government website.
Wales: Statistics on children looked after in Wales, adoptions and outcomes are available for the period up to 31 March 2019 on the Welsh Government website.
Northern Ireland: Statistics on children looked after in Northern Ireland and children adopted from care in Northern Ireland are published on the Department of Health website.
Collection switched to one third sample. CLA100 aggregate return used to gross sample data to overall population.
AD1 return for adopted children introduced.
All codes revamped from numeric to alphanumeric
OC3 return for care leavers aged 19 introduced.
Collection reverted to all children. CLA100 discontinued. Collection became web-based for first time. OC1 return on educational achievements introduced,
Placement code set expanded to include temporary placements. Respite care codes split.
Participation in reviews collected.
Freeing orders replaced by placement orders. Reason episode ceased codes expanded to include residence orders and special guardianship orders.
Code set for adoption placements expanded to incorporate placement with consent or placement order. Gender and relationship status of adopters collected. Activity of care leavers expanded to include full-time/part-time activity.
Reason episode ceased expanded to include adoption orders unopposed/consent dispensed, special guardianship orders split into former foster carers/other, sentenced to custody collected.
OC2 return for outcomes for looked after children introduced. Adoption decision return introduced.
LA of placement code set expanded to include other constituent countries. Placement provider collected. SDQ information introduced.
Ethnicity code set revised to align with school census.
Care leaver accommodation code expanded to incorporate care leavers with former foster carers.
UPNs collected for all children of school age. Activity at 16 discontinued.
OC3 expanded to include care leavers aged 19 to 21. Pervious permanence return introduced.
Episode information expanded to include postcode for all placements. In touch information for care leavers redefined and switched from numeric to alphanumeric.
Missing return for children going missing/away without authorisation for any period introduced.
Reason episode ceased expanded for further details. Foster codes extension to include Fostering for Adoption (FFA) and long-term fostering. Relationship status of adopters expanded to include same sex marriage.
OC3 expanded again to include care leavers aged 17 to 21.
Ofsted URN of placement collected. Reason for placement change collected. Reason for new episode code expanded to incorporate children staying with same carer. Activity of care leavers expanded to include Not in employment, education or training (NEET) due to parenting or pregnancy.
Reason episode ceased expanded to cover more information on special guardianship orders (SGOs).